From 6a2c3f2100d030c30b7d8ac8c95dcece7edb040c Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.de>
Date: Tue, 8 Sep 2009 16:04:23 +0200
Subject: api separated into individual components, adapted for new webservice
 versions

---
 lib/compound.rb | 67 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 67 insertions(+)
 create mode 100644 lib/compound.rb

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
new file mode 100644
index 0000000..4652770
--- /dev/null
+++ b/lib/compound.rb
@@ -0,0 +1,67 @@
+module OpenTox
+
+	# uri: /compound/:inchi
+	class Compound < OpenTox
+
+		attr_reader :inchi
+
+		# Initialize with <tt>:uri => uri</tt>, <tt>:smiles => smiles</tt> or <tt>:name => name</tt> (name can be also an InChI/InChiKey, CAS number, etc)
+		def initialize(params)
+			@@cactus_uri="http://cactus.nci.nih.gov/chemical/structure/"
+			if params[:smiles]
+				@inchi = smiles2inchi(params[:smiles])
+				@uri = File.join(@@config[:services]["opentox-dataset"],"compound",@inchi)
+			elsif params[:inchi]
+				@inchi = inchi
+				@uri = File.join(@@config[:services]["opentox-dataset"],"compound",@inchi)
+			elsif params[:name]
+				@inchi = RestClient.get "#{@@cactus_uri}#{params[:name]}/stdinchi"
+				@uri = File.join(@@config[:services]["opentox-dataset"],"compound",@inchi)
+			elsif params[:uri]
+				@inchi = params[:uri].sub(/^.*InChI/, 'InChI')
+				@uri = params[:uri]
+			end
+		end
+
+		# Get the (canonical) smiles
+		def smiles
+			obconversion(@inchi,'inchi','can')
+		end
+
+		def sdf
+			obconversion(@inchi,'inchi','sdf')
+		end
+
+		# Matchs a smarts string
+		def match?(smarts)
+			obconversion = OpenBabel::OBConversion.new
+			obmol = OpenBabel::OBMol.new
+			obconversion.set_in_format('inchi')
+			obconversion.read_string(obmol,@inchi) 
+			smarts_pattern = OpenBabel::OBSmartsPattern.new
+			smarts_pattern.init(smarts)
+			smarts_pattern.match(obmol)
+		end
+
+		# Match an array of smarts features, returns matching features
+		def match(smarts_features)
+			smarts_features.all_features.collect{ |smarts| smarts if self.match?(smarts.name) }.compact
+		end
+
+		def smiles2inchi(smiles)
+			obconversion(smiles,'smi','inchi')
+		end
+
+		def smiles2cansmi(smiles)
+			obconversion(smiles,'smi','can')
+		end
+
+		def obconversion(identifier,input_format,output_format)
+			obconversion = OpenBabel::OBConversion.new
+			obmol = OpenBabel::OBMol.new
+			obconversion.set_in_and_out_formats input_format, output_format
+			obconversion.read_string obmol, identifier
+			obconversion.write_string(obmol).gsub(/\s/,'').chomp
+		end
+	end
+end
-- 
cgit v1.2.3


From 59249a0febc2f90cd1643ddb7e3baa68e3f49065 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.de>
Date: Thu, 10 Sep 2009 15:21:09 +0200
Subject: initial wrapper for new webservices

---
 lib/compound.rb | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 4652770..67c8004 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -44,8 +44,8 @@ module OpenTox
 		end
 
 		# Match an array of smarts features, returns matching features
-		def match(smarts_features)
-			smarts_features.all_features.collect{ |smarts| smarts if self.match?(smarts.name) }.compact
+		def match(smarts_dataset)
+			smarts_dataset.all_features.collect{ |uri| uri if self.match?(Feature.new(:uri => uri).name) }.compact
 		end
 
 		def smiles2inchi(smiles)
-- 
cgit v1.2.3


From 5d5688fbe43806af8523e5481715b5969feeab60 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.de>
Date: Tue, 6 Oct 2009 10:39:43 +0200
Subject: Passes tests in opentox-test

---
 lib/compound.rb | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 67c8004..c4ba8d9 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -10,13 +10,13 @@ module OpenTox
 			@@cactus_uri="http://cactus.nci.nih.gov/chemical/structure/"
 			if params[:smiles]
 				@inchi = smiles2inchi(params[:smiles])
-				@uri = File.join(@@config[:services]["opentox-dataset"],"compound",@inchi)
+				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:inchi]
-				@inchi = inchi
-				@uri = File.join(@@config[:services]["opentox-dataset"],"compound",@inchi)
+				@inchi = params[:inchi]
+				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:name]
 				@inchi = RestClient.get "#{@@cactus_uri}#{params[:name]}/stdinchi"
-				@uri = File.join(@@config[:services]["opentox-dataset"],"compound",@inchi)
+				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:uri]
 				@inchi = params[:uri].sub(/^.*InChI/, 'InChI')
 				@uri = params[:uri]
-- 
cgit v1.2.3


From adfcc9d572a2122b1b030dc04b3abf46007fb3f7 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.de>
Date: Mon, 23 Nov 2009 18:17:37 +0100
Subject: RDF support added

---
 lib/compound.rb | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index c4ba8d9..416acab 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -15,7 +15,7 @@ module OpenTox
 				@inchi = params[:inchi]
 				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:name]
-				@inchi = RestClient.get "#{@@cactus_uri}#{params[:name]}/stdinchi"
+				@inchi = RestClient.get("#{@@cactus_uri}#{params[:name]}/stdinchi").chomp
 				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:uri]
 				@inchi = params[:uri].sub(/^.*InChI/, 'InChI')
@@ -44,8 +44,8 @@ module OpenTox
 		end
 
 		# Match an array of smarts features, returns matching features
-		def match(smarts_dataset)
-			smarts_dataset.all_features.collect{ |uri| uri if self.match?(Feature.new(:uri => uri).name) }.compact
+		def match(smarts_array)
+			smarts_array.collect{|s| s if match?(s)}.compact
 		end
 
 		def smiles2inchi(smiles)
-- 
cgit v1.2.3


From fc3bda0095ca7e6968edf01bec034a243d294af4 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.de>
Date: Mon, 14 Dec 2009 09:41:59 +0100
Subject: Initial adaptation to RDF/OWL

---
 lib/compound.rb | 19 +++++++++++++++----
 1 file changed, 15 insertions(+), 4 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 416acab..9e1cee7 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -1,9 +1,8 @@
 module OpenTox
 
-	# uri: /compound/:inchi
-	class Compound < OpenTox
+	class Compound #< OpenTox
 
-		attr_reader :inchi
+		attr_reader :inchi, :uri
 
 		# Initialize with <tt>:uri => uri</tt>, <tt>:smiles => smiles</tt> or <tt>:name => name</tt> (name can be also an InChI/InChiKey, CAS number, etc)
 		def initialize(params)
@@ -14,6 +13,9 @@ module OpenTox
 			elsif params[:inchi]
 				@inchi = params[:inchi]
 				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
+			elsif params[:sdf]
+				@inchi = sdf2inchi(params[:sdf])
+				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:name]
 				@inchi = RestClient.get("#{@@cactus_uri}#{params[:name]}/stdinchi").chomp
 				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
@@ -48,6 +50,10 @@ module OpenTox
 			smarts_array.collect{|s| s if match?(s)}.compact
 		end
 
+		def sdf2inchi(sdf)
+			obconversion(sdf,'sdf','inchi')
+		end
+
 		def smiles2inchi(smiles)
 			obconversion(smiles,'smi','inchi')
 		end
@@ -61,7 +67,12 @@ module OpenTox
 			obmol = OpenBabel::OBMol.new
 			obconversion.set_in_and_out_formats input_format, output_format
 			obconversion.read_string obmol, identifier
-			obconversion.write_string(obmol).gsub(/\s/,'').chomp
+			case output_format
+			when /smi|can|inchi/
+				obconversion.write_string(obmol).gsub(/\s/,'').chomp
+			else
+				obconversion.write_string(obmol)
+			end
 		end
 	end
 end
-- 
cgit v1.2.3


From 09974c5a920bbb07475d8e48a725ef3780e60dbe Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.de>
Date: Fri, 22 Jan 2010 11:16:13 +0100
Subject: central config.ru, generic compound API

---
 lib/compound.rb | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 4385a7d..0a90663 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -20,14 +20,19 @@ module OpenTox
 				@inchi = RestClient.get("#{@@cactus_uri}#{params[:name]}/stdinchi").chomp
 				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:uri]
-				@inchi = params[:uri].sub(/^.*InChI/, 'InChI')
+				if params[:uri].match(/InChI/) # shortcut for IST services
+					@inchi = params[:uri].sub(/^.*InChI/, 'InChI')
+				else
+					@inchi = RestClient.get @uri, :accept => 'chemical/x-inchi'
+				end
 				@uri = params[:uri]
 			end
 		end
 
 		# Get the (canonical) smiles
 		def smiles
-			obconversion(@inchi,'inchi','can')
+			RestClient.get(@uri, :accept => 'chemical/x-daylight-smiles').split(/\s+/).first # fix ambit output
+			#obconversion(@inchi,'inchi','can')
 		end
 
 		def sdf
-- 
cgit v1.2.3


From c9d820e767a6baf6ae6fba2bb6b2e3641253b5c4 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.de>
Date: Sun, 31 Jan 2010 19:09:16 +0100
Subject: Upload facilities added, Ambit disabled

---
 lib/compound.rb | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 0a90663..562baaa 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -20,19 +20,22 @@ module OpenTox
 				@inchi = RestClient.get("#{@@cactus_uri}#{params[:name]}/stdinchi").chomp
 				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:uri]
+				@uri = params[:uri]
 				if params[:uri].match(/InChI/) # shortcut for IST services
 					@inchi = params[:uri].sub(/^.*InChI/, 'InChI')
 				else
 					@inchi = RestClient.get @uri, :accept => 'chemical/x-inchi'
+					# AMBIT does not provide InChIs
+					#smiles = RestClient.get(@uri, :accept => 'chemical/x-daylight-smiles').split(/\s+/).first # fix ambit output
+					#@inchi = obconversion(smiles,'smi','inchi')
 				end
-				@uri = params[:uri]
 			end
 		end
 
 		# Get the (canonical) smiles
 		def smiles
-			RestClient.get(@uri, :accept => 'chemical/x-daylight-smiles').split(/\s+/).first # fix ambit output
-			#obconversion(@inchi,'inchi','can')
+			#RestClient.get(@uri, :accept => 'chemical/x-daylight-smiles').split(/\s+/).first # fix ambit output
+			obconversion(@inchi,'inchi','can')
 		end
 
 		def sdf
-- 
cgit v1.2.3


From 3d65a42cf68bd5128666d810ec3c29673624ee5d Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.de>
Date: Tue, 2 Feb 2010 17:00:48 +0100
Subject: centralized database configuration

---
 lib/compound.rb | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 562baaa..56646c0 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -21,20 +21,20 @@ module OpenTox
 				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:uri]
 				@uri = params[:uri]
-				if params[:uri].match(/InChI/) # shortcut for IST services
+				case params[:uri]
+				when /ambit/ # Ambit does not deliver InChIs reliably
+					smiles = RestClient.get @uri, :accept => 'chemical/x-daylight-smiles'
+					@inchi = obconversion(smiles,'smi','inchi')
+				when /InChI/ # shortcut for IST services
 					@inchi = params[:uri].sub(/^.*InChI/, 'InChI')
 				else
 					@inchi = RestClient.get @uri, :accept => 'chemical/x-inchi'
-					# AMBIT does not provide InChIs
-					#smiles = RestClient.get(@uri, :accept => 'chemical/x-daylight-smiles').split(/\s+/).first # fix ambit output
-					#@inchi = obconversion(smiles,'smi','inchi')
 				end
 			end
 		end
 
 		# Get the (canonical) smiles
 		def smiles
-			#RestClient.get(@uri, :accept => 'chemical/x-daylight-smiles').split(/\s+/).first # fix ambit output
 			obconversion(@inchi,'inchi','can')
 		end
 
-- 
cgit v1.2.3


From 7b59d1f07bf8cf0e23f3070f97f940af3bfad843 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.de>
Date: Tue, 23 Feb 2010 17:49:02 +0100
Subject: some test issues resolved

---
 lib/compound.rb | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 56646c0..0ee853d 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -17,7 +17,8 @@ module OpenTox
 				@inchi = sdf2inchi(params[:sdf])
 				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:name]
-				@inchi = RestClient.get("#{@@cactus_uri}#{params[:name]}/stdinchi").chomp
+				# paranoid URI encoding to keep SMILES charges and brackets
+				@inchi = RestClient.get("#{@@cactus_uri}#{URI.encode(params[:name], Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").chomp
 				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:uri]
 				@uri = params[:uri]
-- 
cgit v1.2.3


From 9e356d0d94ea4fe210ea7cefce5d4c1179cb63cd Mon Sep 17 00:00:00 2001
From: mguetlein <martin.guetlein@gmail.com>
Date: Thu, 6 May 2010 11:32:05 +0200
Subject: major change: using literal datatypes in owl

---
 lib/compound.rb | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 0ee853d..cc29fc5 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -18,18 +18,18 @@ module OpenTox
 				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:name]
 				# paranoid URI encoding to keep SMILES charges and brackets
-				@inchi = RestClient.get("#{@@cactus_uri}#{URI.encode(params[:name], Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").chomp
+				@inchi = RestClientWrapper.get("#{@@cactus_uri}#{URI.encode(params[:name], Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").chomp
 				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:uri]
 				@uri = params[:uri]
 				case params[:uri]
 				when /ambit/ # Ambit does not deliver InChIs reliably
-					smiles = RestClient.get @uri, :accept => 'chemical/x-daylight-smiles'
+					smiles = RestClientWrapper.get @uri, :accept => 'chemical/x-daylight-smiles'
 					@inchi = obconversion(smiles,'smi','inchi')
 				when /InChI/ # shortcut for IST services
 					@inchi = params[:uri].sub(/^.*InChI/, 'InChI')
 				else
-					@inchi = RestClient.get @uri, :accept => 'chemical/x-inchi'
+					@inchi = RestClientWrapper.get @uri, :accept => 'chemical/x-inchi'
 				end
 			end
 		end
@@ -44,7 +44,7 @@ module OpenTox
 		end
 
 		def image
-			RestClient.get("#{@@cactus_uri}#{@inchi}/image")
+			RestClientWrapper.get("#{@@cactus_uri}#{@inchi}/image")
 		end
 
 		def image_uri
-- 
cgit v1.2.3


From d5eea102ff343804dde1ccd254bc570596341a89 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Fri, 21 May 2010 15:55:17 +0200
Subject: dynamic From: address from mailer

---
 lib/compound.rb | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index cc29fc5..6ab78e0 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -18,7 +18,9 @@ module OpenTox
 				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:name]
 				# paranoid URI encoding to keep SMILES charges and brackets
-				@inchi = RestClientWrapper.get("#{@@cactus_uri}#{URI.encode(params[:name], Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").chomp
+				@inchi = RestClient.get("#{@@cactus_uri}#{URI.encode(params[:name], Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").body.chomp
+				# this was too hard for me to debug and leads to additional errors (ch)
+				#@inchi = RestClientWrapper.get("#{@@cactus_uri}#{URI.encode(params[:name], Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").chomp
 				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:uri]
 				@uri = params[:uri]
-- 
cgit v1.2.3


From 4902d7d4c2d45ef98d20d5a6269e0dbf3bd01cc4 Mon Sep 17 00:00:00 2001
From: ist <ist@ist.opentox.org>
Date: Fri, 11 Jun 2010 17:35:10 +0200
Subject: Added occurrence function 'match_all'

---
 lib/compound.rb | 6 ++++++
 1 file changed, 6 insertions(+)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 6ab78e0..7c7c3d8 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -69,6 +69,12 @@ module OpenTox
 			smarts_array.collect{|s| s if match?(s)}.compact
 		end
 
+		# AM
+		# Match an array of smarts features, returns (0)1 for (non)matching features at each pos
+		def match_all(smarts_array)
+			smarts_array.collect{|s| match?(s) ? 1 : 0 }
+		end
+
 		def sdf2inchi(sdf)
 			obconversion(sdf,'sdf','inchi')
 		end
-- 
cgit v1.2.3


From 3fe0e8c735abd2cacc138f075b5a4af0c6359724 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Wed, 14 Jul 2010 13:56:50 +0200
Subject: model creation working, validation report generation fails for large
 datasets

---
 lib/compound.rb | 13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 7c7c3d8..a293df0 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -1,3 +1,6 @@
+@@cactus_uri="http://cactus.nci.nih.gov/chemical/structure/"
+@@ambit_uri="http://ambit.uni-plovdiv.bg:8080/ambit2/depict/cdk?search="
+
 module OpenTox
 
 	class Compound #< OpenTox
@@ -6,7 +9,6 @@ module OpenTox
 
 		# Initialize with <tt>:uri => uri</tt>, <tt>:smiles => smiles</tt> or <tt>:name => name</tt> (name can be also an InChI/InChiKey, CAS number, etc)
 		def initialize(params)
-			@@cactus_uri="http://cactus.nci.nih.gov/chemical/structure/"
 			if params[:smiles]
 				@inchi = smiles2inchi(params[:smiles])
 				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
@@ -45,12 +47,17 @@ module OpenTox
 			obconversion(@inchi,'inchi','sdf')
 		end
 
-		def image
+		def gif
 			RestClientWrapper.get("#{@@cactus_uri}#{@inchi}/image")
 		end
 
+		def png
+			RestClientWrapper.get("#{@@ambit_uri}#{smiles}")
+		end
+
 		def image_uri
-			"#{@@cactus_uri}#{@inchi}/image"
+			"#{@@ambit_uri}#{smiles}"
+			#"#{@@cactus_uri}#{@inchi}/image"
 		end
 
 		# Matchs a smarts string
-- 
cgit v1.2.3


From 05084b1127f0d2498240472cca8a475f48f1fcc4 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Thu, 22 Jul 2010 10:23:53 +0200
Subject: modifications from first code review meeting

---
 lib/compound.rb | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index a293df0..45b96b8 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -11,19 +11,19 @@ module OpenTox
 		def initialize(params)
 			if params[:smiles]
 				@inchi = smiles2inchi(params[:smiles])
-				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
+				@uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:inchi]
 				@inchi = params[:inchi]
-				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
+				@uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:sdf]
 				@inchi = sdf2inchi(params[:sdf])
-				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
+				@uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:name]
 				# paranoid URI encoding to keep SMILES charges and brackets
 				@inchi = RestClient.get("#{@@cactus_uri}#{URI.encode(params[:name], Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").body.chomp
 				# this was too hard for me to debug and leads to additional errors (ch)
 				#@inchi = RestClientWrapper.get("#{@@cactus_uri}#{URI.encode(params[:name], Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").chomp
-				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
+				@uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:uri]
 				@uri = params[:uri]
 				case params[:uri]
-- 
cgit v1.2.3


From 34a7b50fb278fc37c937f2fe170b86332eb054c5 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Thu, 22 Jul 2010 17:12:51 +0200
Subject: CONFIG reverted to @@config

---
 lib/compound.rb | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 45b96b8..9366db4 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -11,19 +11,19 @@ module OpenTox
 		def initialize(params)
 			if params[:smiles]
 				@inchi = smiles2inchi(params[:smiles])
-				@uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(@inchi))
+				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:inchi]
 				@inchi = params[:inchi]
-				@uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(@inchi))
+				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:sdf]
 				@inchi = sdf2inchi(params[:sdf])
-				@uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(@inchi))
+				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:name]
 				# paranoid URI encoding to keep SMILES charges and brackets
 				@inchi = RestClient.get("#{@@cactus_uri}#{URI.encode(params[:name], Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").body.chomp
 				# this was too hard for me to debug and leads to additional errors (ch)
 				#@inchi = RestClientWrapper.get("#{@@cactus_uri}#{URI.encode(params[:name], Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").chomp
-				@uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(@inchi))
+				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
 			elsif params[:uri]
 				@uri = params[:uri]
 				case params[:uri]
@@ -56,7 +56,8 @@ module OpenTox
 		end
 
 		def image_uri
-			"#{@@ambit_uri}#{smiles}"
+      File.join @uri, "image"
+			#"#{@@ambit_uri}#{smiles}"
 			#"#{@@cactus_uri}#{@inchi}/image"
 		end
 
-- 
cgit v1.2.3


From d6811507c1c1339cc4fe7cdb429b9b34b97dc422 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Fri, 22 Oct 2010 17:45:19 +0200
Subject: new API with support for external services (initial version)

