rdkit/Code/GraphMol/Wrap/test_cdxml.py

#
#  Copyright (C) 2003-2021  Greg Landrum and other RDKit contributors
#         All Rights Reserved
#
""" This is a rough coverage test of the python wrapper

it's intended to be shallow, but broad

"""

import doctest
import gc
import gzip
import logging
import os
import sys
import tempfile
import unittest
from io import StringIO

from rdkit import Chem


class TestCase(unittest.TestCase):

  def test_cdxml(self):
    cdxml = """<?xml version="1.0" encoding="UTF-8" ?>
        <!DOCTYPE CDXML SYSTEM "http://www.cambridgesoft.com/xml/cdxml.dtd" >
        <CDXML
         CreationProgram="ChemDraw JS 2.0.0.9"
         Name="ACS Document 1996"
         BoundingBox="94.75 178.16 154.89 211.17"
         WindowPosition="0 0"
         WindowSize="0 0"
         FractionalWidths="yes"
         InterpretChemically="yes"
         ShowAtomQuery="yes"
         ShowAtomStereo="no"
         ShowAtomEnhancedStereo="yes"
         ShowAtomNumber="no"
         ShowResidueID="no"
         ShowBondQuery="yes"
         ShowBondRxn="yes"
         ShowBondStereo="no"
         ShowTerminalCarbonLabels="no"
         ShowNonTerminalCarbonLabels="no"
         HideImplicitHydrogens="no"
         Magnification="666"
         LabelFont="24"
         LabelSize="10"
         LabelFace="96"
         CaptionFont="24"
         CaptionSize="10"
         HashSpacing="2.50"
         MarginWidth="1.60"
         LineWidth="0.60"
         BoldWidth="2"
         BondLength="14.40"
         BondSpacing="18"
         ChainAngle="120"
         LabelJustification="Auto"
         CaptionJustification="Left"
         AminoAcidTermini="HOH"
         ShowSequenceTermini="yes"
         ShowSequenceBonds="yes"
         ShowSequenceUnlinkedBranches="no"
         ResidueWrapCount="40"
         ResidueBlockCount="10"
         ResidueZigZag="yes"
         NumberResidueBlocks="no"
         PrintMargins="36 36 36 36"
         MacPrintInfo="0003000001200120000000000B6608A0FF84FF880BE309180367052703FC0002000001200120000000000B6608A0000100000064000000010001010100000001270F000100010000000000000000000000000002001901900000000000400000000000000000000100000000000000000000000000000000"
         ChemPropName=""
         ChemPropFormula="Chemical Formula: "
         ChemPropExactMass="Exact Mass: "
         ChemPropMolWt="Molecular Weight: "
         ChemPropMOverZ="m/z: "
         ChemPropAnalysis="Elemental Analysis: "
         ChemPropBoilingPt="Boiling Point: "
         ChemPropMeltingPt="Melting Point: "
         ChemPropCritTemp="Critical Temp: "
         ChemPropCritPres="Critical Pres: "
         ChemPropCritVol="Critical Vol: "
         ChemPropGibbs="Gibbs Energy: "
         ChemPropLogP="Log P: "
         ChemPropMR="MR: "
         ChemPropHenry="Henry&apos;s Law: "
         ChemPropEForm="Heat of Form: "
         ChemProptPSA="tPSA: "
         ChemPropID=""
         ChemPropFragmentLabel=""
         color="0"
         bgcolor="1"
         RxnAutonumberStart="1"
         RxnAutonumberConditions="no"
         RxnAutonumberStyle="Roman"
        ><colortable>
        <color r="1" g="1" b="1"/>
        <color r="0" g="0" b="0"/>
        <color r="1" g="0" b="0"/>
        <color r="1" g="1" b="0"/>
        <color r="0" g="1" b="0"/>
        <color r="0" g="1" b="1"/>
        <color r="0" g="0" b="1"/>
        <color r="1" g="0" b="1"/>
        </colortable><fonttable>
        <font id="24" charset="utf-8" name="Arial"/>
        </fonttable><page
         id="32"
         BoundingBox="0 0 542 354"
         Width="542"
         Height="354"
