Source Code for Module rdkit.Chem.FragmentCatalog

  1  # $Id: FragmentCatalog.py 997 2009-02-25 06:12:43Z glandrum $ 
  2  # 
  3  #  Copyright (C) 2003-2008 Greg Landrum and Rational Discovery LLC 
  4  # 
  5  #   @@ All Rights Reserved  @@ 
  6  # 
  7  from rdkit import Chem 
  8  from rdfragcatalog import * 
  9  import sys 
 10  import sets 
 11   
 12 -def message(msg,dest=sys.stdout): 
 13    dest.write(msg) 
 14   
 15   
 16 -class BitGainsInfo(object): 
 17    id=-1 
 18    description='' 
 19    gain=0.0 
 20    nPerClass=None 
 21     
 22 -def ProcessGainsFile(fileName,nToDo=-1,delim=',',haveDescriptions=1): 
 23    inFile = open(fileName,'r') 
 24    nRead = 0 
 25    res = [] 
 26    for line in inFile.xreadlines(): 
 27      nRead += 1 
 28      splitL = [x.strip() for x in line.split(delim)] 
 29      if nRead != 1 and len(splitL): 
 30        bit = BitGainsInfo() 
 31        bit.id = int(splitL[0]) 
 32        col = 1 
 33        if haveDescriptions: 
 34          bit.description = splitL[col] 
 35          col += 1 
 36        bit.gain = float(splitL[col]) 
 37        col += 1 
 38        nPerClass = [] 
 39        for entry in splitL[col:]: 
 40          nPerClass.append(int(entry)) 
 41        bit.nPerClass = nPerClass   
 42        res.append(bit) 
 43        if len(res)==nToDo: 
 44          break 
 45    return res    
 46   
 47 -def BuildAdjacencyList(catalog,bits,limitInclusion=1,orderLevels=0): 
 48    adjs = {} 
 49    levels = {} 
 50    bitIds = [bit.id for bit in bits] 
 51    for bitId in bitIds: 
 52      entry = catalog.GetBitEntryId(bitId) 
 53      tmp = [] 
 54      order = catalog.GetEntryOrder(entry) 
 55      s = levels.get(order,sets.Set()) 
 56      s.add(bitId) 
 57      levels[order] = s 
 58      for down in catalog.GetEntryDownIds(entry): 
 59        id = catalog.GetEntryBitId(down) 
 60        if not limitInclusion or id in bitIds: 
 61          tmp.append(id) 
 62          order = catalog.GetEntryOrder(down) 
 63          s = levels.get(order,sets.Set()) 
 64          s.add(id) 
 65          levels[order] = s 
 66      adjs[bitId] = tmp 
 67    if orderLevels: 
 68      # we'll play a little game and sort the indices in each level by 
 69      #  the number of downlinks they have: 
 70      for order in levels.keys(): 
 71        ids = levels[order] 
 72        counts = [len(adjs[id]) for id in ids] 
 73        countOrder = argsort(counts) 
 74        l = [ids[x] for x in countOrder] 
 75        l.reverse() 
 76        levels[order] = l 
 77    return adjs,levels 
 78   
 79 -def GetMolsMatchingBit(mols,bit,fps): 
 80    res = [] 
 81    if isinstance(bit,BitGainsInfo): 
 82      bitId = bit.id 
 83    else: 
 84      bitId = bit 
 85    for i,mol in enumerate(mols): 
 86      fp = fps[i] 
 87      if fp[bitId]: 
 88        res.append(mol) 
 89    return res 
 90   
 91  xl = None 
 92 -def ShowMolsMatchingBit(mols,bit,fps,actName="",wrapper=None,col=1,row=1): 
 93    global xl 
 94    from rdkit.Excel import Molecules 
 95    if wrapper is not None: 
 96      xl = wrapper 
 97    if xl is None: 
 98      xl = Molecules.ExcelWrapper() 
 99      xl.Workbooks.Add() 
100      xl.ChemdrawNewWorksheet() 
101    mols = GetMolsMatchingBit(mols,bit,fps) 
102    i = row 
103    for mol in mols: 
104      xl[i,col] = Chem.MolToSmiles(mol) 
105      if actName and mol.HasProp(actName): 
106        xl[i,col+1] = mol.GetProp(actName) 
107      i+=1 
108    i = row 
109    for mol in mols: 
110      try: 
111        xl.ChemdrawConvertCellsToMols(i,col) 
112        xl.ChemdrawShowPictures(i,col) 
113      except: 
114        pass 
115      i+=1 
116    return mols 
117