---
 lib/compound.rb | 184 +++++++++++++++++++++++++++++++++++---------------------
 1 file changed, 117 insertions(+), 67 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 49c166f..699e4c1 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -3,82 +3,120 @@
 
 module OpenTox
 
-	class Compound #< OpenTox
-
-		attr_reader :inchi, :uri
-
-		# Initialize with <tt>:uri => uri</tt>, <tt>:smiles => smiles</tt> or <tt>:name => name</tt> (name can be also an InChI/InChiKey, CAS number, etc)
-		def initialize(params)
-			if params[:smiles]
-				@inchi = smiles2inchi(params[:smiles])
-				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
-			elsif params[:inchi]
-				@inchi = params[:inchi]
-				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
-			elsif params[:sdf]
-				@inchi = sdf2inchi(params[:sdf])
-				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
-			elsif params[:name]
-				# paranoid URI encoding to keep SMILES charges and brackets
-				@inchi = RestClient.get("#{@@cactus_uri}#{URI.encode(params[:name], Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").body.chomp
-				# this was too hard for me to debug and leads to additional errors (ch)
-				#@inchi = RestClientWrapper.get("#{@@cactus_uri}#{URI.encode(params[:name], Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").chomp
-				@uri = File.join(@@config[:services]["opentox-compound"],URI.escape(@inchi))
-			elsif params[:uri]
-				@uri = params[:uri]
-				case params[:uri]
-				when /ambit/ # Ambit does not deliver InChIs reliably
-					smiles = RestClientWrapper.get @uri, :accept => 'chemical/x-daylight-smiles'
-					@inchi = obconversion(smiles,'smi','inchi')
-				when /InChI/ # shortcut for IST services
-					@inchi = params[:uri].sub(/^.*InChI/, 'InChI')
-				else
-					@inchi = RestClientWrapper.get @uri, :accept => 'chemical/x-inchi'
-				end
-			end
-		end
+  # Ruby wrapper for OpenTox Compound Webservices (http://opentox.org/dev/apis/api-1.2/structure).
+  # 
+  # Examples:
+  #   require "opentox-ruby-api-wrapper"
+  #
+  #   # Creating compounds
+  #
+  #   # from smiles string
+  #   compound = OpenTox::Compound.from_smiles("c1ccccc1")
+  #   # from name
+  #   compound = OpenTox::Compound.from_name("Benzene")
+  #   # from uri
+  #   compound = OpenTox::Compound.new("http://webservices.in-silico.ch/compound/InChI=1S/C6H6/c1-2-4-6-5-3-1/h1-6H"")
+  #
+  #   # Getting compound representations
+  #
+  #   # get InChI
+  #   inchi = compound.inchi
+  #   # get all compound names
+  #   names = compound.names
+  #   # get png image
+  #   image = compound.png
+  #   # get uri
+  #   uri = compound.uri
+  #
+  #   # SMARTS matching
+  #
+  #   # match a smarts string
+  #   compound.match?("cN") # returns false
+  #   # match an array of smarts strings
+  #   compound.match(['cc','cN']) # returns ['cc']
+	class Compound 
+
+		attr_accessor :inchi, :uri
+
+		# Create compound with optional uri
+		def initialize(uri=nil)
+      @uri = uri
+      case @uri
+      when /InChI/ # shortcut for IST services
+        @inchi = @uri.sub(/^.*InChI/, 'InChI')
+      else
+        @inchi = RestClientWrapper.get(@uri, :accept => 'chemical/x-inchi').to_s.chomp if @uri
+      end
+    end
 
-		# Get the (canonical) smiles
+    # Create a compound from smiles string
+    def self.from_smiles(smiles)
+      c = Compound.new
+      c.inchi = Compound.smiles2inchi(smiles)
+      c.uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(c.inchi))
+      c
+    end
+
+    # Create a compound from inchi string
+    def self.from_inchi(inchi)
+      c = Compound.new
+      c.inchi = inchi
+      c.uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(c.inchi))
+      c
+    end
+
+    # Create a compound from sdf string
+    def self.from_sdf(sdf)
+      c = Compound.new
+      c.inchi = Compound.sdf2inchi(sdf)
+      c.uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(c.inchi))
+      c
+    end
+
+    # Create a compound from name (name can be also an InChI/InChiKey, CAS number, etc)
+    def self.from_name(name)
+      c = Compound.new
+      # paranoid URI encoding to keep SMILES charges and brackets
+      c.inchi = RestClientWrapper.get("#{@@cactus_uri}#{URI.encode(name, Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").to_s.chomp
+      c.uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(c.inchi))
+      c
+    end
+
+		# Get (canonical) smiles
 		def smiles
-			obconversion(@inchi,'inchi','can')
+			Compound.obconversion(@inchi,'inchi','can')
 		end
 
+    # Get sdf
 		def sdf
-			obconversion(@inchi,'inchi','sdf')
+			Compound.obconversion(@inchi,'inchi','sdf')
 		end
 
+    # Get gif image
 		def gif
 			RestClientWrapper.get("#{@@cactus_uri}#{@inchi}/image")
 		end
 
+    # Get png image
 		def png
       RestClientWrapper.get(File.join @uri, "image")
 		end
 
+    # Get URI of compound image
+		def image_uri
+      File.join @uri, "image"
+		end
+
+    # Get all known compound names
 		def names
       begin
-        RestClientWrapper.get("#{@@cactus_uri}#{@inchi}/names")
+        RestClientWrapper.get("#{@@cactus_uri}#{@inchi}/names").split("\n")
       rescue
         "not available"
       end
 		end
 
-    def display_smarts_uri(activating, deactivating, highlight = nil)
-      LOGGER.debug activating.to_yaml unless activating.nil?
-      activating_smarts = URI.encode "\"#{activating.join("\"/\"")}\""
-      deactivating_smarts = URI.encode "\"#{deactivating.join("\"/\"")}\""
-      if highlight.nil?
-        File.join @@config[:services]["opentox-compound"], "smiles", URI.encode(smiles), "smarts/activating", URI.encode(activating_smarts),"deactivating", URI.encode(deactivating_smarts)
-      else
-        File.join @@config[:services]["opentox-compound"], "smiles", URI.encode(smiles), "smarts/activating", URI.encode(activating_smarts),"deactivating", URI.encode(deactivating_smarts), "highlight", URI.encode(highlight)
-      end
-    end
-
-		def image_uri
-      File.join @uri, "image"
-		end
-
-		# Matchs a smarts string
+		# Match a smarts string
 		def match?(smarts)
 			obconversion = OpenBabel::OBConversion.new
 			obmol = OpenBabel::OBMol.new
@@ -89,30 +127,42 @@ module OpenTox
 			smarts_pattern.match(obmol)
 		end
 
-		# Match an array of smarts features, returns matching features
+		# Match an array of smarts strings, returns array with matching smarts
 		def match(smarts_array)
 			smarts_array.collect{|s| s if match?(s)}.compact
 		end
 
-		# AM
-		# Match an array of smarts features, returns (0)1 for (non)matching features at each pos
-		def match_all(smarts_array)
-			smarts_array.collect{|s| match?(s) ? 1 : 0 }
-		end
+    # Get URI of compound image with highlighted fragments
+    def matching_smarts_image_uri(activating, deactivating, highlight = nil)
+      activating_smarts = URI.encode "\"#{activating.join("\"/\"")}\""
+      deactivating_smarts = URI.encode "\"#{deactivating.join("\"/\"")}\""
+      if highlight.nil?
+        File.join CONFIG[:services]["opentox-compound"], "smiles", URI.encode(smiles), "smarts/activating", URI.encode(activating_smarts),"deactivating", URI.encode(deactivating_smarts)
+      else
+        File.join CONFIG[:services]["opentox-compound"], "smiles", URI.encode(smiles), "smarts/activating", URI.encode(activating_smarts),"deactivating", URI.encode(deactivating_smarts), "highlight", URI.encode(highlight)
+      end
+    end
+
+
+    private
 
-		def sdf2inchi(sdf)
-			obconversion(sdf,'sdf','inchi')
+    # Convert sdf to inchi
+		def self.sdf2inchi(sdf)
+			Compound.obconversion(sdf,'sdf','inchi')
 		end
 
-		def smiles2inchi(smiles)
-			obconversion(smiles,'smi','inchi')
+    # Convert smiles to inchi
+		def self.smiles2inchi(smiles)
+			Compound.obconversion(smiles,'smi','inchi')
 		end
 
-		def smiles2cansmi(smiles)
-			obconversion(smiles,'smi','can')
+    # Convert smiles to canonical smiles
+		def self.smiles2cansmi(smiles)
+			Compound.obconversion(smiles,'smi','can')
 		end
 
-		def obconversion(identifier,input_format,output_format)
+    # Convert identifier from OpenBabel input_format to OpenBabel output_format
+		def self.obconversion(identifier,input_format,output_format)
 			obconversion = OpenBabel::OBConversion.new
 			obmol = OpenBabel::OBMol.new
 			obconversion.set_in_and_out_formats input_format, output_format
-- 
cgit v1.2.3


From b93002b4ea50ff7e357da08abd10577347ce2d5f Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Thu, 11 Nov 2010 09:31:27 +0100
Subject: first steps towards version 2.0, yard documentation started, passes
 compound, dataset, feature, algorithm, fminer tests

---
 lib/compound.rb | 105 ++++++++++++++++++++++++++++++++------------------------
 1 file changed, 60 insertions(+), 45 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 699e4c1..6834860 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -4,41 +4,15 @@
 module OpenTox
 
   # Ruby wrapper for OpenTox Compound Webservices (http://opentox.org/dev/apis/api-1.2/structure).
-  # 
-  # Examples:
-  #   require "opentox-ruby-api-wrapper"
-  #
-  #   # Creating compounds
-  #
-  #   # from smiles string
-  #   compound = OpenTox::Compound.from_smiles("c1ccccc1")
-  #   # from name
-  #   compound = OpenTox::Compound.from_name("Benzene")
-  #   # from uri
-  #   compound = OpenTox::Compound.new("http://webservices.in-silico.ch/compound/InChI=1S/C6H6/c1-2-4-6-5-3-1/h1-6H"")
-  #
-  #   # Getting compound representations
-  #
-  #   # get InChI
-  #   inchi = compound.inchi
-  #   # get all compound names
-  #   names = compound.names
-  #   # get png image
-  #   image = compound.png
-  #   # get uri
-  #   uri = compound.uri
-  #
-  #   # SMARTS matching
-  #
-  #   # match a smarts string
-  #   compound.match?("cN") # returns false
-  #   # match an array of smarts strings
-  #   compound.match(['cc','cN']) # returns ['cc']
 	class Compound 
 
 		attr_accessor :inchi, :uri
 
 		# Create compound with optional uri
+    # @example
+    #   compound = OpenTox::Compound.new("http://webservices.in-silico.ch/compound/InChI=1S/C6H6/c1-2-4-6-5-3-1/h1-6H"")
+    # @param [optional, String] uri Compound URI
+    # @return [OpenTox::Compound] Compound
 		def initialize(uri=nil)
       @uri = uri
       case @uri
@@ -50,6 +24,10 @@ module OpenTox
     end
 
     # Create a compound from smiles string
+    # @example
+    #   compound = OpenTox::Compound.from_smiles("c1ccccc1")
+    # @param [String] smiles Smiles string
+    # @return [OpenTox::Compound] Compound
     def self.from_smiles(smiles)
       c = Compound.new
       c.inchi = Compound.smiles2inchi(smiles)
@@ -58,6 +36,8 @@ module OpenTox
     end
 
     # Create a compound from inchi string
+    # @param [String] smiles InChI string
+    # @return [OpenTox::Compound] Compound
     def self.from_inchi(inchi)
       c = Compound.new
       c.inchi = inchi
@@ -66,6 +46,8 @@ module OpenTox
     end
 
     # Create a compound from sdf string
+    # @param [String] smiles SDF string
+    # @return [OpenTox::Compound] Compound
     def self.from_sdf(sdf)
       c = Compound.new
       c.inchi = Compound.sdf2inchi(sdf)
@@ -73,7 +55,11 @@ module OpenTox
       c
     end
 
-    # Create a compound from name (name can be also an InChI/InChiKey, CAS number, etc)
+    # Create a compound from name. Relies on an external service for name lookups.
+    # @example
+    #   compound = OpenTox::Compound.from_name("Benzene")
+    # @param [String] name name can be also an InChI/InChiKey, CAS number, etc
+    # @return [OpenTox::Compound] Compound
     def self.from_name(name)
       c = Compound.new
       # paranoid URI encoding to keep SMILES charges and brackets
@@ -83,32 +69,42 @@ module OpenTox
     end
 
 		# Get (canonical) smiles
-		def smiles
+    # @return [String] Smiles string
+		def to_smiles
 			Compound.obconversion(@inchi,'inchi','can')
 		end
 
     # Get sdf
-		def sdf
+    # @return [String] SDF string
+		def to_sdf
 			Compound.obconversion(@inchi,'inchi','sdf')
 		end
 
     # Get gif image
-		def gif
+    # @return [image/gif] Image data
+		def to_gif
 			RestClientWrapper.get("#{@@cactus_uri}#{@inchi}/image")
 		end
 
     # Get png image
-		def png
+    # @example
+    #   image = compound.to_png
+    # @return [image/png] Image data
+		def to_png
       RestClientWrapper.get(File.join @uri, "image")
 		end
 
     # Get URI of compound image
-		def image_uri
+    # @return [String] Compound image URI
+		def to_image_uri
       File.join @uri, "image"
 		end
 
-    # Get all known compound names
-		def names
+    # Get all known compound names. Relies on an external service for name lookups.
+    # @example
+    #   names = compound.to_names
+    # @return [String] Compound names
+		def to_names
       begin
         RestClientWrapper.get("#{@@cactus_uri}#{@inchi}/names").split("\n")
       rescue
@@ -117,6 +113,10 @@ module OpenTox
 		end
 
 		# Match a smarts string
+    # @example
+    #   compound = OpenTox::Compound.from_name("Benzene")
+    #   compound.match?("cN") # returns false
+    # @param [String] smarts Smarts string
 		def match?(smarts)
 			obconversion = OpenBabel::OBConversion.new
 			obmol = OpenBabel::OBMol.new
@@ -128,19 +128,34 @@ module OpenTox
 		end
 
 		# Match an array of smarts strings, returns array with matching smarts
+    # @example
+    #   compound = OpenTox::Compound.from_name("Benzene")
+    #   compound.match(['cc','cN']) # returns ['cc']
+    # @param [Array] smarts_array Array with Smarts strings
+    # @return [Array] Array with matching Smarts strings
 		def match(smarts_array)
-			smarts_array.collect{|s| s if match?(s)}.compact
+      # avoid recreation of OpenBabel objects
+			obconversion = OpenBabel::OBConversion.new
+			obmol = OpenBabel::OBMol.new
+			obconversion.set_in_format('inchi')
+			obconversion.read_string(obmol,@inchi) 
+			smarts_pattern = OpenBabel::OBSmartsPattern.new
+			smarts_array.collect do |smarts|
+        smarts_pattern.init(smarts)
+        smarts if smarts_pattern.match(obmol)
+      end.compact
+      #smarts_array.collect { |s| s if match?(s)}.compact
 		end
 
     # Get URI of compound image with highlighted fragments
-    def matching_smarts_image_uri(activating, deactivating, highlight = nil)
+    #
+    # @param [Array] activating Array with activating Smarts strings
+    # @param [Array] deactivating Array with deactivating Smarts strings
+    # @return [String] URI for compound image with highlighted fragments
+    def matching_smarts_image_uri(activating, deactivating)
       activating_smarts = URI.encode "\"#{activating.join("\"/\"")}\""
       deactivating_smarts = URI.encode "\"#{deactivating.join("\"/\"")}\""
-      if highlight.nil?
-        File.join CONFIG[:services]["opentox-compound"], "smiles", URI.encode(smiles), "smarts/activating", URI.encode(activating_smarts),"deactivating", URI.encode(deactivating_smarts)
-      else
-        File.join CONFIG[:services]["opentox-compound"], "smiles", URI.encode(smiles), "smarts/activating", URI.encode(activating_smarts),"deactivating", URI.encode(deactivating_smarts), "highlight", URI.encode(highlight)
-      end
+      File.join @uri, "smarts/activating", URI.encode(activating_smarts),"deactivating", URI.encode(deactivating_smarts)
     end
 
 
-- 
cgit v1.2.3


From 64e5d2890f42cef112fbe768f6bfd54b746686de Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Fri, 3 Dec 2010 14:37:09 +0100
Subject: underscore method added to String class

---
 lib/compound.rb | 6 ++++++
 1 file changed, 6 insertions(+)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 6834860..a85507b 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -68,6 +68,12 @@ module OpenTox
       c
     end
 
+		# Get InChI
+    # @return [String] InChI string
+		def to_inchi
+      @inchi
+		end
+
 		# Get (canonical) smiles
     # @return [String] Smiles string
 		def to_smiles
-- 
cgit v1.2.3


From 64354959e04fcac11bcf70e75099691b74573033 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Tue, 16 Aug 2011 14:46:31 +0200
Subject: initial minimal version

---
 lib/compound.rb | 21 +++++++--------------
 1 file changed, 7 insertions(+), 14 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index a85507b..189db7b 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -29,30 +29,21 @@ module OpenTox
     # @param [String] smiles Smiles string
     # @return [OpenTox::Compound] Compound
     def self.from_smiles(smiles)
-      c = Compound.new
-      c.inchi = Compound.smiles2inchi(smiles)
-      c.uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(c.inchi))
-      c
+      Compound.new  RestClientWrapper.post(CONFIG[:services]["opentox-compound"], smiles, :content_type => 'chemical/x-daylight-smiles').to_s.chomp 
     end
 
     # Create a compound from inchi string
     # @param [String] smiles InChI string
     # @return [OpenTox::Compound] Compound
     def self.from_inchi(inchi)
-      c = Compound.new
-      c.inchi = inchi
-      c.uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(c.inchi))
-      c
+      Compound.new  RestClientWrapper.post(CONFIG[:services]["opentox-compound"], inchi, :content_type => 'chemical/x-inchi').to_s.chomp 
     end
 
     # Create a compound from sdf string
     # @param [String] smiles SDF string
     # @return [OpenTox::Compound] Compound
     def self.from_sdf(sdf)
-      c = Compound.new
-      c.inchi = Compound.sdf2inchi(sdf)
-      c.uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(c.inchi))
-      c
+      Compound.new  RestClientWrapper.post(CONFIG[:services]["opentox-compound"], sdf, :content_type => 'chemical/x-mdl-sdfile').to_s.chomp 
     end
 
     # Create a compound from name. Relies on an external service for name lookups.
@@ -77,13 +68,13 @@ module OpenTox
 		# Get (canonical) smiles
     # @return [String] Smiles string
 		def to_smiles
-			Compound.obconversion(@inchi,'inchi','can')
+      RestClientWrapper.get(@uri, :accept => 'chemical/x-daylight-smiles').chomp
 		end
 
     # Get sdf
     # @return [String] SDF string
 		def to_sdf
-			Compound.obconversion(@inchi,'inchi','sdf')
+      RestClientWrapper.get(@uri, :accept => 'chemical/x-mdl-sdfile').chomp
 		end
 
     # Get gif image
@@ -118,6 +109,7 @@ module OpenTox
       end
 		end
 
+=begin
 		# Match a smarts string
     # @example
     #   compound = OpenTox::Compound.from_name("Benzene")
@@ -195,5 +187,6 @@ module OpenTox
 				obconversion.write_string(obmol)
 			end
 		end
+=end
 	end
 end
-- 
cgit v1.2.3


From ac54997dccc571471a0cdf62939e2fcbc42e06e2 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Fri, 13 Jan 2012 12:49:09 +0100
Subject: compound added

---
 lib/compound.rb | 39 ++++++++++-----------------------------
 1 file changed, 10 insertions(+), 29 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 189db7b..974e46e 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -6,44 +6,27 @@ module OpenTox
   # Ruby wrapper for OpenTox Compound Webservices (http://opentox.org/dev/apis/api-1.2/structure).
 	class Compound 
 
-		attr_accessor :inchi, :uri
-
-		# Create compound with optional uri
-    # @example
-    #   compound = OpenTox::Compound.new("http://webservices.in-silico.ch/compound/InChI=1S/C6H6/c1-2-4-6-5-3-1/h1-6H"")
-    # @param [optional, String] uri Compound URI
-    # @return [OpenTox::Compound] Compound
-		def initialize(uri=nil)
-      @uri = uri
-      case @uri
-      when /InChI/ # shortcut for IST services
-        @inchi = @uri.sub(/^.*InChI/, 'InChI')
-      else
-        @inchi = RestClientWrapper.get(@uri, :accept => 'chemical/x-inchi').to_s.chomp if @uri
-      end
-    end
-
     # Create a compound from smiles string
     # @example
     #   compound = OpenTox::Compound.from_smiles("c1ccccc1")
     # @param [String] smiles Smiles string
     # @return [OpenTox::Compound] Compound
-    def self.from_smiles(smiles)
-      Compound.new  RestClientWrapper.post(CONFIG[:services]["opentox-compound"], smiles, :content_type => 'chemical/x-daylight-smiles').to_s.chomp 
+    def self.from_smiles service_uri, smiles, subjectid=nil
+      Compound.new  RestClientWrapper.post(service_uri, smiles, :content_type => 'chemical/x-daylight-smiles').to_s.chomp 
     end
 