         HeaderPosition="36"
         FooterPosition="36"
         PageOverlap="0"
         PrintTrimMarks="yes"
         HeightPages="1"
         WidthPages="1"
         DrawingSpace="poster"
        ><fragment
         id="10"
         BoundingBox="94.75 178.16 154.89 211.17"
         Z="4"
        ><n
         id="7"
         p="95.05 187.47"
         Z="1"
         AS="N"
        /><n
         id="9"
         p="95.05 201.87"
         Z="3"
         AS="N"
        /><n
         id="11"
         p="106.31 210.84"
         Z="5"
         AS="N"
        /><n
         id="13"
         p="120.35 207.64"
         Z="7"
         AS="N"
        /><n
         id="15"
         p="126.59 194.67"
         Z="9"
         AS="N"
        /><n
         id="17"
         p="120.35 181.69"
         Z="11"
         AS="N"
        /><n
         id="19"
         p="106.31 178.49"
         Z="13"
         AS="N"
        /><n
         id="28"
         p="140.99 194.67"
         Z="22"
         NodeType="Nickname"
         NeedsClean="yes"
         AS="N"
        ><fragment
         id="33"
        ><n
         id="34"
         p="148.17 207.09"
         Element="8"
         NumHydrogens="0"
        /><n
         id="35"
         p="162.52 207.09"
        /><n
         id="36"
         p="176.87 207.09"
        /><n
         id="37"
         p="169.69 194.67"
        /><n
         id="38"
         p="169.69 219.52"
        /><n
         id="39"
         p="140.99 194.67"
        /><n
         id="40"
         p="148.17 182.24"
         Element="8"
         NumHydrogens="0"
        /><n
         id="41"
         p="126.64 194.67"
         NodeType="ExternalConnectionPoint"
        /><b
         id="42"
         B="39"
         E="40"
         Order="2"
        /><b
         id="43"
         B="35"
         E="38"
        /><b
         id="44"
         B="35"
         E="36"
        /><b
         id="45"
         B="35"
         E="37"
        /><b
         id="46"
         B="34"
         E="35"
        /><b
         id="47"
         B="34"
         E="39"
        /><b
         id="48"
         B="41"
         E="39"
        /></fragment><t
         p="137.66 198.28"
         BoundingBox="137.66 189.64 154.89 198.28"
         LabelJustification="Left"
         LabelAlignment="Left"
        ><s font="24" size="9.95" color="0" face="96">Boc</s></t></n><b
         id="21"
         Z="15"
         B="7"
         E="9"
         BS="N"
        /><b
         id="22"
         Z="16"
         B="9"
         E="11"
         BS="N"
        /><b
         id="23"
         Z="17"
         B="11"
         E="13"
         BS="N"
        /><b
         id="24"
         Z="18"
         B="13"
         E="15"
         BS="N"
        /><b
         id="25"
         Z="19"
         B="15"
         E="17"
         BS="N"
        /><b
         id="26"
         Z="20"
         B="17"
         E="19"
         BS="N"
        /><b
         id="27"
         Z="21"
         B="19"
         E="7"
         BS="N"
        /><b
         id="29"
         Z="23"
         B="15"
         E="28"
         BS="N"
        /></fragment></page></CDXML>"""
    mols = Chem.MolsFromCDXML(cdxml)
    self.assertEqual(len(mols), 1)
    self.assertEqual(Chem.MolToSmiles(mols[0]), "CC(C)(C)OC(=O)C1CCCCCC1")


if __name__ == '__main__':
  if "RDTESTCASE" in os.environ:
    suite = unittest.TestSuite()
    testcases = os.environ["RDTESTCASE"]
    for name in testcases.split(':'):
      suite.addTest(TestCase(name))

    runner = unittest.TextTestRunner()
    runner.run(suite)
  else:
    unittest.main()