changeset 0:764340994e71

Uploaded

--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/rdkit_descriptors.py	Sat Apr 27 09:01:41 2013 -0400
@@ -0,0 +1,71 @@
+#!/usr/bin/env python
+
+from rdkit.Chem import Descriptors
+from rdkit import Chem
+import sys, os, re
+import argparse
+import inspect
+
+def get_supplier( infile, format = 'smiles' ):
+    with open(infile) as handle:
+        for line in handle:
+            line = line.strip()
+            if format == 'smiles':
+                mol = Chem.MolFromSmiles( line, sanitize=True )
+            elif format == 'inchi':
+                mol = Chem.inchi.MolFromInchi( line, sanitize=True, removeHs=True, logLevel=None, treatWarningAsError=False )
+            if mol is None:
+                yield False
+            else:
+                mol.SetProp( '_Name', line.split('\t')[0] )
+                yield mol
+
+
+def get_rdkit_descriptor_functions():
+    ret = [ (name, f) for name, f in inspect.getmembers( Descriptors ) if inspect.isfunction( f ) and not name.startswith( '_' ) ]
+    ret.sort()
+    return ret
+
+
+def descriptors( mol, functions ):
+    """
+    Calculates the descriptors of a given molecule.
+    """
+
+    for name, function in functions:
+        yield (name, function( mol ))
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument('-i', '--infile', required=True, help='Path to the input file.')
+    parser.add_argument("--iformat", help="Specify the input file format.")
+
+    parser.add_argument('-o', '--outfile', type=argparse.FileType('w+'), 
+        default=sys.stdout, help="path to the result file, default it sdtout")
+
+    parser.add_argument("--header", dest="header", action="store_true",
+                    default=False,
+                    help="Write header line.")
+
+    args = parser.parse_args()
+
+    if args.iformat == 'sdf':
+        supplier = Chem.SDMolSupplier( args.infile )
+    elif args.iformat =='smi':
+        supplier = get_supplier( args.infile, format = 'smiles' )
+    elif args.iformat == 'inchi':
+        supplier = get_supplier( args.infile, format = 'inchi' )
+
+    functions = get_rdkit_descriptor_functions()
+
+    if args.header:
+        args.outfile.write( '%s\n' % '\t'.join( [name for name, f in functions] ) )
+
+    for mol in supplier:
+        if not mol:
+            continue
+        descs = descriptors( mol, functions )
+        name = mol.GetProp("_Name")
+        args.outfile.write( "%s\n" % '\t'.join( [name]+ [str(res) for name, res in descs] ) )
+

--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/rdkit_descriptors.xml	Sat Apr 27 09:01:41 2013 -0400
@@ -0,0 +1,49 @@
+<tool id="ctb_rdkit_describtors" name="Descriptors" version="0.1">
+    <description>calculated with RDKit</description>
+    <parallelism method="multi" split_inputs="infile" split_mode="to_size" split_size="10000" shared_inputs="" merge_outputs="outfile"></parallelism>
+    <requirements>
+        <requirement type="package" version="2012_12_1">rdkit</requirement>
+    </requirements>
+    <command interpreter="python">rdkit_descriptors.py -i "${infile}" --iformat "${infile.ext}" -o "${outfile}" $header 2>&#38;1</command>
+    <inputs>
+        <param format="smi,sdf" name="infile" type="data" label="Molecule data in SD- or SMILES-format" help="Dataset missing? See TIP below"/>
+        <param name="header" type="boolean" label="Include the descriptor name as header" truevalue="--header" falsevalue="" checked="false" />
+    </inputs>
+    <outputs>
+        <data format="tabular" name="outfile" />
+    </outputs>
+    <tests>
+    </tests>
+    <help>
+
+.. class:: infomark
+
+**TIP:** If your data is not in SD- or SMILES format, use the pencil icon on your dataset to convert your data
+
+-----
+
+**Example**
+
+- Input file::
+
+    - SD-Format (http://en.wikipedia.org/wiki/Chemical_table_file)
+    - SMILES Format (http://en.wikipedia.org/wiki/Simplified_molecular_input_line_entry_specification)
+
+- Result::
+
+
+http://code.google.com/p/rdkit/wiki/DescriptorsInTheRDKit
+http://rdkit.org/Python_Docs/rdkit.Chem.Descriptors-module.html
+
+
+-----
+
+
+**Cite**
+
+Greg Landrum
+[1] RDKit: Open-source cheminformatics; http://www.rdkit.org
+
+
+    </help>
+</tool>

--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/repository_dependencies.xml	Sat Apr 27 09:01:41 2013 -0400
@@ -0,0 +1,4 @@
+<?xml version="1.0"?>
+<repositories description="This requires the Molecule datatype definitions (e.g. SMILES, InChI, SD-format).">
+    <repository toolshed="http://testtoolshed.g2.bx.psu.edu/" name="molecule_datatypes" owner="bgruening" changeset_revision="e533de975501" />
+</repositories> 

--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/tool_dependencies.xml	Sat Apr 27 09:01:41 2013 -0400
@@ -0,0 +1,5 @@
+<tool_dependency>
+    <package name="rdkit" version="2012_12_1">
+        <repository toolshed="http://testtoolshed.g2.bx.psu.edu/" name="package_rdkit_2012_12" owner="bgruening" changeset_revision="0adc830e7d71" prior_installation_required="True" />
+    </package>
+</tool_dependency>