     # Create a compound from inchi string
     # @param [String] smiles InChI string
     # @return [OpenTox::Compound] Compound
-    def self.from_inchi(inchi)
-      Compound.new  RestClientWrapper.post(CONFIG[:services]["opentox-compound"], inchi, :content_type => 'chemical/x-inchi').to_s.chomp 
+    def self.from_inchi(service_uri, inchi)
+      Compound.new  RestClientWrapper.post(service_uri, inchi, :content_type => 'chemical/x-inchi').to_s.chomp 
     end
 
     # Create a compound from sdf string
     # @param [String] smiles SDF string
     # @return [OpenTox::Compound] Compound
-    def self.from_sdf(sdf)
-      Compound.new  RestClientWrapper.post(CONFIG[:services]["opentox-compound"], sdf, :content_type => 'chemical/x-mdl-sdfile').to_s.chomp 
+    def self.from_sdf(service_uri, sdf)
+      Compound.new  RestClientWrapper.post(service_uri, sdf, :content_type => 'chemical/x-mdl-sdfile').to_s.chomp 
     end
 
     # Create a compound from name. Relies on an external service for name lookups.
@@ -51,18 +34,16 @@ module OpenTox
     #   compound = OpenTox::Compound.from_name("Benzene")
     # @param [String] name name can be also an InChI/InChiKey, CAS number, etc
     # @return [OpenTox::Compound] Compound
-    def self.from_name(name)
-      c = Compound.new
+    def self.from_name(service_uri, name)
       # paranoid URI encoding to keep SMILES charges and brackets
-      c.inchi = RestClientWrapper.get("#{@@cactus_uri}#{URI.encode(name, Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").to_s.chomp
-      c.uri = File.join(CONFIG[:services]["opentox-compound"],URI.escape(c.inchi))
-      c
+      inchi = RestClientWrapper.get("#{@@cactus_uri}#{URI.encode(name, Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").to_s.chomp
+      Compound.new File.join(service_uri,URI.escape(inchi))
     end
 
 		# Get InChI
     # @return [String] InChI string
 		def to_inchi
-      @inchi
+      RestClientWrapper.get(@uri, :accept => 'chemical/x-inchi').to_s.chomp if @uri
 		end
 
 		# Get (canonical) smiles
-- 
cgit v1.2.3


From 354aaa649e9eeed5d81793e09d9714b45063c147 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Wed, 8 Feb 2012 13:14:11 +0100
Subject: toxbank-investigation compatible version

---
 lib/compound.rb | 31 ++++++++++++++-----------------
 1 file changed, 14 insertions(+), 17 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 974e46e..8761d50 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -1,5 +1,4 @@
-@@cactus_uri="http://cactus.nci.nih.gov/chemical/structure/"
-@@ambit_uri="http://ambit.uni-plovdiv.bg:8080/ambit2/depict/cdk?search="
+CACTUS_URI="http://cactus.nci.nih.gov/chemical/structure/"
 
 module OpenTox
 
@@ -12,21 +11,21 @@ module OpenTox
     # @param [String] smiles Smiles string
     # @return [OpenTox::Compound] Compound
     def self.from_smiles service_uri, smiles, subjectid=nil
-      Compound.new  RestClientWrapper.post(service_uri, smiles, :content_type => 'chemical/x-daylight-smiles').to_s.chomp 
+      Compound.new RestClient.post(service_uri, smiles, {:content_type => 'chemical/x-daylight-smiles', :subjectid => subjectid})
     end
 
     # Create a compound from inchi string
     # @param [String] smiles InChI string
     # @return [OpenTox::Compound] Compound
-    def self.from_inchi(service_uri, inchi)
-      Compound.new  RestClientWrapper.post(service_uri, inchi, :content_type => 'chemical/x-inchi').to_s.chomp 
+    def self.from_inchi service_uri, inchi, subjectid=nil
+      Compound.new RestClient.post(service_uri, inchi, {:content_type => 'chemical/x-inchi', :subjectid => subjectid})
     end
 
     # Create a compound from sdf string
     # @param [String] smiles SDF string
     # @return [OpenTox::Compound] Compound
-    def self.from_sdf(service_uri, sdf)
-      Compound.new  RestClientWrapper.post(service_uri, sdf, :content_type => 'chemical/x-mdl-sdfile').to_s.chomp 
+    def self.from_sdf service_uri, sdf, subjectid=nil
+      Compound.new RestClient.post(service_uri, sdf, {:content_type => 'chemical/x-mdl-sdfile', :subjectid => subjectid})
     end
 
     # Create a compound from name. Relies on an external service for name lookups.
@@ -34,34 +33,32 @@ module OpenTox
     #   compound = OpenTox::Compound.from_name("Benzene")
     # @param [String] name name can be also an InChI/InChiKey, CAS number, etc
     # @return [OpenTox::Compound] Compound
-    def self.from_name(service_uri, name)
-      # paranoid URI encoding to keep SMILES charges and brackets
-      inchi = RestClientWrapper.get("#{@@cactus_uri}#{URI.encode(name, Regexp.new("[^#{URI::PATTERN::UNRESERVED}]"))}/stdinchi").to_s.chomp
-      Compound.new File.join(service_uri,URI.escape(inchi))
+    def self.from_name service_uri, name, subjectid=nil
+      Compound.new RestClient.post(service_uri, name, {:content_type => 'text/plain', :subjectid => subjectid})
     end
 
 		# Get InChI
     # @return [String] InChI string
 		def to_inchi
-      RestClientWrapper.get(@uri, :accept => 'chemical/x-inchi').to_s.chomp if @uri
+      get(:accept => 'chemical/x-inchi').to_s.chomp if @uri
 		end
 
 		# Get (canonical) smiles
     # @return [String] Smiles string
 		def to_smiles
-      RestClientWrapper.get(@uri, :accept => 'chemical/x-daylight-smiles').chomp
+      get(:accept => 'chemical/x-daylight-smiles').chomp
 		end
 
     # Get sdf
     # @return [String] SDF string
 		def to_sdf
-      RestClientWrapper.get(@uri, :accept => 'chemical/x-mdl-sdfile').chomp
+      get(:accept => 'chemical/x-mdl-sdfile').chomp
 		end
 
     # Get gif image
     # @return [image/gif] Image data
 		def to_gif
-			RestClientWrapper.get("#{@@cactus_uri}#{@inchi}/image")
+			get("#{CACTUS_URI}#{to_inchi}/image")
 		end
 
     # Get png image
@@ -69,7 +66,7 @@ module OpenTox
     #   image = compound.to_png
     # @return [image/png] Image data
 		def to_png
-      RestClientWrapper.get(File.join @uri, "image")
+      get(File.join @uri, "image")
 		end
 
     # Get URI of compound image
@@ -84,7 +81,7 @@ module OpenTox
     # @return [String] Compound names
 		def to_names
       begin
-        RestClientWrapper.get("#{@@cactus_uri}#{@inchi}/names").split("\n")
+        get("#{CACTUS_URI}#{to_inchi}/names").split("\n")
       rescue
         "not available"
       end
-- 
cgit v1.2.3


From fa9069e13fb6b1c8bb4ebcdf82f1cf1c04ad71ca Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Thu, 23 Feb 2012 17:56:46 +0000
Subject: (partially) switched back to RestClientWrapper

---
 lib/compound.rb | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 8761d50..ce0fdbf 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -11,21 +11,21 @@ module OpenTox
     # @param [String] smiles Smiles string
     # @return [OpenTox::Compound] Compound
     def self.from_smiles service_uri, smiles, subjectid=nil
-      Compound.new RestClient.post(service_uri, smiles, {:content_type => 'chemical/x-daylight-smiles', :subjectid => subjectid})
+      Compound.new RestClientWrapper.post(service_uri, smiles, {:content_type => 'chemical/x-daylight-smiles', :subjectid => subjectid})
     end
 
     # Create a compound from inchi string
     # @param [String] smiles InChI string
     # @return [OpenTox::Compound] Compound
     def self.from_inchi service_uri, inchi, subjectid=nil
-      Compound.new RestClient.post(service_uri, inchi, {:content_type => 'chemical/x-inchi', :subjectid => subjectid})
+      Compound.new RestClientWrapper.post(service_uri, inchi, {:content_type => 'chemical/x-inchi', :subjectid => subjectid})
     end
 
     # Create a compound from sdf string
     # @param [String] smiles SDF string
     # @return [OpenTox::Compound] Compound
     def self.from_sdf service_uri, sdf, subjectid=nil
-      Compound.new RestClient.post(service_uri, sdf, {:content_type => 'chemical/x-mdl-sdfile', :subjectid => subjectid})
+      Compound.new RestClientWrapper.post(service_uri, sdf, {:content_type => 'chemical/x-mdl-sdfile', :subjectid => subjectid})
     end
 
     # Create a compound from name. Relies on an external service for name lookups.
@@ -34,7 +34,7 @@ module OpenTox
     # @param [String] name name can be also an InChI/InChiKey, CAS number, etc
     # @return [OpenTox::Compound] Compound
     def self.from_name service_uri, name, subjectid=nil
-      Compound.new RestClient.post(service_uri, name, {:content_type => 'text/plain', :subjectid => subjectid})
+      Compound.new RestClientWrapper.post(service_uri, name, {:content_type => 'text/plain', :subjectid => subjectid})
     end
 
 		# Get InChI
-- 
cgit v1.2.3


From bf6834445feb6f93f0a20359462dbd1e7e89f4b8 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Thu, 12 Jul 2012 16:38:03 +0200
Subject: all opentox-client tests pass

---
 lib/compound.rb | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index ce0fdbf..5992ee3 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -40,25 +40,25 @@ module OpenTox
 		# Get InChI
     # @return [String] InChI string
 		def to_inchi
-      get(:accept => 'chemical/x-inchi').to_s.chomp if @uri
+      RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-inchi'}).chomp
 		end
 
 		# Get (canonical) smiles
     # @return [String] Smiles string
 		def to_smiles
-      get(:accept => 'chemical/x-daylight-smiles').chomp
+      RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-daylight-smiles'}).chomp
 		end
 
     # Get sdf
     # @return [String] SDF string
 		def to_sdf
-      get(:accept => 'chemical/x-mdl-sdfile').chomp
+      RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-mdl-sdfile'}).chomp
 		end
 
     # Get gif image
     # @return [image/gif] Image data
 		def to_gif
-			get("#{CACTUS_URI}#{to_inchi}/image")
+      RestClientWrapper.get("#{CACTUS_URI}#{to_inchi}/image")
 		end
 
     # Get png image
@@ -66,7 +66,7 @@ module OpenTox
     #   image = compound.to_png
     # @return [image/png] Image data
 		def to_png
-      get(File.join @uri, "image")
+      RestClientWrapper.get(File.join @uri, "image")
 		end
 
     # Get URI of compound image
@@ -81,7 +81,7 @@ module OpenTox
     # @return [String] Compound names
 		def to_names
       begin
-        get("#{CACTUS_URI}#{to_inchi}/names").split("\n")
+        RestClientWrapper.get("#{CACTUS_URI}#{to_inchi}/names").split("\n")
       rescue
         "not available"
       end
-- 
cgit v1.2.3


From 32ad3c8f6e1e16cfe9fd59a47df6b560ffb13ddd Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Wed, 1 Aug 2012 14:56:08 +0200
Subject: task error handling improved

---
 lib/compound.rb | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 5992ee3..7d84a3c 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -83,7 +83,7 @@ module OpenTox
       begin
         RestClientWrapper.get("#{CACTUS_URI}#{to_inchi}/names").split("\n")
       rescue
-        "not available"
+        "CACTVS service not responding."
       end
 		end
 
-- 
cgit v1.2.3


From fdad1b78a4ec89c160bff710f0374614c22cda56 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Mon, 20 Aug 2012 14:36:39 +0200
Subject: cached compound representations, initial pubchem and chembl methods

---
 lib/compound.rb | 56 ++++++++++++++++++++++++++++++++++++++------------------
 1 file changed, 38 insertions(+), 18 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 7d84a3c..127a6ee 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -11,6 +11,7 @@ module OpenTox
     # @param [String] smiles Smiles string
     # @return [OpenTox::Compound] Compound
     def self.from_smiles service_uri, smiles, subjectid=nil
+      @smiles = smiles
       Compound.new RestClientWrapper.post(service_uri, smiles, {:content_type => 'chemical/x-daylight-smiles', :subjectid => subjectid})
     end
 
@@ -18,6 +19,7 @@ module OpenTox
     # @param [String] smiles InChI string
     # @return [OpenTox::Compound] Compound
     def self.from_inchi service_uri, inchi, subjectid=nil
+      @inchi = inchi
       Compound.new RestClientWrapper.post(service_uri, inchi, {:content_type => 'chemical/x-inchi', :subjectid => subjectid})
     end
 
@@ -33,60 +35,78 @@ module OpenTox
     #   compound = OpenTox::Compound.from_name("Benzene")
     # @param [String] name name can be also an InChI/InChiKey, CAS number, etc
     # @return [OpenTox::Compound] Compound
+    #
     def self.from_name service_uri, name, subjectid=nil
-      Compound.new RestClientWrapper.post(service_uri, name, {:content_type => 'text/plain', :subjectid => subjectid})
+      @inchi = RestClientWrapper.get File.join(CACTUS_URI,URI.escape(name),"stdinchi")
+      Compound.new RestClientWrapper.post(service_uri, @inchi, {:content_type => 'chemical/x-inchi', :subjectid => subjectid})
     end
 
 		# Get InChI
     # @return [String] InChI string
-		def to_inchi
-      RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-inchi'}).chomp
+		def inchi
+      @inchi ||= RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-inchi'}).chomp
+		end
+
+		# Get InChIKey
+    # @return [String] InChI string
+		def inchikey
+      @inchikey ||= RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-inchikey'}).chomp
 		end
 
 		# Get (canonical) smiles
     # @return [String] Smiles string
-		def to_smiles
-      RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-daylight-smiles'}).chomp
+		def smiles
+      @smiles ||= RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-daylight-smiles'}).chomp
 		end
 
     # Get sdf
     # @return [String] SDF string
-		def to_sdf
+		def sdf
       RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-mdl-sdfile'}).chomp
 		end
 
     # Get gif image
     # @return [image/gif] Image data
-		def to_gif
-      RestClientWrapper.get("#{CACTUS_URI}#{to_inchi}/image")
+		def gif
+      RestClientWrapper.get File.join(CACTUS_URI,inchi,"image")
 		end
 
     # Get png image
     # @example
-    #   image = compound.to_png
+    #   image = compound.png
     # @return [image/png] Image data
-		def to_png
+		def png
       RestClientWrapper.get(File.join @uri, "image")
 		end
 
     # Get URI of compound image
     # @return [String] Compound image URI
-		def to_image_uri
+		def image_uri
       File.join @uri, "image"
 		end
 
     # Get all known compound names. Relies on an external service for name lookups.
     # @example
-    #   names = compound.to_names
+    #   names = compound.names
     # @return [String] Compound names
-		def to_names
-      begin
-        RestClientWrapper.get("#{CACTUS_URI}#{to_inchi}/names").split("\n")
-      rescue
-        "CACTVS service not responding."
-      end
+		def names
+      RestClientWrapper.get("#{CACTUS_URI}#{inchi}/names").split("\n")
 		end
 
+    def cid
+      pug_uri = "http://pubchem.ncbi.nlm.nih.gov/rest/pug/"
+      @cid ||= RestClientWrapper.post(File.join(pug_uri, "compound", "inchi", "cids", "TXT"),{:inchi => inchi}).strip
+    end
+
+    def chebi
+    end
+
+    def chemblid
+      # https://www.ebi.ac.uk/chembldb/ws#individualCompoundByInChiKey
+      uri = "http://www.ebi.ac.uk/chemblws/compounds/smiles/#{smiles}.json"
+      @chemblid = JSON.parse(RestClientWrapper.get(uri))["compounds"].first["chemblId"]
+    end
+
 =begin
 		# Match a smarts string
     # @example
-- 
cgit v1.2.3


From 781e7f714130dcd054d5d437f9c9f95c02cf9f7f Mon Sep 17 00:00:00 2001
From: Andreas Maunz <andreas@maunz.de>
Date: Thu, 6 Sep 2012 16:59:57 +0200
Subject: Added match routines

---
 lib/compound.rb | 39 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 39 insertions(+)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 127a6ee..19d00fb 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -186,5 +186,44 @@ module OpenTox
 			end
 		end
 =end
+
+    # Match an array of smarts strings, returns array with matching smarts
+    # @example
+    #   compound = Compound.from_name("Benzene")
+    #   compound.match(['cc','cN']) # returns ['cc']
+    # @param [Array] smarts_array Array with Smarts strings
+    # @return [Array] Array with matching Smarts strings
+    def match(smarts_array)
+      obconversion = OpenBabel::OBConversion.new
+      obmol = OpenBabel::OBMol.new
+      obconversion.set_in_format('inchi')
+      obconversion.read_string(obmol,inchi)
+      smarts_pattern = OpenBabel::OBSmartsPattern.new
+      smarts_array.collect { |smarts|
+        smarts_pattern.init(smarts)
+        smarts if smarts_pattern.match(obmol)
+      }.compact
+    end
+
+    # Match an array of smarts strings, returns hash with matching smarts as key and number of non-unique hits as value 
+    # @param [Array] smarts_array Array with Smarts strings
+    # @return [Hash] Hash with matching smarts as key and number of non-unique hits as value
+    def match_hits(smarts_array)
+      obconversion = OpenBabel::OBConversion.new
+      obmol = OpenBabel::OBMol.new
+      obconversion.set_in_format('inchi')
+      obconversion.read_string(obmol,inchi)
+      smarts_pattern = OpenBabel::OBSmartsPattern.new
+      smarts_hits = {}
+      smarts_array.collect do |smarts|
+        smarts_pattern.init(smarts)
+        if smarts_pattern.match(obmol)
+          hits = smarts_pattern.get_map_list
+          smarts_hits[smarts] = hits.to_a.size
+        end
+      end
+      return smarts_hits
+    end
+
 	end
 end
-- 
cgit v1.2.3


From fcb0574da3d9e52ad68669b7c84e0a8e029ad240 Mon Sep 17 00:00:00 2001
From: Andreas Maunz <andreas@maunz.de>
Date: Tue, 11 Sep 2012 12:57:00 +0200
Subject: Updated match routines

---
 lib/compound.rb | 43 +++++++++++++++++++++----------------------
 1 file changed, 21 insertions(+), 22 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 19d00fb..da16c85 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -187,28 +187,18 @@ module OpenTox
 		end
 =end
 
-    # Match an array of smarts strings, returns array with matching smarts
-    # @example
-    #   compound = Compound.from_name("Benzene")
-    #   compound.match(['cc','cN']) # returns ['cc']
-    # @param [Array] smarts_array Array with Smarts strings
-    # @return [Array] Array with matching Smarts strings
-    def match(smarts_array)
-      obconversion = OpenBabel::OBConversion.new
-      obmol = OpenBabel::OBMol.new
-      obconversion.set_in_format('inchi')
-      obconversion.read_string(obmol,inchi)
-      smarts_pattern = OpenBabel::OBSmartsPattern.new
-      smarts_array.collect { |smarts|
-        smarts_pattern.init(smarts)
-        smarts if smarts_pattern.match(obmol)
-      }.compact
-    end
 
-    # Match an array of smarts strings, returns hash with matching smarts as key and number of non-unique hits as value 
+
+    # Match an array of smarts strings, returns hash
+    # Keys: matching smarts, values: number of non-unique hits, or 1
     # @param [Array] smarts_array Array with Smarts strings
-    # @return [Hash] Hash with matching smarts as key and number of non-unique hits as value
-    def match_hits(smarts_array)
+    # @param [Boolean] Whether non-unique hits or 1 should be produced
+    # @return [Array] Array with matching Smarts strings
+    # @example {
+    #   compound = Compound.from_name("Benzene")
+    #   compound.match(['cc','cN']) # returns { 'cc' => 12, 'cN' => 0 }
+    # }
+    def match_hits(smarts_array, use_hits=true)
       obconversion = OpenBabel::OBConversion.new
       obmol = OpenBabel::OBMol.new
       obconversion.set_in_format('inchi')
@@ -218,12 +208,21 @@ module OpenTox
       smarts_array.collect do |smarts|
         smarts_pattern.init(smarts)
         if smarts_pattern.match(obmol)
-          hits = smarts_pattern.get_map_list
-          smarts_hits[smarts] = hits.to_a.size
+          if use_hits 
+            hits = smarts_pattern.get_map_list
+            smarts_hits[smarts] = hits.to_a.size
+          else
+            smarts_hits[smarts] = 1
+          end
         end
       end
       return smarts_hits
     end
 
+    # Provided for backward compatibility
+    def match(smarts_array)
+      match_hits(smarts_array,false)
+    end
+
 	end
 end
-- 
cgit v1.2.3


From 147e1b7351975f19697e1e6ff4a6106fd38174ae Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Wed, 23 Jan 2013 12:53:43 +0100
Subject: url_for removed auto insertion of RDF::DC.modified

---
 lib/compound.rb | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index da16c85..a0d6ec7 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -11,7 +11,7 @@ module OpenTox
     # @param [String] smiles Smiles string
     # @return [OpenTox::Compound] Compound
     def self.from_smiles service_uri, smiles, subjectid=nil
-      @smiles = smiles
+      #@smiles = smiles
       Compound.new RestClientWrapper.post(service_uri, smiles, {:content_type => 'chemical/x-daylight-smiles', :subjectid => subjectid})
     end
 
@@ -19,7 +19,7 @@ module OpenTox
     # @param [String] smiles InChI string
     # @return [OpenTox::Compound] Compound
     def self.from_inchi service_uri, inchi, subjectid=nil
-      @inchi = inchi
+      #@inchi = inchi
       Compound.new RestClientWrapper.post(service_uri, inchi, {:content_type => 'chemical/x-inchi', :subjectid => subjectid})
     end
 
-- 
cgit v1.2.3


From a54db46684680d98311631804eca367cc949a715 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Tue, 26 Mar 2013 10:56:04 +0100
Subject: code cleanup and refactoring.

---
 lib/compound.rb | 12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index a0d6ec7..5d77dec 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -10,23 +10,21 @@ module OpenTox
     #   compound = OpenTox::Compound.from_smiles("c1ccccc1")
     # @param [String] smiles Smiles string
     # @return [OpenTox::Compound] Compound
-    def self.from_smiles service_uri, smiles, subjectid=nil
-      #@smiles = smiles
+    def self.from_smiles smiles, subjectid=nil
       Compound.new RestClientWrapper.post(service_uri, smiles, {:content_type => 'chemical/x-daylight-smiles', :subjectid => subjectid})
     end
 
     # Create a compound from inchi string
     # @param [String] smiles InChI string
     # @return [OpenTox::Compound] Compound
-    def self.from_inchi service_uri, inchi, subjectid=nil
-      #@inchi = inchi
+    def self.from_inchi inchi, subjectid=nil
       Compound.new RestClientWrapper.post(service_uri, inchi, {:content_type => 'chemical/x-inchi', :subjectid => subjectid})
     end
 
     # Create a compound from sdf string
     # @param [String] smiles SDF string
     # @return [OpenTox::Compound] Compound
-    def self.from_sdf service_uri, sdf, subjectid=nil
+    def self.from_sdf sdf, subjectid=nil
       Compound.new RestClientWrapper.post(service_uri, sdf, {:content_type => 'chemical/x-mdl-sdfile', :subjectid => subjectid})
     end
 
@@ -36,7 +34,7 @@ module OpenTox
     # @param [String] name name can be also an InChI/InChiKey, CAS number, etc
     # @return [OpenTox::Compound] Compound
     #
-    def self.from_name service_uri, name, subjectid=nil
+    def self.from_name name, subjectid=nil
       @inchi = RestClientWrapper.get File.join(CACTUS_URI,URI.escape(name),"stdinchi")
       Compound.new RestClientWrapper.post(service_uri, @inchi, {:content_type => 'chemical/x-inchi', :subjectid => subjectid})
     end
@@ -216,7 +214,7 @@ module OpenTox
           end
         end
       end
-      return smarts_hits
+      smarts_hits
     end
 
     # Provided for backward compatibility
-- 
cgit v1.2.3


From 7def60bb82474c5e12bbc0362476752524e25d53 Mon Sep 17 00:00:00 2001
From: gebele <gebele@in-silico.ch>
Date: Wed, 15 May 2013 13:43:20 +0200
Subject: require openbabel

---
 lib/compound.rb | 1 +
 1 file changed, 1 insertion(+)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 5d77dec..294aa90 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -1,3 +1,4 @@
+require "openbabel"
 CACTUS_URI="http://cactus.nci.nih.gov/chemical/structure/"
 
 module OpenTox
-- 
cgit v1.2.3


From 0df97a547ad623cff5d714076addbaa3cc73c53b Mon Sep 17 00:00:00 2001
From: davor <vorgrimmlerdavid@gmx.de>
Date: Tue, 28 May 2013 12:15:36 +0200
Subject: Replacing tab with two spaces.

---
 lib/compound.rb | 126 ++++++++++++++++++++++++++++----------------------------
 1 file changed, 63 insertions(+), 63 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 294aa90..2eccdcd 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -4,7 +4,7 @@ CACTUS_URI="http://cactus.nci.nih.gov/chemical/structure/"
 module OpenTox
 
   # Ruby wrapper for OpenTox Compound Webservices (http://opentox.org/dev/apis/api-1.2/structure).
-	class Compound 
+  class Compound
 
     # Create a compound from smiles string
     # @example
@@ -40,57 +40,57 @@ module OpenTox
       Compound.new RestClientWrapper.post(service_uri, @inchi, {:content_type => 'chemical/x-inchi', :subjectid => subjectid})
     end
 
-		# Get InChI
+    # Get InChI
     # @return [String] InChI string
-		def inchi
+    def inchi
       @inchi ||= RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-inchi'}).chomp
-		end
+    end
 
-		# Get InChIKey
+    # Get InChIKey
     # @return [String] InChI string
-		def inchikey
+    def inchikey
       @inchikey ||= RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-inchikey'}).chomp
-		end
+    end
 
-		# Get (canonical) smiles
+    # Get (canonical) smiles
     # @return [String] Smiles string
-		def smiles
+    def smiles
       @smiles ||= RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-daylight-smiles'}).chomp
-		end
+    end
 
     # Get sdf
     # @return [String] SDF string
-		def sdf
+    def sdf
       RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-mdl-sdfile'}).chomp
-		end
+    end
 
     # Get gif image
     # @return [image/gif] Image data
-		def gif
+    def gif
       RestClientWrapper.get File.join(CACTUS_URI,inchi,"image")
-		end
+    end
 
     # Get png image
     # @example
     #   image = compound.png
     # @return [image/png] Image data
-		def png
+    def png
       RestClientWrapper.get(File.join @uri, "image")
-		end
+    end
 
     # Get URI of compound image
     # @return [String] Compound image URI
-		def image_uri
+    def image_uri
       File.join @uri, "image"
-		end
+    end
 
     # Get all known compound names. Relies on an external service for name lookups.
     # @example
     #   names = compound.names
     # @return [String] Compound names
-		def names
+    def names
       RestClientWrapper.get("#{CACTUS_URI}#{inchi}/names").split("\n")
-		end
+    end
 
     def cid
       pug_uri = "http://pubchem.ncbi.nlm.nih.gov/rest/pug/"
@@ -107,40 +107,40 @@ module OpenTox
     end
 
 =begin
-		# Match a smarts string
+    # Match a smarts string
     # @example
     #   compound = OpenTox::Compound.from_name("Benzene")
     #   compound.match?("cN") # returns false
     # @param [String] smarts Smarts string
-		def match?(smarts)
-			obconversion = OpenBabel::OBConversion.new
-			obmol = OpenBabel::OBMol.new
-			obconversion.set_in_format('inchi')
-			obconversion.read_string(obmol,@inchi) 
-			smarts_pattern = OpenBabel::OBSmartsPattern.new
-			smarts_pattern.init(smarts)
-			smarts_pattern.match(obmol)
-		end
-
-		# Match an array of smarts strings, returns array with matching smarts
+    def match?(smarts)
+      obconversion = OpenBabel::OBConversion.new
+      obmol = OpenBabel::OBMol.new
+      obconversion.set_in_format('inchi')
+      obconversion.read_string(obmol,@inchi)
+      smarts_pattern = OpenBabel::OBSmartsPattern.new
+      smarts_pattern.init(smarts)
+      smarts_pattern.match(obmol)
+    end
+
+    # Match an array of smarts strings, returns array with matching smarts
     # @example
     #   compound = OpenTox::Compound.from_name("Benzene")
     #   compound.match(['cc','cN']) # returns ['cc']
     # @param [Array] smarts_array Array with Smarts strings
     # @return [Array] Array with matching Smarts strings
-		def match(smarts_array)
+    def match(smarts_array)
       # avoid recreation of OpenBabel objects
-			obconversion = OpenBabel::OBConversion.new
-			obmol = OpenBabel::OBMol.new
-			obconversion.set_in_format('inchi')
-			obconversion.read_string(obmol,@inchi) 
-			smarts_pattern = OpenBabel::OBSmartsPattern.new
-			smarts_array.collect do |smarts|
+      obconversion = OpenBabel::OBConversion.new
+      obmol = OpenBabel::OBMol.new
+      obconversion.set_in_format('inchi')
+      obconversion.read_string(obmol,@inchi)
+      smarts_pattern = OpenBabel::OBSmartsPattern.new
+      smarts_array.collect do |smarts|
         smarts_pattern.init(smarts)
         smarts if smarts_pattern.match(obmol)
       end.compact
       #smarts_array.collect { |s| s if match?(s)}.compact
-		end
+    end
 
     # Get URI of compound image with highlighted fragments
     #
@@ -157,33 +157,33 @@ module OpenTox
     private
 
     # Convert sdf to inchi
-		def self.sdf2inchi(sdf)
-			Compound.obconversion(sdf,'sdf','inchi')
-		end
+    def self.sdf2inchi(sdf)
+      Compound.obconversion(sdf,'sdf','inchi')
+    end
 
     # Convert smiles to inchi
-		def self.smiles2inchi(smiles)
-			Compound.obconversion(smiles,'smi','inchi')
-		end
+    def self.smiles2inchi(smiles)
+      Compound.obconversion(smiles,'smi','inchi')
+    end
 
     # Convert smiles to canonical smiles
-		def self.smiles2cansmi(smiles)
-			Compound.obconversion(smiles,'smi','can')
-		end
+    def self.smiles2cansmi(smiles)
+      Compound.obconversion(smiles,'smi','can')
+    end
 
     # Convert identifier from OpenBabel input_format to OpenBabel output_format
-		def self.obconversion(identifier,input_format,output_format)
-			obconversion = OpenBabel::OBConversion.new
-			obmol = OpenBabel::OBMol.new
-			obconversion.set_in_and_out_formats input_format, output_format
-			obconversion.read_string obmol, identifier
-			case output_format
-			when /smi|can|inchi/
-				obconversion.write_string(obmol).gsub(/\s/,'').chomp
-			else
-				obconversion.write_string(obmol)
-			end
-		end
+    def self.obconversion(identifier,input_format,output_format)
+      obconversion = OpenBabel::OBConversion.new
+      obmol = OpenBabel::OBMol.new
+      obconversion.set_in_and_out_formats input_format, output_format
+      obconversion.read_string obmol, identifier
+      case output_format
+      when /smi|can|inchi/
+        obconversion.write_string(obmol).gsub(/\s/,'').chomp
+      else
+        obconversion.write_string(obmol)
+      end
+    end
 =end
 
 
@@ -207,7 +207,7 @@ module OpenTox
       smarts_array.collect do |smarts|
         smarts_pattern.init(smarts)
         if smarts_pattern.match(obmol)
-          if use_hits 
+          if use_hits
             hits = smarts_pattern.get_map_list
             smarts_hits[smarts] = hits.to_a.size
           else
@@ -223,5 +223,5 @@ module OpenTox
       match_hits(smarts_array,false)
     end
 
-	end
+  end
 end
-- 
cgit v1.2.3


From 0f770837cbfe17fe17fc86aa0d360c55e281c7f7 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Wed, 29 May 2013 15:10:46 +0200
Subject: modification date inserted at client to improve performance

---
 lib/compound.rb | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 5d77dec..2bef69e 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -183,7 +183,6 @@ module OpenTox
 				obconversion.write_string(obmol)
 			end
 		end
-=end
 
 
 
@@ -221,6 +220,7 @@ module OpenTox
     def match(smarts_array)
       match_hits(smarts_array,false)
     end
+=end
 
 	end
 end
-- 
cgit v1.2.3


From 962773d59d94f8d16dae11b6290fdf30aaf09d7b Mon Sep 17 00:00:00 2001
From: rautenberg <rautenberg@in-silico.ch>
Date: Mon, 3 Jun 2013 15:25:33 +0200
Subject: make yard documentation full yard 0.8.6.1 compatible

---
 lib/compound.rb | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 2eccdcd..a278244 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -16,14 +16,14 @@ module OpenTox
     end
 
     # Create a compound from inchi string
-    # @param [String] smiles InChI string
+    # @param inchi [String] smiles InChI string
     # @return [OpenTox::Compound] Compound
     def self.from_inchi inchi, subjectid=nil
       Compound.new RestClientWrapper.post(service_uri, inchi, {:content_type => 'chemical/x-inchi', :subjectid => subjectid})
     end
 
     # Create a compound from sdf string
-    # @param [String] smiles SDF string
+    # @param sdf [String] smiles SDF string
     # @return [OpenTox::Compound] Compound
     def self.from_sdf sdf, subjectid=nil
       Compound.new RestClientWrapper.post(service_uri, sdf, {:content_type => 'chemical/x-mdl-sdfile', :subjectid => subjectid})
@@ -32,9 +32,8 @@ module OpenTox
     # Create a compound from name. Relies on an external service for name lookups.
     # @example
     #   compound = OpenTox::Compound.from_name("Benzene")
-    # @param [String] name name can be also an InChI/InChiKey, CAS number, etc
+    # @param name [String] can be also an InChI/InChiKey, CAS number, etc
     # @return [OpenTox::Compound] Compound
-    #
     def self.from_name name, subjectid=nil
       @inchi = RestClientWrapper.get File.join(CACTUS_URI,URI.escape(name),"stdinchi")
       Compound.new RestClientWrapper.post(service_uri, @inchi, {:content_type => 'chemical/x-inchi', :subjectid => subjectid})
@@ -191,7 +190,7 @@ module OpenTox
     # Match an array of smarts strings, returns hash
     # Keys: matching smarts, values: number of non-unique hits, or 1
     # @param [Array] smarts_array Array with Smarts strings
-    # @param [Boolean] Whether non-unique hits or 1 should be produced
+    # @param use_hits [Boolean] Whether non-unique hits or 1 should be produced
     # @return [Array] Array with matching Smarts strings
     # @example {
     #   compound = Compound.from_name("Benzene")
-- 
cgit v1.2.3


From 498ec077e6b8290cf5efee08aaa6af7759ddccc1 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Martin=20G=C3=BCtlein?= <martin.guetlein@gmail.com>
Date: Thu, 6 Jun 2013 12:58:26 +0200
Subject: commenting cid, chemblid, and match_hits

---
 lib/compound.rb | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index a278244..51ef96d 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -91,14 +91,18 @@ module OpenTox
       RestClientWrapper.get("#{CACTUS_URI}#{inchi}/names").split("\n")
     end
 
+    # @return [String] PubChem Compound Identifier (CID), derieved via restcall to pubchem
     def cid
       pug_uri = "http://pubchem.ncbi.nlm.nih.gov/rest/pug/"
       @cid ||= RestClientWrapper.post(File.join(pug_uri, "compound", "inchi", "cids", "TXT"),{:inchi => inchi}).strip
     end
 
+    # @todo
     def chebi
+      raise_internal_error "not yet implemented"
     end
 
+    # @return [String] ChEMBL database compound id, derieved via restcall to chembl
     def chemblid
       # https://www.ebi.ac.uk/chembldb/ws#individualCompoundByInChiKey
       uri = "http://www.ebi.ac.uk/chemblws/compounds/smiles/#{smiles}.json"
@@ -191,11 +195,11 @@ module OpenTox
     # Keys: matching smarts, values: number of non-unique hits, or 1
     # @param [Array] smarts_array Array with Smarts strings
     # @param use_hits [Boolean] Whether non-unique hits or 1 should be produced
-    # @return [Array] Array with matching Smarts strings
-    # @example {
+    # @return [Hash] Hash with matching Smarts as keys, nr-of-hits/1 as values
+    # @example
     #   compound = Compound.from_name("Benzene")
-    #   compound.match(['cc','cN']) # returns { 'cc' => 12, 'cN' => 0 }
-    # }
+    #   compound.match(['cc','cN'],true) # returns { 'cc' => 12 }, 'cN' is not included because it does not match
+    #   compound.match(['cc','cN'],false) # returns { 'cc' => 1 }
     def match_hits(smarts_array, use_hits=true)
       obconversion = OpenBabel::OBConversion.new
       obmol = OpenBabel::OBMol.new
-- 
cgit v1.2.3


From bc9d5b9266e18d39ab5fd83db272bffde98c4161 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Wed, 19 Jun 2013 16:29:43 +0200
Subject: openbabel requirement removed. initial descriptor classes.

---
 lib/compound.rb | 23 ++++++++++++++++++++++-
 1 file changed, 22 insertions(+), 1 deletion(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 89bf840..daeabb9 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -1,4 +1,4 @@
-require "openbabel"
+#require "openbabel"
 CACTUS_URI="http://cactus.nci.nih.gov/chemical/structure/"
 
 module OpenTox
@@ -110,6 +110,27 @@ module OpenTox
     end
 
 =begin
+    # Match a smarts string
+    # @example
+    #   compound = OpenTox::Compound.from_name("Benzene")
+    #   compound.match?("cN") # returns false
+    # @param [String] smarts Smarts string
+    def match?(smarts)
+      matcher = Algorithm.new File.join($algorithm[:uri],"descriptor","smarts")
+      matcher.run :compound_uri => @uri, :smarts => smarts, :count => false
+    end
+
+    # Match an array of smarts strings, returns array with matching smarts
+    # @example
+    #   compound = OpenTox::Compound.from_name("Benzene")
+    #   compound.match(['cc','cN']) # returns ['cc']
+    # @param [Array] smarts_array Array with Smarts strings
+    # @return [Array] Array with matching Smarts strings
+    def match(smarts_array)
+      matcher = Algorithm.new File.join($algorithm[:uri],"descriptor","smarts")
+      matcher.run :compound_uri => @uri, :smarts => smarts_array, :count => false
+    end
+
     # Match a smarts string
     # @example
     #   compound = OpenTox::Compound.from_name("Benzene")
-- 
cgit v1.2.3


From 6c35e3f8369ba96cb251eac487424bd949fdcf6c Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Tue, 2 Jul 2013 19:23:01 +0200
Subject: Algorithms and Models are modules instead of classes.

---
 lib/compound.rb | 4 ++++
 1 file changed, 4 insertions(+)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index daeabb9..9bd2066 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -6,6 +6,10 @@ module OpenTox
   # Ruby wrapper for OpenTox Compound Webservices (http://opentox.org/dev/apis/api-1.2/structure).
   class Compound
 
+    def ==(c)
+      @uri == c.uri
+    end
+
     # Create a compound from smiles string
     # @example
     #   compound = OpenTox::Compound.from_smiles("c1ccccc1")
-- 
cgit v1.2.3


From 3262e3cf2ed1ce896dcd11d9c1bde5222ea720ce Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Thu, 4 Jul 2013 18:17:56 +0200
Subject: Algorithm::Generic and Model::Generic classes.

---
 lib/compound.rb | 139 --------------------------------------------------------
 1 file changed, 139 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 9bd2066..82f0d73 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -112,144 +112,5 @@ module OpenTox
       uri = "http://www.ebi.ac.uk/chemblws/compounds/smiles/#{smiles}.json"
       @chemblid = JSON.parse(RestClientWrapper.get(uri))["compounds"].first["chemblId"]
     end
-
-=begin
-    # Match a smarts string
-    # @example
-    #   compound = OpenTox::Compound.from_name("Benzene")
-    #   compound.match?("cN") # returns false
-    # @param [String] smarts Smarts string
-    def match?(smarts)
-      matcher = Algorithm.new File.join($algorithm[:uri],"descriptor","smarts")
-      matcher.run :compound_uri => @uri, :smarts => smarts, :count => false
-    end
-
-    # Match an array of smarts strings, returns array with matching smarts
-    # @example
-    #   compound = OpenTox::Compound.from_name("Benzene")
-    #   compound.match(['cc','cN']) # returns ['cc']
-    # @param [Array] smarts_array Array with Smarts strings
-    # @return [Array] Array with matching Smarts strings
-    def match(smarts_array)
-      matcher = Algorithm.new File.join($algorithm[:uri],"descriptor","smarts")
-      matcher.run :compound_uri => @uri, :smarts => smarts_array, :count => false
-    end
-
-    # Match a smarts string
-    # @example
-    #   compound = OpenTox::Compound.from_name("Benzene")
-    #   compound.match?("cN") # returns false
-    # @param [String] smarts Smarts string
-    def match?(smarts)
-      obconversion = OpenBabel::OBConversion.new
-      obmol = OpenBabel::OBMol.new
-      obconversion.set_in_format('inchi')
-      obconversion.read_string(obmol,@inchi)
-      smarts_pattern = OpenBabel::OBSmartsPattern.new
-      smarts_pattern.init(smarts)
-      smarts_pattern.match(obmol)
-    end
-
-    # Match an array of smarts strings, returns array with matching smarts
-    # @example
-    #   compound = OpenTox::Compound.from_name("Benzene")
-    #   compound.match(['cc','cN']) # returns ['cc']
-    # @param [Array] smarts_array Array with Smarts strings
-    # @return [Array] Array with matching Smarts strings
-    def match(smarts_array)
-      # avoid recreation of OpenBabel objects
-      obconversion = OpenBabel::OBConversion.new
-      obmol = OpenBabel::OBMol.new
-      obconversion.set_in_format('inchi')
-      obconversion.read_string(obmol,@inchi)
-      smarts_pattern = OpenBabel::OBSmartsPattern.new
-      smarts_array.collect do |smarts|
-        smarts_pattern.init(smarts)
-        smarts if smarts_pattern.match(obmol)
-      end.compact
-      #smarts_array.collect { |s| s if match?(s)}.compact
-    end
-
-    # Get URI of compound image with highlighted fragments
-    #
-    # @param [Array] activating Array with activating Smarts strings
-    # @param [Array] deactivating Array with deactivating Smarts strings
-    # @return [String] URI for compound image with highlighted fragments
-    def matching_smarts_image_uri(activating, deactivating)
-      activating_smarts = URI.encode "\"#{activating.join("\"/\"")}\""
-      deactivating_smarts = URI.encode "\"#{deactivating.join("\"/\"")}\""
-      File.join @uri, "smarts/activating", URI.encode(activating_smarts),"deactivating", URI.encode(deactivating_smarts)
-    end
-
-
-    private
-
-    # Convert sdf to inchi
-    def self.sdf2inchi(sdf)
-      Compound.obconversion(sdf,'sdf','inchi')
-    end
-
-    # Convert smiles to inchi
-    def self.smiles2inchi(smiles)
-      Compound.obconversion(smiles,'smi','inchi')
-    end
-
-    # Convert smiles to canonical smiles
-    def self.smiles2cansmi(smiles)
-      Compound.obconversion(smiles,'smi','can')
-    end
-
-    # Convert identifier from OpenBabel input_format to OpenBabel output_format
-		def self.obconversion(identifier,input_format,output_format)
-			obconversion = OpenBabel::OBConversion.new
-			obmol = OpenBabel::OBMol.new
-			obconversion.set_in_and_out_formats input_format, output_format
-			obconversion.read_string obmol, identifier
-			case output_format
-			when /smi|can|inchi/
-				obconversion.write_string(obmol).gsub(/\s/,'').chomp
-			else
-				obconversion.write_string(obmol)
-			end
-		end
-
-
-
-    # Match an array of smarts strings, returns hash
-    # Keys: matching smarts, values: number of non-unique hits, or 1
-    # @param [Array] smarts_array Array with Smarts strings
-    # @param use_hits [Boolean] Whether non-unique hits or 1 should be produced
-    # @return [Hash] Hash with matching Smarts as keys, nr-of-hits/1 as values
-    # @example
-    #   compound = Compound.from_name("Benzene")
-    #   compound.match(['cc','cN'],true) # returns { 'cc' => 12 }, 'cN' is not included because it does not match
-    #   compound.match(['cc','cN'],false) # returns { 'cc' => 1 }
-    def match_hits(smarts_array, use_hits=true)
-      obconversion = OpenBabel::OBConversion.new
-      obmol = OpenBabel::OBMol.new
-      obconversion.set_in_format('inchi')
-      obconversion.read_string(obmol,inchi)
-      smarts_pattern = OpenBabel::OBSmartsPattern.new
-      smarts_hits = {}
-      smarts_array.collect do |smarts|
-        smarts_pattern.init(smarts)
-        if smarts_pattern.match(obmol)
-          if use_hits
-            hits = smarts_pattern.get_map_list
-            smarts_hits[smarts] = hits.to_a.size
-          else
-            smarts_hits[smarts] = 1
-          end
-        end
-      end
-      smarts_hits
-    end
-
-    # Provided for backward compatibility
-    def match(smarts_array)
-      match_hits(smarts_array,false)
-    end
-=end
-
   end
 end
-- 
cgit v1.2.3


From ddb83b4302e8628b333402d24e3e05fc90b3faef Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Wed, 10 Jul 2013 14:36:39 +0200
Subject: subjectids partially removed

---
 lib/compound.rb | 17 ++++++++---------
 1 file changed, 8 insertions(+), 9 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 82f0d73..9d99ae9 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -1,4 +1,3 @@
-#require "openbabel"
 CACTUS_URI="http://cactus.nci.nih.gov/chemical/structure/"
 
 module OpenTox
@@ -15,22 +14,22 @@ module OpenTox
     #   compound = OpenTox::Compound.from_smiles("c1ccccc1")
     # @param [String] smiles Smiles string
     # @return [OpenTox::Compound] Compound
-    def self.from_smiles smiles, subjectid=nil
-      Compound.new RestClientWrapper.post(service_uri, smiles, {:content_type => 'chemical/x-daylight-smiles', :subjectid => subjectid})
+    def self.from_smiles smiles
+      Compound.new RestClientWrapper.post(service_uri, smiles, {:content_type => 'chemical/x-daylight-smiles'})
     end
 
     # Create a compound from inchi string
     # @param inchi [String] smiles InChI string
     # @return [OpenTox::Compound] Compound
-    def self.from_inchi inchi, subjectid=nil
-      Compound.new RestClientWrapper.post(service_uri, inchi, {:content_type => 'chemical/x-inchi', :subjectid => subjectid})
+    def self.from_inchi inchi
+      Compound.new RestClientWrapper.post(service_uri, inchi, {:content_type => 'chemical/x-inchi'})
     end
 
     # Create a compound from sdf string
     # @param sdf [String] smiles SDF string
     # @return [OpenTox::Compound] Compound
-    def self.from_sdf sdf, subjectid=nil
-      Compound.new RestClientWrapper.post(service_uri, sdf, {:content_type => 'chemical/x-mdl-sdfile', :subjectid => subjectid})
+    def self.from_sdf sdf
+      Compound.new RestClientWrapper.post(service_uri, sdf, {:content_type => 'chemical/x-mdl-sdfile'})
     end
 
     # Create a compound from name. Relies on an external service for name lookups.
@@ -38,9 +37,9 @@ module OpenTox
     #   compound = OpenTox::Compound.from_name("Benzene")
     # @param name [String] can be also an InChI/InChiKey, CAS number, etc
     # @return [OpenTox::Compound] Compound
-    def self.from_name name, subjectid=nil
+    def self.from_name name
       @inchi = RestClientWrapper.get File.join(CACTUS_URI,URI.escape(name),"stdinchi")
-      Compound.new RestClientWrapper.post(service_uri, @inchi, {:content_type => 'chemical/x-inchi', :subjectid => subjectid})
+      Compound.new RestClientWrapper.post(service_uri, @inchi, {:content_type => 'chemical/x-inchi'})
     end
 
     # Get InChI
-- 
cgit v1.2.3


From 0172f39bf3638919735d0bcd2b6597c5885a05d7 Mon Sep 17 00:00:00 2001
From: mguetlein <martin.guetlein@gmail.com>
Date: Fri, 10 Oct 2014 12:35:50 +0200
Subject: fix error msg

---
 lib/compound.rb | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 9d99ae9..f63e127 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -102,7 +102,7 @@ module OpenTox
 
     # @todo
     def chebi
-      raise_internal_error "not yet implemented"
+      internal_server_error "not yet implemented"
     end
 
     # @return [String] ChEMBL database compound id, derieved via restcall to chembl
-- 
cgit v1.2.3


From f850712765a67bf31b1327401e3eddf59e3e6f50 Mon Sep 17 00:00:00 2001
From: ch <ch@ch.in-silico.ch>
Date: Sun, 12 Jul 2015 12:14:13 +0200
Subject: initial dataset tests

---
 lib/compound.rb | 19 ++++++++++++-------
 1 file changed, 12 insertions(+), 7 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index f63e127..82ea94e 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -5,8 +5,13 @@ module OpenTox
   # Ruby wrapper for OpenTox Compound Webservices (http://opentox.org/dev/apis/api-1.2/structure).
   class Compound
 
+    def initialize uri
+      @data = {}
+      @data["uri"] = uri
+    end
+
     def ==(c)
-      @uri == c.uri
+      @data["uri"] == c.uri
     end
 
     # Create a compound from smiles string
@@ -45,25 +50,25 @@ module OpenTox
     # Get InChI
     # @return [String] InChI string
     def inchi
-      @inchi ||= RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-inchi'}).chomp
+      @inchi ||= RestClientWrapper.get(@data["uri"],{},{:accept => 'chemical/x-inchi'}).chomp
     end
 
     # Get InChIKey
     # @return [String] InChI string
     def inchikey
-      @inchikey ||= RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-inchikey'}).chomp
+      @inchikey ||= RestClientWrapper.get(@data["uri"],{},{:accept => 'chemical/x-inchikey'}).chomp
     end
 
     # Get (canonical) smiles
     # @return [String] Smiles string
     def smiles
-      @smiles ||= RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-daylight-smiles'}).chomp
+      @smiles ||= RestClientWrapper.get(@data["uri"],{},{:accept => 'chemical/x-daylight-smiles'}).chomp
     end
 
     # Get sdf
     # @return [String] SDF string
     def sdf
-      RestClientWrapper.get(@uri,{},{:accept => 'chemical/x-mdl-sdfile'}).chomp
+      RestClientWrapper.get(@data["uri"],{},{:accept => 'chemical/x-mdl-sdfile'}).chomp
     end
 
     # Get gif image
@@ -77,13 +82,13 @@ module OpenTox
     #   image = compound.png
     # @return [image/png] Image data
     def png
-      RestClientWrapper.get(File.join @uri, "image")
+      RestClientWrapper.get(File.join @data["uri"], "image")
     end
 
     # Get URI of compound image
     # @return [String] Compound image URI
     def image_uri
-      File.join @uri, "image"
+      File.join @data["uri"], "image"
     end
 
     # Get all known compound names. Relies on an external service for name lookups.
-- 
cgit v1.2.3


From 1c74a684b8b2992d7e6fce5f1a4677d397ec4dd6 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Sat, 18 Jul 2015 08:52:39 +0200
Subject: dataset tests pass

---
 lib/compound.rb | 55 ++++++++++++++++++++++++++++++++++---------------------
 1 file changed, 34 insertions(+), 21 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 82ea94e..b588c75 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -1,17 +1,15 @@
 CACTUS_URI="http://cactus.nci.nih.gov/chemical/structure/"
+require 'openbabel'
 
 module OpenTox
 
   # Ruby wrapper for OpenTox Compound Webservices (http://opentox.org/dev/apis/api-1.2/structure).
   class Compound
 
-    def initialize uri
-      @data = {}
-      @data["uri"] = uri
-    end
+    attr_reader :inchi
 
-    def ==(c)
-      @data["uri"] == c.uri
+    def initialize inchi
+      @inchi = inchi
     end
 
     # Create a compound from smiles string
@@ -20,21 +18,21 @@ module OpenTox
     # @param [String] smiles Smiles string
     # @return [OpenTox::Compound] Compound
     def self.from_smiles smiles
-      Compound.new RestClientWrapper.post(service_uri, smiles, {:content_type => 'chemical/x-daylight-smiles'})
+      OpenTox::Compound.new obconversion(smiles,"smi","inchi")
     end
 
     # Create a compound from inchi string
     # @param inchi [String] smiles InChI string
     # @return [OpenTox::Compound] Compound
     def self.from_inchi inchi
-      Compound.new RestClientWrapper.post(service_uri, inchi, {:content_type => 'chemical/x-inchi'})
+      OpenTox::Compound.new inchi
     end
 
     # Create a compound from sdf string
     # @param sdf [String] smiles SDF string
     # @return [OpenTox::Compound] Compound
     def self.from_sdf sdf
-      Compound.new RestClientWrapper.post(service_uri, sdf, {:content_type => 'chemical/x-mdl-sdfile'})
+      OpenTox::Compound.new obconversion(sdf,"sdf","inchi")
     end
 
     # Create a compound from name. Relies on an external service for name lookups.
@@ -43,32 +41,25 @@ module OpenTox
     # @param name [String] can be also an InChI/InChiKey, CAS number, etc
     # @return [OpenTox::Compound] Compound
     def self.from_name name
-      @inchi = RestClientWrapper.get File.join(CACTUS_URI,URI.escape(name),"stdinchi")
-      Compound.new RestClientWrapper.post(service_uri, @inchi, {:content_type => 'chemical/x-inchi'})
-    end
-
-    # Get InChI
-    # @return [String] InChI string
-    def inchi
-      @inchi ||= RestClientWrapper.get(@data["uri"],{},{:accept => 'chemical/x-inchi'}).chomp
+      OpenTox::Compound.new RestClientWrapper.get File.join(CACTUS_URI,URI.escape(name),"stdinchi")
     end
 
     # Get InChIKey
     # @return [String] InChI string
     def inchikey
-      @inchikey ||= RestClientWrapper.get(@data["uri"],{},{:accept => 'chemical/x-inchikey'}).chomp
+      obconversion(@inchi,"inchi","inchikey")
     end
 
     # Get (canonical) smiles
     # @return [String] Smiles string
     def smiles
-      @smiles ||= RestClientWrapper.get(@data["uri"],{},{:accept => 'chemical/x-daylight-smiles'}).chomp
+      obconversion(@inchi,"inchi","smi") # "can" gives nonn-canonical smiles??
     end
 
     # Get sdf
     # @return [String] SDF string
     def sdf
-      RestClientWrapper.get(@data["uri"],{},{:accept => 'chemical/x-mdl-sdfile'}).chomp
+      obconversion(@inchi,"inchi","sdf")
     end
 
     # Get gif image
@@ -82,14 +73,16 @@ module OpenTox
     #   image = compound.png
     # @return [image/png] Image data
     def png
-      RestClientWrapper.get(File.join @data["uri"], "image")
+      obconversion(@inchi,"inchi","_png2")
     end
 
+=begin
     # Get URI of compound image
     # @return [String] Compound image URI
     def image_uri
       File.join @data["uri"], "image"
     end
+=end
 
     # Get all known compound names. Relies on an external service for name lookups.
     # @example
@@ -116,5 +109,25 @@ module OpenTox
       uri = "http://www.ebi.ac.uk/chemblws/compounds/smiles/#{smiles}.json"
       @chemblid = JSON.parse(RestClientWrapper.get(uri))["compounds"].first["chemblId"]
     end
+
+    private
+
+    def self.obconversion(identifier,input_format,output_format,option=nil)
+      obconversion = OpenBabel::OBConversion.new
+      obconversion.set_options(option, OpenBabel::OBConversion::OUTOPTIONS) if option
+      obmol = OpenBabel::OBMol.new
+      obconversion.set_in_and_out_formats input_format, output_format
+      obconversion.read_string obmol, identifier
+      case output_format
+      when /smi|can|inchi/
+        obconversion.write_string(obmol).gsub(/\s/,'').chomp
+      else
+        obconversion.write_string(obmol)
+      end
+    end
+
+    def obconversion(identifier,input_format,output_format,option=nil)
+      self.class.obconversion(identifier,input_format,output_format,option=nil)
+    end
   end
 end
-- 
cgit v1.2.3


From 51e7a29c8aac568ff8ef04de5b15d2f6db8f66da Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Sun, 19 Jul 2015 21:26:03 +0200
Subject: lazar predictions working in principle

---
 lib/compound.rb | 4 ++++
 1 file changed, 4 insertions(+)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index b588c75..6cc4707 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -12,6 +12,10 @@ module OpenTox
       @inchi = inchi
     end
 
+    def  == compound
+      self.inchi == compound.inchi
+    end
+
     # Create a compound from smiles string
     # @example
     #   compound = OpenTox::Compound.from_smiles("c1ccccc1")
-- 
cgit v1.2.3


From cd1cba67830505cd2d23ec83e64c0beed42a9f28 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Wed, 22 Jul 2015 20:08:12 +0200
Subject: mongo batch import workaround

---
 lib/compound.rb | 86 ++++++++++++++++++++++++++++++++++-----------------------
 1 file changed, 51 insertions(+), 35 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 6cc4707..4a8089b 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -1,16 +1,20 @@
 CACTUS_URI="http://cactus.nci.nih.gov/chemical/structure/"
 require 'openbabel'
+require "base64"
 
 module OpenTox
 
-  # Ruby wrapper for OpenTox Compound Webservices (http://opentox.org/dev/apis/api-1.2/structure).
   class Compound
 
-    attr_reader :inchi
-
-    def initialize inchi
-      @inchi = inchi
-    end
+    field :inchi, type: String
+    attr_readonly :inchi
+    field :smiles, type: String
+    field :inchikey, type: String
+    field :names, type: Array
+    field :cid, type: String
+    field :chemblid, type: String
+    field :image_id, type: BSON::ObjectId
+    field :sdf_id, type: BSON::ObjectId
 
     def  == compound
       self.inchi == compound.inchi
@@ -22,21 +26,23 @@ module OpenTox
     # @param [String] smiles Smiles string
     # @return [OpenTox::Compound] Compound
     def self.from_smiles smiles
-      OpenTox::Compound.new obconversion(smiles,"smi","inchi")
+      # do not store smiles because it might be noncanonical
+      Compound.find_or_create_by :inchi => obconversion(smiles,"smi","inchi")
     end
 
     # Create a compound from inchi string
     # @param inchi [String] smiles InChI string
     # @return [OpenTox::Compound] Compound
     def self.from_inchi inchi
-      OpenTox::Compound.new inchi
+      Compound.find_or_create_by :inchi => inchi
     end
 
     # Create a compound from sdf string
     # @param sdf [String] smiles SDF string
     # @return [OpenTox::Compound] Compound
     def self.from_sdf sdf
-      OpenTox::Compound.new obconversion(sdf,"sdf","inchi")
+      # do not store sdf because it might be 2D
+      Compound.find_or_create_by :inchi => obconversion(sdf,"sdf","inchi")
     end
 
     # Create a compound from name. Relies on an external service for name lookups.
@@ -45,31 +51,32 @@ module OpenTox
     # @param name [String] can be also an InChI/InChiKey, CAS number, etc
     # @return [OpenTox::Compound] Compound
     def self.from_name name
-      OpenTox::Compound.new RestClientWrapper.get File.join(CACTUS_URI,URI.escape(name),"stdinchi")
+      Compound.find_or_create_by :inchi => RestClientWrapper.get(File.join(CACTUS_URI,URI.escape(name),"stdinchi"))
     end
 
     # Get InChIKey
     # @return [String] InChI string
     def inchikey
-      obconversion(@inchi,"inchi","inchikey")
+      update(:inchikey => obconversion(inchi,"inchi","inchikey")) unless self["inchikey"]
+      self["inchikey"]
     end
 
     # Get (canonical) smiles
     # @return [String] Smiles string
     def smiles
-      obconversion(@inchi,"inchi","smi") # "can" gives nonn-canonical smiles??
+      update(:smiles => obconversion(inchi,"inchi","smi")) unless self["smiles"] # should give canonical smiles, "can" seems to give incorrect results
+      self["smiles"]
     end
 
     # Get sdf
     # @return [String] SDF string
     def sdf
-      obconversion(@inchi,"inchi","sdf")
-    end
-
-    # Get gif image
-    # @return [image/gif] Image data
-    def gif
-      RestClientWrapper.get File.join(CACTUS_URI,inchi,"image")
+      if sdf_id.nil?
+        sdf = obconversion(inchi,"inchi","sdf")
+        file = Mongo::Grid::File.new(sdf, :filename => "#{id}.sdf",:content_type => "chemical/x-mdl-sdfile")
+        sdf_id = $gridfs.insert_one file
+      end
+      $gridfs.find_one(_id: sdf_id).data
     end
 
     # Get png image
@@ -77,41 +84,37 @@ module OpenTox
     #   image = compound.png
     # @return [image/png] Image data
     def png
-      obconversion(@inchi,"inchi","_png2")
-    end
+      if image_id.nil?
+       png = obconversion(inchi,"inchi","_png2")
+       file = Mongo::Grid::File.new(Base64.encode64(png), :filename => "#{id}.png", :content_type => "image/png")
+       update(:image_id => $gridfs.insert_one(file))
+      end
+      Base64.decode64($gridfs.find_one(_id: image_id).data)
 
-=begin
-    # Get URI of compound image
-    # @return [String] Compound image URI
-    def image_uri
-      File.join @data["uri"], "image"
     end
-=end
 
     # Get all known compound names. Relies on an external service for name lookups.
     # @example
     #   names = compound.names
     # @return [String] Compound names
     def names
-      RestClientWrapper.get("#{CACTUS_URI}#{inchi}/names").split("\n")
+      update(:names => RestClientWrapper.get("#{CACTUS_URI}#{inchi}/names").split("\n")) unless self["names"] 
+      self["names"]
     end
 
     # @return [String] PubChem Compound Identifier (CID), derieved via restcall to pubchem
     def cid
       pug_uri = "http://pubchem.ncbi.nlm.nih.gov/rest/pug/"
-      @cid ||= RestClientWrapper.post(File.join(pug_uri, "compound", "inchi", "cids", "TXT"),{:inchi => inchi}).strip
-    end
-
-    # @todo
-    def chebi
-      internal_server_error "not yet implemented"
+      update(:cid => RestClientWrapper.post(File.join(pug_uri, "compound", "inchi", "cids", "TXT"),{:inchi => inchi}).strip) unless self["cid"] 
+      self["cid"]
     end
 
     # @return [String] ChEMBL database compound id, derieved via restcall to chembl
     def chemblid
       # https://www.ebi.ac.uk/chembldb/ws#individualCompoundByInChiKey
       uri = "http://www.ebi.ac.uk/chemblws/compounds/smiles/#{smiles}.json"
-      @chemblid = JSON.parse(RestClientWrapper.get(uri))["compounds"].first["chemblId"]
+      update(:chemblid => JSON.parse(RestClientWrapper.get(uri))["compounds"].first["chemblId"]) unless self["chemblid"] 
+      self["chemblid"]
     end
 
     private
@@ -125,6 +128,19 @@ module OpenTox
       case output_format
       when /smi|can|inchi/
         obconversion.write_string(obmol).gsub(/\s/,'').chomp
+      when /sdf/
+        OpenBabel::OBOp.find_type("Gen3D").do(obmol) 
+        sdf = obconversion.write_string(obmol)
+        if sdf.match(/.nan/)
+          $logger.warn "3D generation failed for compound #{compound.inchi}, trying to calculate 2D structure"
+          OpenBabel::OBOp.find_type("Gen2D").do(obmol) 
+          sdf = obconversion.write_string(obmol)
+          if sdf.match(/.nan/)
+            $logger.warn "2D generation failed for compound #{compound.inchi}"
+            sdf = nil
+          end
+        end
+        sdf
       else
         obconversion.write_string(obmol)
       end
-- 
cgit v1.2.3


From 3cea6abe3606ea586b733e943737f77d58f215f9 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Mon, 27 Jul 2015 20:51:25 +0200
Subject: reasonable query performace for data_entries

---
 lib/compound.rb | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 4a8089b..93f609f 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -15,10 +15,12 @@ module OpenTox
     field :chemblid, type: String
     field :image_id, type: BSON::ObjectId
     field :sdf_id, type: BSON::ObjectId
+    #belongs_to :dataset
+    #belongs_to :data_entry
 
-    def  == compound
-      self.inchi == compound.inchi
-    end
+    #def  == compound
+      #self.inchi == compound.inchi
+    #end
 
     # Create a compound from smiles string
     # @example
-- 
cgit v1.2.3


From e68c69f90036bb7c47c57acb1ee3652b73c835c1 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Fri, 31 Jul 2015 10:59:53 +0200
Subject: descriptor tests working

---
 lib/compound.rb | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 93f609f..4d36915 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -73,12 +73,13 @@ module OpenTox
     # Get sdf
     # @return [String] SDF string
     def sdf
-      if sdf_id.nil?
+      if self.sdf_id.nil? 
         sdf = obconversion(inchi,"inchi","sdf")
         file = Mongo::Grid::File.new(sdf, :filename => "#{id}.sdf",:content_type => "chemical/x-mdl-sdfile")
         sdf_id = $gridfs.insert_one file
+        update :sdf_id => sdf_id
       end
-      $gridfs.find_one(_id: sdf_id).data
+      $gridfs.find_one(_id: self.sdf_id).data
     end
 
     # Get png image
@@ -86,12 +87,12 @@ module OpenTox
     #   image = compound.png
     # @return [image/png] Image data
     def png
-      if image_id.nil?
+      if self.image_id.nil?
        png = obconversion(inchi,"inchi","_png2")
        file = Mongo::Grid::File.new(Base64.encode64(png), :filename => "#{id}.png", :content_type => "image/png")
        update(:image_id => $gridfs.insert_one(file))
       end
-      Base64.decode64($gridfs.find_one(_id: image_id).data)
+      Base64.decode64($gridfs.find_one(_id: self.image_id).data)
 
     end
 
@@ -134,11 +135,11 @@ module OpenTox
         OpenBabel::OBOp.find_type("Gen3D").do(obmol) 
         sdf = obconversion.write_string(obmol)
         if sdf.match(/.nan/)
-          $logger.warn "3D generation failed for compound #{compound.inchi}, trying to calculate 2D structure"
+          $logger.warn "3D generation failed for compound #{identifier}, trying to calculate 2D structure"
           OpenBabel::OBOp.find_type("Gen2D").do(obmol) 
           sdf = obconversion.write_string(obmol)
           if sdf.match(/.nan/)
-            $logger.warn "2D generation failed for compound #{compound.inchi}"
+            $logger.warn "2D generation failed for compound #{identifier}"
             sdf = nil
           end
         end
-- 
cgit v1.2.3


From f93333d7f5df30875eae9135f4b5e65a98d923b6 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Fri, 31 Jul 2015 19:25:13 +0200
Subject: intermediary commit

---
 lib/compound.rb | 4 ++++
 1 file changed, 4 insertions(+)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 4d36915..0c27553 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -1,3 +1,7 @@
+# TODO: check
+# *** Open Babel Error  in ParseFile
+#    Could not find contribution data file.
+# 3d creation??
 CACTUS_URI="http://cactus.nci.nih.gov/chemical/structure/"
 require 'openbabel'
 require "base64"
-- 
cgit v1.2.3


From 04af01b8135ea147e9ce253e5526e3ee3adcc675 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Fri, 7 Aug 2015 19:50:09 +0200
Subject: initial k-nn weighted average implementation

---
 lib/compound.rb | 71 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 70 insertions(+), 1 deletion(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 0c27553..4d292f1 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -1,7 +1,7 @@
 # TODO: check
 # *** Open Babel Error  in ParseFile
 #    Could not find contribution data file.
-# 3d creation??
+
 CACTUS_URI="http://cactus.nci.nih.gov/chemical/structure/"
 require 'openbabel'
 require "base64"
@@ -9,6 +9,34 @@ require "base64"
 module OpenTox
 
   class Compound
+    include OpenTox
+
+# OpenBabel FP4 fingerprints
+# OpenBabel http://open-babel.readthedocs.org/en/latest/Fingerprints/intro.html
+# TODO store in DB
+fp4 = FingerprintSmarts.find
+unless fp4
+  fp4 = []
+  File.open(File.join(File.dirname(__FILE__),"SMARTS_InteLigand.txt")).each do |l| 
+    l.strip!
+    unless l.empty? or l.match /^#/
+      name,smarts = l.split(': ')
+      fp4 << OpenTox::FingerprintSmarts.find_or_create_by(:name => name, :smarts => smarts) unless smarts.nil?
+    end
+  end
+end
+FP4 = fp4
+
+# TODO investigate other types of fingerprints (MACCS)
+# OpenBabel http://open-babel.readthedocs.org/en/latest/Fingerprints/intro.html
+# http://www.dalkescientific.com/writings/diary/archive/2008/06/26/fingerprint_background.html
+# OpenBabel MNA http://openbabel.org/docs/dev/FileFormats/Multilevel_Neighborhoods_of_Atoms_(MNA).html#multilevel-neighborhoods-of-atoms-mna
+# Morgan ECFP, FCFP
+# http://cdk.github.io/cdk/1.5/docs/api/org/openscience/cdk/fingerprint/CircularFingerprinter.html
+# http://www.rdkit.org/docs/GettingStartedInPython.html
+# Chemfp
+# https://chemfp.readthedocs.org/en/latest/using-tools.html
+# CACTVS/PubChem
 
     field :inchi, type: String
     attr_readonly :inchi
@@ -19,6 +47,8 @@ module OpenTox
     field :chemblid, type: String
     field :image_id, type: BSON::ObjectId
     field :sdf_id, type: BSON::ObjectId
+    field :fp4, type: Array
+    field :fp4_size, type: Integer
     #belongs_to :dataset
     #belongs_to :data_entry
 
@@ -26,6 +56,22 @@ module OpenTox
       #self.inchi == compound.inchi
     #end
 
+    def self.find_or_create_by params
+      compound = self.find_or_initialize_by params
+      unless compound.fp4
+        compound.fp4_size = 0
+        compound.fp4 = []
+        Algorithm::Descriptor.smarts_match(compound, FP4.collect{|f| f.smarts}).each_with_index do |m,i|
+          if m > 0
+            compound.fp4 << FP4[i].id
+            compound.fp4_size += 1
+          end
+        end
+      end
+      compound.save
+      compound
+    end
+
     # Create a compound from smiles string
     # @example
     #   compound = OpenTox::Compound.from_smiles("c1ccccc1")
@@ -124,6 +170,29 @@ module OpenTox
       self["chemblid"]
     end
 
+    def neighbors threshold=0.3
+      # from http://blog.matt-swain.com/post/87093745652/chemical-similarity-search-in-mongodb
+      qn = fp4.size
+      qmin = qn * threshold
+      qmax = qn / threshold
+      #not sure if it is worth the effort of keeping feature counts up to date (compound deletions, additions, ...)
+      #reqbits = [count['_id'] for count in db.mfp_counts.find({'_id': {'$in': qfp}}).sort('count', 1).limit(qn - qmin + 1)]
+      aggregate = [
+        #{'$match': {'mfp.count': {'$gte': qmin, '$lte': qmax}, 'mfp.bits': {'$in': reqbits}}},
+        {'$project': {
+          'tanimoto': {'$let': {
+            'vars': {'common': {'$size': {'$setIntersection': ['$fp4', fp4]}}},
+            'in': {'$divide': ['$$common', {'$subtract': [{'$add': [qn, '$fp4_size']}, '$$common']}]}
+          }},
+          '_id': 1
+        }},
+        {'$match':  {'tanimoto': {'$gte': threshold}}},
+        {'$sort': {'tanimoto': -1}}
+      ]
+      
+      $mongo["compounds"].aggregate(aggregate).collect { |r| [Compound.find(r["_id"]), r["tanimoto"]]}
+    end
+
     private
 
     def self.obconversion(identifier,input_format,output_format,option=nil)
-- 
cgit v1.2.3


From ef76c077fd39d31fc795b842c32575f6afb9fdb2 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Sun, 9 Aug 2015 13:42:54 +0200
Subject: customized prediction algorithms implemented

---
 lib/compound.rb | 59 +++++++++++++++++++++++++++++----------------------------
 1 file changed, 30 insertions(+), 29 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 4d292f1..4e29938 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -11,32 +11,31 @@ module OpenTox
   class Compound
     include OpenTox
 
-# OpenBabel FP4 fingerprints
-# OpenBabel http://open-babel.readthedocs.org/en/latest/Fingerprints/intro.html
-# TODO store in DB
-fp4 = FingerprintSmarts.find
-unless fp4
-  fp4 = []
-  File.open(File.join(File.dirname(__FILE__),"SMARTS_InteLigand.txt")).each do |l| 
-    l.strip!
-    unless l.empty? or l.match /^#/
-      name,smarts = l.split(': ')
-      fp4 << OpenTox::FingerprintSmarts.find_or_create_by(:name => name, :smarts => smarts) unless smarts.nil?
+    # OpenBabel FP4 fingerprints
+    # OpenBabel http://open-babel.readthedocs.org/en/latest/Fingerprints/intro.html
+    fp4 = FingerprintSmarts.all
+    unless fp4
+      fp4 = []
+      File.open(File.join(File.dirname(__FILE__),"SMARTS_InteLigand.txt")).each do |l| 
+        l.strip!
+        unless l.empty? or l.match /^#/
+          name,smarts = l.split(': ')
+          fp4 << OpenTox::FingerprintSmarts.find_or_create_by(:name => name, :smarts => smarts) unless smarts.nil?
+        end
+      end
     end
-  end
-end
-FP4 = fp4
-
-# TODO investigate other types of fingerprints (MACCS)
-# OpenBabel http://open-babel.readthedocs.org/en/latest/Fingerprints/intro.html
-# http://www.dalkescientific.com/writings/diary/archive/2008/06/26/fingerprint_background.html
-# OpenBabel MNA http://openbabel.org/docs/dev/FileFormats/Multilevel_Neighborhoods_of_Atoms_(MNA).html#multilevel-neighborhoods-of-atoms-mna
-# Morgan ECFP, FCFP
-# http://cdk.github.io/cdk/1.5/docs/api/org/openscience/cdk/fingerprint/CircularFingerprinter.html
-# http://www.rdkit.org/docs/GettingStartedInPython.html
-# Chemfp
-# https://chemfp.readthedocs.org/en/latest/using-tools.html
-# CACTVS/PubChem
+    FP4 = fp4
+
+    # TODO investigate other types of fingerprints (MACCS)
+    # OpenBabel http://open-babel.readthedocs.org/en/latest/Fingerprints/intro.html
+    # http://www.dalkescientific.com/writings/diary/archive/2008/06/26/fingerprint_background.html
+    # OpenBabel MNA http://openbabel.org/docs/dev/FileFormats/Multilevel_Neighborhoods_of_Atoms_(MNA).html#multilevel-neighborhoods-of-atoms-mna
+    # Morgan ECFP, FCFP
+    # http://cdk.github.io/cdk/1.5/docs/api/org/openscience/cdk/fingerprint/CircularFingerprinter.html
+    # http://www.rdkit.org/docs/GettingStartedInPython.html
+    # Chemfp
+    # https://chemfp.readthedocs.org/en/latest/using-tools.html
+    # CACTVS/PubChem
 
     field :inchi, type: String
     attr_readonly :inchi
@@ -170,15 +169,16 @@ FP4 = fp4
       self["chemblid"]
     end
 
-    def neighbors threshold=0.3
+    def neighbors threshold=0.7
       # from http://blog.matt-swain.com/post/87093745652/chemical-similarity-search-in-mongodb
       qn = fp4.size
-      qmin = qn * threshold
-      qmax = qn / threshold
+      #qmin = qn * threshold
+      #qmax = qn / threshold
       #not sure if it is worth the effort of keeping feature counts up to date (compound deletions, additions, ...)
       #reqbits = [count['_id'] for count in db.mfp_counts.find({'_id': {'$in': qfp}}).sort('count', 1).limit(qn - qmin + 1)]
       aggregate = [
         #{'$match': {'mfp.count': {'$gte': qmin, '$lte': qmax}, 'mfp.bits': {'$in': reqbits}}},
+        {'$match':  {'_id': {'$ne': self.id}}}, # remove self
         {'$project': {
           'tanimoto': {'$let': {
             'vars': {'common': {'$size': {'$setIntersection': ['$fp4', fp4]}}},
@@ -190,7 +190,8 @@ FP4 = fp4
         {'$sort': {'tanimoto': -1}}
       ]
       
-      $mongo["compounds"].aggregate(aggregate).collect { |r| [Compound.find(r["_id"]), r["tanimoto"]]}
+      $mongo["compounds"].aggregate(aggregate).collect{ |r| [r["_id"], r["tanimoto"]] }
+        
     end
 
     private
-- 
cgit v1.2.3


From 23ecfc6fa5ae4913e5cd17b7d58432d1f88d780c Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Mon, 10 Aug 2015 09:48:57 +0200
Subject: transfer to new git project started

---
 lib/compound.rb | 40 +++++-----------------------------------
 1 file changed, 5 insertions(+), 35 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 4e29938..3ba1670 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -11,32 +11,6 @@ module OpenTox
   class Compound
     include OpenTox
 
-    # OpenBabel FP4 fingerprints
-    # OpenBabel http://open-babel.readthedocs.org/en/latest/Fingerprints/intro.html
-    fp4 = FingerprintSmarts.all
-    unless fp4
-      fp4 = []
-      File.open(File.join(File.dirname(__FILE__),"SMARTS_InteLigand.txt")).each do |l| 
-        l.strip!
-        unless l.empty? or l.match /^#/
-          name,smarts = l.split(': ')
-          fp4 << OpenTox::FingerprintSmarts.find_or_create_by(:name => name, :smarts => smarts) unless smarts.nil?
-        end
-      end
-    end
-    FP4 = fp4
-
-    # TODO investigate other types of fingerprints (MACCS)
-    # OpenBabel http://open-babel.readthedocs.org/en/latest/Fingerprints/intro.html
-    # http://www.dalkescientific.com/writings/diary/archive/2008/06/26/fingerprint_background.html
-    # OpenBabel MNA http://openbabel.org/docs/dev/FileFormats/Multilevel_Neighborhoods_of_Atoms_(MNA).html#multilevel-neighborhoods-of-atoms-mna
-    # Morgan ECFP, FCFP
-    # http://cdk.github.io/cdk/1.5/docs/api/org/openscience/cdk/fingerprint/CircularFingerprinter.html
-    # http://www.rdkit.org/docs/GettingStartedInPython.html
-    # Chemfp
-    # https://chemfp.readthedocs.org/en/latest/using-tools.html
-    # CACTVS/PubChem
-
     field :inchi, type: String
     attr_readonly :inchi
     field :smiles, type: String
@@ -48,21 +22,17 @@ module OpenTox
     field :sdf_id, type: BSON::ObjectId
     field :fp4, type: Array
     field :fp4_size, type: Integer
-    #belongs_to :dataset
-    #belongs_to :data_entry
-
-    #def  == compound
-      #self.inchi == compound.inchi
-    #end
 
+    # Overwrites standard Mongoid method to create fingerprints before database insertion
     def self.find_or_create_by params
       compound = self.find_or_initialize_by params
-      unless compound.fp4
+      unless compound.fp4 and !compound.fp4.empty?
         compound.fp4_size = 0
         compound.fp4 = []
-        Algorithm::Descriptor.smarts_match(compound, FP4.collect{|f| f.smarts}).each_with_index do |m,i|
+        fingerprint = FingerprintSmarts.fingerprint
+        Algorithm::Descriptor.smarts_match(compound, fingerprint).each_with_index do |m,i|
           if m > 0
-            compound.fp4 << FP4[i].id
+            compound.fp4 << fingerprint[i].id
             compound.fp4_size += 1
           end
         end
-- 
cgit v1.2.3


From b7cd3ebbb858a8891c35c45896f1bdd525f3534e Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Mon, 10 Aug 2015 13:26:06 +0200
Subject: algorithm libraries added, fminer tests pass

---
 lib/compound.rb | 2 --
 1 file changed, 2 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 3ba1670..3418fcc 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -3,8 +3,6 @@
 #    Could not find contribution data file.
 
 CACTUS_URI="http://cactus.nci.nih.gov/chemical/structure/"
-require 'openbabel'
-require "base64"
 
 module OpenTox
 
-- 
cgit v1.2.3


From 6ab86c253ba0eb79b9e6a20effa2d18626accf2b Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Thu, 13 Aug 2015 11:56:40 +0200
Subject: OpenBabel can (canonical smiles) instead of inchi as internal
 identifier to avoid OpenBabel InChi bug.

---
 lib/compound.rb | 54 ++++++++++++++++++++++++++++++++++++++++--------------
 1 file changed, 40 insertions(+), 14 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 3418fcc..5343aa0 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -10,13 +10,13 @@ module OpenTox
     include OpenTox
 
     field :inchi, type: String
-    attr_readonly :inchi
     field :smiles, type: String
     field :inchikey, type: String
     field :names, type: Array
     field :cid, type: String
     field :chemblid, type: String
-    field :image_id, type: BSON::ObjectId
+    field :png_id, type: BSON::ObjectId
+    field :svg_id, type: BSON::ObjectId
     field :sdf_id, type: BSON::ObjectId
     field :fp4, type: Array
     field :fp4_size, type: Integer
@@ -46,14 +46,18 @@ module OpenTox
     # @return [OpenTox::Compound] Compound
     def self.from_smiles smiles
       # do not store smiles because it might be noncanonical
-      Compound.find_or_create_by :inchi => obconversion(smiles,"smi","inchi")
+      Compound.find_or_create_by :smiles => obconversion(smiles,"smi","can")
     end
 
     # Create a compound from inchi string
     # @param inchi [String] smiles InChI string
     # @return [OpenTox::Compound] Compound
     def self.from_inchi inchi
-      Compound.find_or_create_by :inchi => inchi
+      # Temporary workaround for OpenBabels Inchi bug
+      # http://sourceforge.net/p/openbabel/bugs/957/
+      # bug has not been fixed in latest git/development version
+      smiles = `echo "#{inchi}" | babel -iinchi - -ocan`.chomp.strip
+      smiles.empty? ? nil : Compound.find_or_create_by(:smiles => smiles, :inchi => inchi)
     end
 
     # Create a compound from sdf string
@@ -61,7 +65,7 @@ module OpenTox
     # @return [OpenTox::Compound] Compound
     def self.from_sdf sdf
       # do not store sdf because it might be 2D
-      Compound.find_or_create_by :inchi => obconversion(sdf,"sdf","inchi")
+      Compound.find_or_create_by :smiles => obconversion(sdf,"sdf","can")
     end
 
     # Create a compound from name. Relies on an external service for name lookups.
@@ -70,20 +74,30 @@ module OpenTox
     # @param name [String] can be also an InChI/InChiKey, CAS number, etc
     # @return [OpenTox::Compound] Compound
     def self.from_name name
-      Compound.find_or_create_by :inchi => RestClientWrapper.get(File.join(CACTUS_URI,URI.escape(name),"stdinchi"))
+      Compound.find_or_create_by :smiles => RestClientWrapper.get(File.join(CACTUS_URI,URI.escape(name),"smiles"))
     end
 
-    # Get InChIKey
+    # Get InChI
     # @return [String] InChI string
+    def inchi
+      unless self["inchi"]
+        result = `echo "#{self.smiles}" | babel -ismi - -oinchi`.chomp
+        update(:inchi => result.chomp) unless result.empty?
+      end
+      self["inchi"]
+    end
+
+    # Get InChIKey
+    # @return [String] InChIKey string
     def inchikey
-      update(:inchikey => obconversion(inchi,"inchi","inchikey")) unless self["inchikey"]
+      update(:inchikey => obconversion(smiles,"smi","inchikey")) unless self["inchikey"]
       self["inchikey"]
     end
 
     # Get (canonical) smiles
     # @return [String] Smiles string
     def smiles
-      update(:smiles => obconversion(inchi,"inchi","smi")) unless self["smiles"] # should give canonical smiles, "can" seems to give incorrect results
+      update(:smiles => obconversion(self["smiles"],"smi","can")) #unless self["smiles"] # should give canonical smiles, "can" seems to give incorrect results
       self["smiles"]
     end
 
@@ -91,7 +105,7 @@ module OpenTox
     # @return [String] SDF string
     def sdf
       if self.sdf_id.nil? 
-        sdf = obconversion(inchi,"inchi","sdf")
+        sdf = obconversion(smiles,"smi","sdf")
         file = Mongo::Grid::File.new(sdf, :filename => "#{id}.sdf",:content_type => "chemical/x-mdl-sdfile")
         sdf_id = $gridfs.insert_one file
         update :sdf_id => sdf_id
@@ -99,17 +113,29 @@ module OpenTox
       $gridfs.find_one(_id: self.sdf_id).data
     end
 
+    # Get SVG image
+    # @return [image/svg] Image data
+    def svg
+      if self.svg_id.nil?
+       svg = obconversion(smiles,"smi","svg")
+       file = Mongo::Grid::File.new(svg, :filename => "#{id}.svg", :content_type => "image/svg")
+       update(:image_id => $gridfs.insert_one(file))
+      end
+      $gridfs.find_one(_id: self.svg_id).data
+
+    end
+
     # Get png image
     # @example
     #   image = compound.png
     # @return [image/png] Image data
     def png
-      if self.image_id.nil?
-       png = obconversion(inchi,"inchi","_png2")
+      if self.png_id.nil?
+       png = obconversion(smiles,"smi","_png2")
        file = Mongo::Grid::File.new(Base64.encode64(png), :filename => "#{id}.png", :content_type => "image/png")
-       update(:image_id => $gridfs.insert_one(file))
+       update(:png_id => $gridfs.insert_one(file))
       end
-      Base64.decode64($gridfs.find_one(_id: self.image_id).data)
+      Base64.decode64($gridfs.find_one(_id: self.png_id).data)
 
     end
 
-- 
cgit v1.2.3


From d0850e2983a219da214a67190fe881c7650f532f Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Thu, 13 Aug 2015 18:57:11 +0200
Subject: majority of tests working

---
 lib/compound.rb | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 5343aa0..10deabc 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -13,6 +13,7 @@ module OpenTox
     field :smiles, type: String
     field :inchikey, type: String
     field :names, type: Array
+    field :warning, type: String
     field :cid, type: String
     field :chemblid, type: String
     field :png_id, type: BSON::ObjectId
@@ -46,7 +47,12 @@ module OpenTox
     # @return [OpenTox::Compound] Compound
     def self.from_smiles smiles
       # do not store smiles because it might be noncanonical
-      Compound.find_or_create_by :smiles => obconversion(smiles,"smi","can")
+      smiles = obconversion(smiles,"smi","can")
+      if smiles.empty?
+        Compound.find_or_create_by(:warning => "SMILES parsing failed for '#{smiles}', this may be caused by an incorrect SMILES string.")
+      else
+        Compound.find_or_create_by :smiles => obconversion(smiles,"smi","can")
+      end
     end
 
     # Create a compound from inchi string
@@ -57,7 +63,11 @@ module OpenTox
       # http://sourceforge.net/p/openbabel/bugs/957/
       # bug has not been fixed in latest git/development version
       smiles = `echo "#{inchi}" | babel -iinchi - -ocan`.chomp.strip
-      smiles.empty? ? nil : Compound.find_or_create_by(:smiles => smiles, :inchi => inchi)
+      if smiles.empty?
+        Compound.find_or_create_by(:warning => "InChi parsing failed for #{inchi}, this may be caused by an incorrect InChi string or a bug in OpenBabel libraries.")
+      else
+        Compound.find_or_create_by(:smiles => smiles, :inchi => inchi)
+      end
     end
 
     # Create a compound from sdf string
-- 
cgit v1.2.3


From 4067d78c78c1f503b5f21d043ece8042260a3d78 Mon Sep 17 00:00:00 2001
From: ch <ch@ch-test2>
Date: Tue, 18 Aug 2015 16:37:09 +0200
Subject: initial installation fixes, installs on Debian 3.16.7

---
 lib/compound.rb | 28 ++++++++++++++++------------
 1 file changed, 16 insertions(+), 12 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 10deabc..e51e9ad 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -62,7 +62,7 @@ module OpenTox
       # Temporary workaround for OpenBabels Inchi bug
       # http://sourceforge.net/p/openbabel/bugs/957/
       # bug has not been fixed in latest git/development version
-      smiles = `echo "#{inchi}" | babel -iinchi - -ocan`.chomp.strip
+      smiles = `echo "#{inchi}" | "#{File.join(File.dirname(__FILE__),"..","openbabel","bin","babel")}" -iinchi - -ocan`.chomp.strip
       if smiles.empty?
         Compound.find_or_create_by(:warning => "InChi parsing failed for #{inchi}, this may be caused by an incorrect InChi string or a bug in OpenBabel libraries.")
       else
@@ -91,7 +91,8 @@ module OpenTox
     # @return [String] InChI string
     def inchi
       unless self["inchi"]
-        result = `echo "#{self.smiles}" | babel -ismi - -oinchi`.chomp
+
+        result = `echo "#{self.smiles}" | "#{File.join(File.dirname(__FILE__),"..","openbabel","bin","babel")}" -ismi - -oinchi`.chomp
         update(:inchi => result.chomp) unless result.empty?
       end
       self["inchi"]
@@ -182,21 +183,23 @@ module OpenTox
       #reqbits = [count['_id'] for count in db.mfp_counts.find({'_id': {'$in': qfp}}).sort('count', 1).limit(qn - qmin + 1)]
       aggregate = [
         #{'$match': {'mfp.count': {'$gte': qmin, '$lte': qmax}, 'mfp.bits': {'$in': reqbits}}},
-        {'$match':  {'_id': {'$ne': self.id}}}, # remove self
-        {'$project': {
-          'tanimoto': {'$let': {
-            'vars': {'common': {'$size': {'$setIntersection': ['$fp4', fp4]}}},
-            'in': {'$divide': ['$$common', {'$subtract': [{'$add': [qn, '$fp4_size']}, '$$common']}]}
+        {'$match' =>  {'_id' => {'$ne' => self.id}}}, # remove self
+        {'$project' => {
+          'tanimoto' => {'$let' => {
+            'vars' => {'common' => {'$size' => {'$setIntersection' => ['$fp4', fp4]}}},
+            'in' => {'$divide' => ['$$common', {'$subtract' => [{'$add' => [qn, '$fp4_size']}, '$$common']}]}
           }},
-          '_id': 1
+          '_id' => 1
         }},
-        {'$match':  {'tanimoto': {'$gte': threshold}}},
-        {'$sort': {'tanimoto': -1}}
+        {'$match' =>  {'tanimoto' => {'$gte' => threshold}}},
+        {'$sort' => {'tanimoto' => -1}}
       ]
       
       $mongo["compounds"].aggregate(aggregate).collect{ |r| [r["_id"], r["tanimoto"]] }
         
     end
+=begin
+=end
 
     private
 
@@ -210,11 +213,12 @@ module OpenTox
       when /smi|can|inchi/
         obconversion.write_string(obmol).gsub(/\s/,'').chomp
       when /sdf/
-        OpenBabel::OBOp.find_type("Gen3D").do(obmol) 
+# TODO set 3D
+        #OpenBabel::OBOp.find_type("Gen3D").do(obmol) 
         sdf = obconversion.write_string(obmol)
         if sdf.match(/.nan/)
           $logger.warn "3D generation failed for compound #{identifier}, trying to calculate 2D structure"
-          OpenBabel::OBOp.find_type("Gen2D").do(obmol) 
+          #OpenBabel::OBOp.find_type("Gen2D").do(obmol) 
           sdf = obconversion.write_string(obmol)
           if sdf.match(/.nan/)
             $logger.warn "2D generation failed for compound #{identifier}"
-- 
cgit v1.2.3


From 36e10aa98a3d19a78a1889857262abe338a05eef Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Wed, 19 Aug 2015 12:57:23 +0200
Subject: libfminer rpath fixed, installation instructions

---
 lib/compound.rb | 13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index e51e9ad..fdec019 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -213,11 +213,22 @@ module OpenTox
       when /smi|can|inchi/
         obconversion.write_string(obmol).gsub(/\s/,'').chomp
       when /sdf/
-# TODO set 3D
+p "SDF conversion"
+        # has no effect
+	#obconversion.add_option("gen3D", OpenBabel::OBConversion::GENOPTIONS)
+        # segfaults with openbabel git master 
         #OpenBabel::OBOp.find_type("Gen3D").do(obmol) 
+
+        builder = OpenBabel::OBBuilder.new
+        builder.build(obmol);
+
         sdf = obconversion.write_string(obmol)
+print sdf
         if sdf.match(/.nan/)
+          
+# TODO: fix or eliminate 2d generation
           $logger.warn "3D generation failed for compound #{identifier}, trying to calculate 2D structure"
+          obconversion.set_options("gen2D", OpenBabel::OBConversion::GENOPTIONS)
           #OpenBabel::OBOp.find_type("Gen2D").do(obmol) 
           sdf = obconversion.write_string(obmol)
           if sdf.match(/.nan/)
-- 
cgit v1.2.3


From 57bb9c2c691d350dce8c30b862030973043818cc Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Wed, 19 Aug 2015 14:00:55 +0200
Subject: hydrogens added for 3d generation

---
 lib/compound.rb | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index fdec019..6d755da 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -46,7 +46,6 @@ module OpenTox
     # @param [String] smiles Smiles string
     # @return [OpenTox::Compound] Compound
     def self.from_smiles smiles
-      # do not store smiles because it might be noncanonical
       smiles = obconversion(smiles,"smi","can")
       if smiles.empty?
         Compound.find_or_create_by(:warning => "SMILES parsing failed for '#{smiles}', this may be caused by an incorrect SMILES string.")
@@ -62,7 +61,8 @@ module OpenTox
       # Temporary workaround for OpenBabels Inchi bug
       # http://sourceforge.net/p/openbabel/bugs/957/
       # bug has not been fixed in latest git/development version
-      smiles = `echo "#{inchi}" | "#{File.join(File.dirname(__FILE__),"..","openbabel","bin","babel")}" -iinchi - -ocan`.chomp.strip
+      #smiles = `echo "#{inchi}" | "#{File.join(File.dirname(__FILE__),"..","openbabel","bin","babel")}" -iinchi - -ocan`.chomp.strip
+      smiles = obconversion(inchi,"inchi","can")
       if smiles.empty?
         Compound.find_or_create_by(:warning => "InChi parsing failed for #{inchi}, this may be caused by an incorrect InChi string or a bug in OpenBabel libraries.")
       else
@@ -92,7 +92,8 @@ module OpenTox
     def inchi
       unless self["inchi"]
 
-        result = `echo "#{self.smiles}" | "#{File.join(File.dirname(__FILE__),"..","openbabel","bin","babel")}" -ismi - -oinchi`.chomp
+        result = obconversion(smiles,"smi","inchi")
+        #result = `echo "#{self.smiles}" | "#{File.join(File.dirname(__FILE__),"..","openbabel","bin","babel")}" -ismi - -oinchi`.chomp
         update(:inchi => result.chomp) unless result.empty?
       end
       self["inchi"]
@@ -108,7 +109,7 @@ module OpenTox
     # Get (canonical) smiles
     # @return [String] Smiles string
     def smiles
-      update(:smiles => obconversion(self["smiles"],"smi","can")) #unless self["smiles"] # should give canonical smiles, "can" seems to give incorrect results
+      update(:smiles => obconversion(self["smiles"],"smi","can")) unless self["smiles"] 
       self["smiles"]
     end
 
@@ -219,8 +220,12 @@ p "SDF conversion"
         # segfaults with openbabel git master 
         #OpenBabel::OBOp.find_type("Gen3D").do(obmol) 
 
+        # TODO: find disconnected structures
+        # strip_salts
+        # separate
+        obmol.add_hydrogens
         builder = OpenBabel::OBBuilder.new
-        builder.build(obmol);
+        builder.build(obmol)
 
         sdf = obconversion.write_string(obmol)
 print sdf
-- 
cgit v1.2.3


From 4fb010f4c0f30e5b23d39e191ce36587827f8f4e Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Wed, 19 Aug 2015 14:17:44 +0200
Subject: name conversion fixed

---
 lib/compound.rb | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 6d755da..c0d6536 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -75,7 +75,7 @@ module OpenTox
     # @return [OpenTox::Compound] Compound
     def self.from_sdf sdf
       # do not store sdf because it might be 2D
-      Compound.find_or_create_by :smiles => obconversion(sdf,"sdf","can")
+      Compound.from_smiles obconversion(sdf,"sdf","can")
     end
 
     # Create a compound from name. Relies on an external service for name lookups.
@@ -84,7 +84,7 @@ module OpenTox
     # @param name [String] can be also an InChI/InChiKey, CAS number, etc
     # @return [OpenTox::Compound] Compound
     def self.from_name name
-      Compound.find_or_create_by :smiles => RestClientWrapper.get(File.join(CACTUS_URI,URI.escape(name),"smiles"))
+      Compound.from_smiles RestClientWrapper.get(File.join(CACTUS_URI,URI.escape(name),"smiles"))
     end
 
     # Get InChI
-- 
cgit v1.2.3


From 8c6c59980bc82dc2177147f2fe34adf8bfbc1539 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Mon, 24 Aug 2015 12:17:07 +0200
Subject: Model::Prediction with tests, initial indexes

---
 lib/compound.rb | 2 ++
 1 file changed, 2 insertions(+)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index c0d6536..fa57aff 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -22,6 +22,8 @@ module OpenTox
     field :fp4, type: Array
     field :fp4_size, type: Integer
 
+    index({smiles: 1}, {unique: true})
+
     # Overwrites standard Mongoid method to create fingerprints before database insertion
     def self.find_or_create_by params
       compound = self.find_or_initialize_by params
-- 
cgit v1.2.3


From f8faf510b4574df1a00fa61a9f0a1681fc2f4857 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Tue, 25 Aug 2015 17:20:55 +0200
Subject: Experiments added

---
 lib/compound.rb | 1 +
 1 file changed, 1 insertion(+)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index fa57aff..a819f56 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -178,6 +178,7 @@ module OpenTox
     end
 
     def neighbors threshold=0.7
+      # TODO restrict to dataset
       # from http://blog.matt-swain.com/post/87093745652/chemical-similarity-search-in-mongodb
       qn = fp4.size
       #qmin = qn * threshold
-- 
cgit v1.2.3


From 66c34e4682965272060a121b6e362af67ed4be5f Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Thu, 27 Aug 2015 20:28:25 +0200
Subject: Initial GUI for Nestec models

---
 lib/compound.rb | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index a819f56..8f393f5 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -133,7 +133,7 @@ module OpenTox
       if self.svg_id.nil?
        svg = obconversion(smiles,"smi","svg")
        file = Mongo::Grid::File.new(svg, :filename => "#{id}.svg", :content_type => "image/svg")
-       update(:image_id => $gridfs.insert_one(file))
+       update(:svg_id => $gridfs.insert_one(file))
       end
       $gridfs.find_one(_id: self.svg_id).data
 
-- 
cgit v1.2.3


From 96a476a2331daa4d1d6b5ac444bbdbd2ac221a5f Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Thu, 10 Sep 2015 12:54:18 +0200
Subject: tests fixed (crossvalidations may fail due to memory constraints)

---
 lib/compound.rb | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 8f393f5..6adf3c0 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -218,11 +218,6 @@ module OpenTox
         obconversion.write_string(obmol).gsub(/\s/,'').chomp
       when /sdf/
 p "SDF conversion"
-        # has no effect
-	#obconversion.add_option("gen3D", OpenBabel::OBConversion::GENOPTIONS)
-        # segfaults with openbabel git master 
-        #OpenBabel::OBOp.find_type("Gen3D").do(obmol) 
-
         # TODO: find disconnected structures
         # strip_salts
         # separate
@@ -234,14 +229,13 @@ p "SDF conversion"
 print sdf
         if sdf.match(/.nan/)
           
-# TODO: fix or eliminate 2d generation
           $logger.warn "3D generation failed for compound #{identifier}, trying to calculate 2D structure"
           obconversion.set_options("gen2D", OpenBabel::OBConversion::GENOPTIONS)
-          #OpenBabel::OBOp.find_type("Gen2D").do(obmol) 
           sdf = obconversion.write_string(obmol)
           if sdf.match(/.nan/)
-            $logger.warn "2D generation failed for compound #{identifier}"
-            sdf = nil
+            $logger.warn "2D generation failed for compound #{identifier}, rendering without coordinates."
+            obconversion.remove_option("gen2D", OpenBabel::OBConversion::GENOPTIONS)
+            sdf = obconversion.write_string(obmol)
           end
         end
         sdf
-- 
cgit v1.2.3


From 2fdecbed76c4db8dfe3f10f825fed9772e653197 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Wed, 16 Sep 2015 16:52:18 +0200
Subject: generic openbabel fingerprints

---
 lib/compound.rb | 31 +++++++++++++++++++++++++++++--
 1 file changed, 29 insertions(+), 2 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 6adf3c0..7f175ca 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -42,6 +42,35 @@ module OpenTox
       compound
     end
 
+    def openbabel_fingerprint type="FP2"
+      fp = OpenBabel::OBFingerprint.find_fingerprint(type)
+      obmol = OpenBabel::OBMol.new
+      obconversion = OpenBabel::OBConversion.new
+      obconversion.set_in_format "smi"
+      obconversion.read_string obmol, smiles
+      result = OpenBabel::VectorUnsignedInt.new
+      fp.get_fingerprint(obmol,result)
+      # TODO: %ignore *::DescribeBits @ line 163 openbabel/scripts/openbabel-ruby.i
+      #p OpenBabel::OBFingerprint.describe_bits(result)
+      result = result.to_a
+      # convert result to a list of the bits that are set
+      # from openbabel/scripts/python/pybel.py line 830
+      # see also http://openbabel.org/docs/dev/UseTheLibrary/Python_Pybel.html#fingerprints
+      bitsperint = OpenBabel::OBFingerprint.getbitsperint()
+      bits_set = []
+      start = 1
+      result.each do |x|
+        i = start
+        while x > 0 do
+          bits_set << i if (x % 2) == 1
+          x >>= 1
+          i += 1
+        end
+        start += bitsperint
+      end
+      bits_set
+    end
+
     # Create a compound from smiles string
     # @example
     #   compound = OpenTox::Compound.from_smiles("c1ccccc1")
@@ -202,8 +231,6 @@ module OpenTox
       $mongo["compounds"].aggregate(aggregate).collect{ |r| [r["_id"], r["tanimoto"]] }
         
     end
-=begin
-=end
 
     private
 
-- 
cgit v1.2.3


From 6ac119c32cef094d4f1c2fb5c2daa4e274401f70 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Thu, 17 Sep 2015 14:56:25 +0200
Subject: neighbor calculation moved to Compound class

---
 lib/compound.rb | 84 ++++++++++++++++++++++++++++++++++++++++-----------------
 1 file changed, 60 insertions(+), 24 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 7f175ca..7abd913 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -19,8 +19,11 @@ module OpenTox
     field :png_id, type: BSON::ObjectId
     field :svg_id, type: BSON::ObjectId
     field :sdf_id, type: BSON::ObjectId
+    field :fp2, type: Array
+    field :fp3, type: Array
     field :fp4, type: Array
     field :fp4_size, type: Integer
+    field :maccs, type: Array
 
     index({smiles: 1}, {unique: true})
 
@@ -43,32 +46,35 @@ module OpenTox
     end
 
     def openbabel_fingerprint type="FP2"
-      fp = OpenBabel::OBFingerprint.find_fingerprint(type)
-      obmol = OpenBabel::OBMol.new
-      obconversion = OpenBabel::OBConversion.new
-      obconversion.set_in_format "smi"
-      obconversion.read_string obmol, smiles
-      result = OpenBabel::VectorUnsignedInt.new
-      fp.get_fingerprint(obmol,result)
-      # TODO: %ignore *::DescribeBits @ line 163 openbabel/scripts/openbabel-ruby.i
-      #p OpenBabel::OBFingerprint.describe_bits(result)
-      result = result.to_a
-      # convert result to a list of the bits that are set
-      # from openbabel/scripts/python/pybel.py line 830
-      # see also http://openbabel.org/docs/dev/UseTheLibrary/Python_Pybel.html#fingerprints
-      bitsperint = OpenBabel::OBFingerprint.getbitsperint()
-      bits_set = []
-      start = 1
-      result.each do |x|
-        i = start
-        while x > 0 do
-          bits_set << i if (x % 2) == 1
-          x >>= 1
-          i += 1
+      unless self.send(type.downcase.to_sym) # stored fingerprint
+        fp = OpenBabel::OBFingerprint.find_fingerprint(type)
+        obmol = OpenBabel::OBMol.new
+        obconversion = OpenBabel::OBConversion.new
+        obconversion.set_in_format "smi"
+        obconversion.read_string obmol, smiles
+        result = OpenBabel::VectorUnsignedInt.new
+        fp.get_fingerprint(obmol,result)
+        # TODO: %ignore *::DescribeBits @ line 163 openbabel/scripts/openbabel-ruby.i
+        #p OpenBabel::OBFingerprint.describe_bits(result)
+        # convert result to a list of the bits that are set
+        # from openbabel/scripts/python/pybel.py line 830
+        # see also http://openbabel.org/docs/dev/UseTheLibrary/Python_Pybel.html#fingerprints
+        result = result.to_a
+        bitsperint = OpenBabel::OBFingerprint.getbitsperint()
+        bits_set = []
+        start = 1
+        result.each do |x|
+          i = start
+          while x > 0 do
+            bits_set << i if (x % 2) == 1
+            x >>= 1
+            i += 1
+          end
+          start += bitsperint
         end
-        start += bitsperint
+        update type.downcase.to_sym, bits_set
       end
-      bits_set
+      self.send(type.downcase.to_sym) 
     end
 
     # Create a compound from smiles string
@@ -206,6 +212,36 @@ module OpenTox
       self["chemblid"]
     end
 
+    def fingerprint_neighbors params
+      bad_request_error "Incorrect parameters '#{params}' for Compound#fingerprint_neighbors. Please provide :type, :training_dataset_id, :min_sim." unless params[:type] and params[:training_dataset_id] and params[:min_sim]
+      neighbors = []
+      query_fingerprint = self.openbabel_fingerprint params[:type]
+      training_dataset = Dataset.find(params[:training_dataset_id]).compounds.each do |compound|
+        unless self == compound
+          fingerprint = compound.openbabel_fingerprint params[:type]
+          sim = (query_fingerprint & fingerprint).size/(query_fingerprint | fingerprint).size.to_f
+          neighbors << [compound.id, sim] if sim >= params[:min_sim]
+        end
+      end
+      neighbors.sort{|a,b| b.last <=> a.last}
+    end
+
+    def fminer_neighbors params
+      bad_request_error "Incorrect parameters for Compound#fminer_neighbors. Please provide :feature_dataset_id, :min_sim." unless params[:feature_dataset_id] and params[:min_sim]
+      feature_dataset = Dataset.find params[:feature_dataset_id]
+      query_fingerprint = Algorithm::Descriptor.smarts_match(self, feature_dataset.features)
+      neighbors = []
+
+      # find neighbors
+      feature_dataset.data_entries.each_with_index do |fingerprint, i|
+        sim = Algorithm::Similarity.tanimoto fingerprint, query_fingerprint
+        if sim >= params[:min_sim]
+          neighbors << [feature_dataset.compound_ids[i],sim] # use compound_ids, instantiation of Compounds is too time consuming
+        end
+      end
+      neighbors
+    end
+
     def neighbors threshold=0.7
       # TODO restrict to dataset
       # from http://blog.matt-swain.com/post/87093745652/chemical-similarity-search-in-mongodb
-- 
cgit v1.2.3


From d5bf97c2cb999539c56bf59aa1d7d3286745be84 Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Wed, 23 Sep 2015 14:51:41 +0200
Subject: validations fixed (all models were executed with default parameters)

---
 lib/compound.rb | 41 +++++++++++++++++++++++++++++++++++++++--
 1 file changed, 39 insertions(+), 2 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index 7abd913..d3df125 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -44,6 +44,21 @@ module OpenTox
       compound.save
       compound
     end
+    
+    #http://openbabel.org/docs/dev/FileFormats/MolPrint2D_format.html#molprint2d-format
+    def mpd 
+      smarts = obconversion(smiles,"smi","mpd").strip.split("\t")
+      smarts.shift # remove Title
+      smarts
+
+    end
+
+    #http://openbabel.org/docs/dev/FileFormats/Multilevel_Neighborhoods_of_Atoms_(MNA).html
+    def mna level=2
+      smarts = obconversion(smiles,"smi","mna","xL\"#{level}\"").split("\n")
+      smarts.shift # remove Title
+      smarts
+    end
 
     def openbabel_fingerprint type="FP2"
       unless self.send(type.downcase.to_sym) # stored fingerprint
@@ -72,7 +87,7 @@ module OpenTox
           end
           start += bitsperint
         end
-        update type.downcase.to_sym, bits_set
+        update_attribute type.downcase.to_sym, bits_set
       end
       self.send(type.downcase.to_sym) 
     end
@@ -242,6 +257,28 @@ module OpenTox
       neighbors
     end
 
+    def physchem_neighbors params
+      feature_dataset = Dataset.find params[:feature_dataset_id]
+      query_fingerprint = Algorithm.run params[:feature_calculation_algorithm], self, params[:descriptors]
+      neighbors = []
+      feature_dataset.data_entries.each_with_index do |fingerprint, i|
+        # TODO implement pearson and cosine similarity separatly
+        R.assign "x", query_fingerprint
+        R.assign "y", fingerprint
+        # pearson r
+        #sim = R.eval("cor(x,y,use='complete.obs',method='pearson')").to_ruby
+        #p "pearson"
+        #p sim
+        #p "cosine"
+        sim = R.eval("x %*% y / sqrt(x%*%x * y%*%y)").to_ruby.first
+        #p sim
+        if sim >= params[:min_sim]
+          neighbors << [feature_dataset.compound_ids[i],sim] # use compound_ids, instantiation of Compounds is too time consuming
+        end
+      end
+      neighbors
+    end
+
     def neighbors threshold=0.7
       # TODO restrict to dataset
       # from http://blog.matt-swain.com/post/87093745652/chemical-similarity-search-in-mongodb
@@ -308,7 +345,7 @@ print sdf
     end
 
     def obconversion(identifier,input_format,output_format,option=nil)
-      self.class.obconversion(identifier,input_format,output_format,option=nil)
+      self.class.obconversion(identifier,input_format,output_format,option)
     end
   end
 end
-- 
cgit v1.2.3


From 8d2f1c8a0f6cc9f7a481d1117bf8b3351130b1ea Mon Sep 17 00:00:00 2001
From: Christoph Helma <helma@in-silico.ch>
Date: Wed, 7 Oct 2015 12:34:02 +0200
Subject: generalised fingerprints

---
 lib/compound.rb | 171 +++++++++++++++++++++++++++++++-------------------------
 1 file changed, 95 insertions(+), 76 deletions(-)

(limited to 'lib/compound.rb')

diff --git a/lib/compound.rb b/lib/compound.rb
index d3df125..7a3dc5c 100644
--- a/lib/compound.rb
+++ b/lib/compound.rb
@@ -9,6 +9,8 @@ module OpenTox
   class Compound
     include OpenTox
 
+    DEFAULT_FINGERPRINT = "MP2D"
+
     field :inchi, type: String
     field :smiles, type: String
     field :inchikey, type: String
@@ -19,77 +21,64 @@ module OpenTox
     field :png_id, type: BSON::ObjectId
     field :svg_id, type: BSON::ObjectId
     field :sdf_id, type: BSON::ObjectId
-    field :fp2, type: Array
-    field :fp3, type: Array
-    field :fp4, type: Array
-    field :fp4_size, type: Integer
-    field :maccs, type: Array
+    field :fingerprints, type: Hash, default: {}
+    field :default_fingerprint_size, type: Integer
 
     index({smiles: 1}, {unique: true})
 
     # Overwrites standard Mongoid method to create fingerprints before database insertion
     def self.find_or_create_by params
       compound = self.find_or_initialize_by params
-      unless compound.fp4 and !compound.fp4.empty?
-        compound.fp4_size = 0
-        compound.fp4 = []
-        fingerprint = FingerprintSmarts.fingerprint
-        Algorithm::Descriptor.smarts_match(compound, fingerprint).each_with_index do |m,i|
-          if m > 0
-            compound.fp4 << fingerprint[i].id
-            compound.fp4_size += 1
-          end
-        end
-      end
+      compound.default_fingerprint_size = compound.fingerprint(DEFAULT_FINGERPRINT)
       compound.save
       compound
     end
-    
-    #http://openbabel.org/docs/dev/FileFormats/MolPrint2D_format.html#molprint2d-format
-    def mpd 
-      smarts = obconversion(smiles,"smi","mpd").strip.split("\t")
-      smarts.shift # remove Title
-      smarts
-
-    end
-
-    #http://openbabel.org/docs/dev/FileFormats/Multilevel_Neighborhoods_of_Atoms_(MNA).html
-    def mna level=2
-      smarts = obconversion(smiles,"smi","mna","xL\"#{level}\"").split("\n")
-      smarts.shift # remove Title
-      smarts
-    end
 
-    def openbabel_fingerprint type="FP2"
-      unless self.send(type.downcase.to_sym) # stored fingerprint
-        fp = OpenBabel::OBFingerprint.find_fingerprint(type)
-        obmol = OpenBabel::OBMol.new
-        obconversion = OpenBabel::OBConversion.new
-        obconversion.set_in_format "smi"
-        obconversion.read_string obmol, smiles
-        result = OpenBabel::VectorUnsignedInt.new
-        fp.get_fingerprint(obmol,result)
-        # TODO: %ignore *::DescribeBits @ line 163 openbabel/scripts/openbabel-ruby.i
-        #p OpenBabel::OBFingerprint.describe_bits(result)
-        # convert result to a list of the bits that are set
-        # from openbabel/scripts/python/pybel.py line 830
-        # see also http://openbabel.org/docs/dev/UseTheLibrary/Python_Pybel.html#fingerprints
-        result = result.to_a
-        bitsperint = OpenBabel::OBFingerprint.getbitsperint()
-        bits_set = []
-        start = 1
-        result.each do |x|
-          i = start
-          while x > 0 do
-            bits_set << i if (x % 2) == 1
-            x >>= 1
-            i += 1
+    def fingerprint type="MP2D"
+      unless fingerprints[type]
+        return [] unless self.smiles
+        #http://openbabel.org/docs/dev/FileFormats/MolPrint2D_format.html#molprint2d-format
+        if type == "MP2D"
+          fp = obconversion(smiles,"smi","mpd").strip.split("\t")
+          name = fp.shift # remove Title
+          fingerprints[type] = fp
+        #http://openbabel.org/docs/dev/FileFormats/Multilevel_Neighborhoods_of_Atoms_(MNA).html
+        elsif type== "MNA"
+          level = 2 # TODO: level as parameter, evaluate level 1, see paper
+          fp = obconversion(smiles,"smi","mna","xL\"#{level}\"").split("\n")
+          fp.shift # remove Title
+          fingerprints[type] = fp
+        else # standard fingerprints
+          fp = OpenBabel::OBFingerprint.find_fingerprint(type)
+          obmol = OpenBabel::OBMol.new
+          obconversion = OpenBabel::OBConversion.new
+          obconversion.set_in_format "smi"
+          obconversion.read_string obmol, self.smiles
+          result = OpenBabel::VectorUnsignedInt.new
+          fp.get_fingerprint(obmol,result)
+          # TODO: %ignore *::DescribeBits @ line 163 openbabel/scripts/openbabel-ruby.i
+          #p OpenBabel::OBFingerprint.describe_bits(result)
+          # convert result to a list of the bits that are set
+          # from openbabel/scripts/python/pybel.py line 830
+          # see also http://openbabel.org/docs/dev/UseTheLibrary/Python_Pybel.html#fingerprints
+          result = result.to_a
+          bitsperint = OpenBabel::OBFingerprint.getbitsperint()
+          bits_set = []
+          start = 1
+          result.each do |x|
+            i = start
+            while x > 0 do
+              bits_set << i if (x % 2) == 1
+              x >>= 1
+              i += 1
+            end
+            start += bitsperint
           end
-          start += bitsperint
+          fingerprints[type] = bits_set
         end
-        update_attribute type.downcase.to_sym, bits_set
+        save
       end
-      self.send(type.downcase.to_sym) 
+      fingerprints[type]
     end
 
     # Create a compound from smiles string
@@ -100,7 +89,8 @@ module OpenTox
     def self.from_smiles smiles
       smiles = obconversion(smiles,"smi","can")
       if smiles.empty?
-        Compound.find_or_create_by(:warning => "SMILES parsing failed for '#{smiles}', this may be caused by an incorrect SMILES string.")
+        return nil
+        #Compound.find_or_create_by(:warning => "SMILES parsing failed for '#{smiles}', this may be caused by an incorrect SMILES string.")
       else
         Compound.find_or_create_by :smiles => obconversion(smiles,"smi","can")
       end
@@ -146,7 +136,7 @@ module OpenTox
 
         result = obconversion(smiles,"smi","inchi")
         #result = `echo "#{self.smiles}" | "#{File.join(File.dirname(__FILE__),"..","openbabel","bin","babel")}" -ismi - -oinchi`.chomp
-        update(:inchi => result.chomp) unless result.empty?
+        update(:inchi => result.chomp) if result and !result.empty?
       end
       self["inchi"]
     end
@@ -227,20 +217,47 @@ module OpenTox
       self["chemblid"]
     end
 
-    def fingerprint_neighbors params
-      bad_request_error "Incorrect parameters '#{params}' for Compound#fingerprint_neighbors. Please provide :type, :training_dataset_id, :min_sim." unless params[:type] and params[:training_dataset_id] and params[:min_sim]
+    def fingerprint_count_neighbors params
+      # TODO fix
       neighbors = []
-      query_fingerprint = self.openbabel_fingerprint params[:type]
+      query_fingerprint = self.fingerprint params[:type]
       training_dataset = Dataset.find(params[:training_dataset_id]).compounds.each do |compound|
         unless self == compound
-          fingerprint = compound.openbabel_fingerprint params[:type]
-          sim = (query_fingerprint & fingerprint).size/(query_fingerprint | fingerprint).size.to_f
-          neighbors << [compound.id, sim] if sim >= params[:min_sim]
+          candidate_fingerprint = compound.fingerprint params[:type]
+          features = (query_fingerprint + candidate_fingerprint).uniq
+          min_sum = 0
+          max_sum = 0
+          features.each do |f|
+            min,max = [query_fingerprint.count(f),candidate_fingerprint.count(f)].minmax
+            min_sum += min
+            max_sum += max
+          end
+          max_sum == 0 ? sim = 0 : sim = min_sum/max_sum.to_f
+          neighbors << [compound.id, sim] if sim and sim >= params[:min_sim]
         end
       end
       neighbors.sort{|a,b| b.last <=> a.last}
     end
 
+    def fingerprint_neighbors params
+      bad_request_error "Incorrect parameters '#{params}' for Compound#fingerprint_neighbors. Please provide :type, :training_dataset_id, :min_sim." unless params[:type] and params[:training_dataset_id] and params[:min_sim]
+      neighbors = []
+      #if params[:type] == DEFAULT_FINGERPRINT
+        #neighbors = db_neighbors params
+        #p neighbors
+      #else
+        query_fingerprint = self.fingerprint params[:type]
+        training_dataset = Dataset.find(params[:training_dataset_id]).compounds.each do |compound|
+          unless self == compound
+            candidate_fingerprint = compound.fingerprint params[:type]
+            sim = (query_fingerprint & candidate_fingerprint).size/(query_fingerprint | candidate_fingerprint).size.to_f
+            neighbors << [compound.id, sim] if sim >= params[:min_sim]
+          end
+        end
+      #end
+      neighbors.sort{|a,b| b.last <=> a.last}
+    end
+
     def fminer_neighbors params
       bad_request_error "Incorrect parameters for Compound#fminer_neighbors. Please provide :feature_dataset_id, :min_sim." unless params[:feature_dataset_id] and params[:min_sim]
       feature_dataset = Dataset.find params[:feature_dataset_id]
@@ -248,8 +265,8 @@ module OpenTox
       neighbors = []
 
       # find neighbors
-      feature_dataset.data_entries.each_with_index do |fingerprint, i|
-        sim = Algorithm::Similarity.tanimoto fingerprint, query_fingerprint
+      feature_dataset.data_entries.each_with_index do |candidate_fingerprint, i|
+        sim = Algorithm::Similarity.tanimoto candidate_fingerprint, query_fingerprint
         if sim >= params[:min_sim]
           neighbors << [feature_dataset.compound_ids[i],sim] # use compound_ids, instantiation of Compounds is too time consuming
         end
@@ -261,10 +278,10 @@ module OpenTox
       feature_dataset = Dataset.find params[:feature_dataset_id]
       query_fingerprint = Algorithm.run params[:feature_calculation_algorithm], self, params[:descriptors]
       neighbors = []
-      feature_dataset.data_entries.each_with_index do |fingerprint, i|
+      feature_dataset.data_entries.each_with_index do |candidate_fingerprint, i|
         # TODO implement pearson and cosine similarity separatly
         R.assign "x", query_fingerprint
-        R.assign "y", fingerprint
+        R.assign "y", candidate_fingerprint
         # pearson r
         #sim = R.eval("cor(x,y,use='complete.obs',method='pearson')").to_ruby
         #p "pearson"
@@ -279,10 +296,12 @@ module OpenTox
       neighbors
     end
 
-    def neighbors threshold=0.7
+    def db_neighbors params
+      p "DB NEIGHBORS"
+      p params
       # TODO restrict to dataset
       # from http://blog.matt-swain.com/post/87093745652/chemical-similarity-search-in-mongodb
-      qn = fp4.size
+      qn = fingerprint(params[:type]).size
       #qmin = qn * threshold
       #qmax = qn / threshold
       #not sure if it is worth the effort of keeping feature counts up to date (compound deletions, additions, ...)
@@ -292,12 +311,12 @@ module OpenTox
         {'$match' =>  {'_id' => {'$ne' => self.id}}}, # remove self
         {'$project' => {
           'tanimoto' => {'$let' => {
-            'vars' => {'common' => {'$size' => {'$setIntersection' => ['$fp4', fp4]}}},
-            'in' => {'$divide' => ['$$common', {'$subtract' => [{'$add' => [qn, '$fp4_size']}, '$$common']}]}
+            'vars' => {'common' => {'$size' => {'$setIntersection' => ["'$#{DEFAULT_FINGERPRINT}'", DEFAULT_FINGERPRINT]}}},
+            'in' => {'$divide' => ['$$common', {'$subtract' => [{'$add' => [qn, '$default_fingerprint_size']}, '$$common']}]}
           }},
           '_id' => 1
         }},
-        {'$match' =>  {'tanimoto' => {'$gte' => threshold}}},
+        {'$match' =>  {'tanimoto' => {'$gte' => params[:min_sim]}}},
         {'$sort' => {'tanimoto' => -1}}
       ]
       
@@ -312,12 +331,12 @@ module OpenTox
       obconversion.set_options(option, OpenBabel::OBConversion::OUTOPTIONS) if option
       obmol = OpenBabel::OBMol.new
       obconversion.set_in_and_out_formats input_format, output_format
+      return nil if identifier.nil?
       obconversion.read_string obmol, identifier
       case output_format
       when /smi|can|inchi/
         obconversion.write_string(obmol).gsub(/\s/,'').chomp
       when /sdf/
-p "SDF conversion"
         # TODO: find disconnected structures
         # strip_salts
         # separate
-- 
cgit v1.2.3