B 3Rc_@s6dZddlmZddlZddlZddlZddlZddlmZddlm Z ddl m Z ddl m Z ddl m Z dd l mZdd lmZdZdZeed rydd lmZmZWnek rYnXdd lmZddlmZeeZiZdZdZdZ dZ!ddZ"ddZ#ddZ$y"ddl%m&Z&e$e&_$e!e&_!ee&_Wn e'k rPe(dYnXdSddZ)y ddl*Z+Wnek r~YnXdTd!d"Z,dUd#d$Z-d%d&Z.ydd'l/m0Z0d(d)Z1Wnek rd*d)Z1YnXd+d,Z2dVd.d/Z3dWd1d2Z4dXd3d4Z5da6dYd5d6Z7dZd8d9Z8d[d;d<Z9d\d=d>Z:d]d@dAZ;dBdCZdHdIZ?d_dJdKZ@e>edLkr2ddlAZAy ddlBZBWnek rdZBYnXyddl*Z+dMdNZCWnek rdZ+YnXGdOdPdPeAjDZDdZEye+eCdQk ZEWneFk rd ZEYnXeEr(e@eAGn e(dRdS)`a Importing pandasTools enables several features that allow for using RDKit molecules as columns of a Pandas dataframe. If the dataframe is containing a molecule format in a column (e.g. smiles), like in this example: >>> from rdkit.Chem import PandasTools >>> import pandas as pd >>> import os >>> from rdkit import RDConfig >>> antibiotics = pd.DataFrame(columns=['Name','Smiles']) >>> antibiotics = antibiotics.append({'Smiles':'CC1(C(N2C(S1)C(C2=O)NC(=O)CC3=CC=CC=C3)C(=O)O)C', ... 'Name':'Penicilline G'}, ignore_index=True)#Penicilline G >>> antibiotics = antibiotics.append({ ... 'Smiles':'CC1(C2CC3C(C(=O)C(=C(C3(C(=O)C2=C(C4=C1C=CC=C4O)O)O)O)C(=O)N)N(C)C)O', ... 'Name':'Tetracycline'}, ignore_index=True)#Tetracycline >>> antibiotics = antibiotics.append({ ... 'Smiles':'CC1(C(N2C(S1)C(C2=O)NC(=O)C(C3=CC=CC=C3)N)C(=O)O)C', ... 'Name':'Ampicilline'}, ignore_index=True)#Ampicilline >>> print([str(x) for x in antibiotics.columns]) ['Name', 'Smiles'] >>> print(antibiotics) Name Smiles 0 Penicilline G CC1(C(N2C(S1)C(C2=O)NC(=O)CC3=CC=CC=C3)C(=O)O)C 1 Tetracycline CC1(C2CC3C(C(=O)C(=C(C3(C(=O)C2=C(C4=C1C=CC=C4... 2 Ampicilline CC1(C(N2C(S1)C(C2=O)NC(=O)C(C3=CC=CC=C3)N)C(=O... a new column can be created holding the respective RDKit molecule objects. The fingerprint can be included to accelerate substructure searches on the dataframe. >>> PandasTools.AddMoleculeColumnToFrame(antibiotics,'Smiles','Molecule',includeFingerprints=True) >>> print([str(x) for x in antibiotics.columns]) ['Name', 'Smiles', 'Molecule'] A substructure filter can be applied on the dataframe using the RDKit molecule column, because the ">=" operator has been modified to work as a substructure check. Such the antibiotics containing the beta-lactam ring "C1C(=O)NC1" can be obtained by >>> beta_lactam = Chem.MolFromSmiles('C1C(=O)NC1') >>> beta_lactam_antibiotics = antibiotics[antibiotics['Molecule'] >= beta_lactam] >>> print(beta_lactam_antibiotics[['Name','Smiles']]) Name Smiles 0 Penicilline G CC1(C(N2C(S1)C(C2=O)NC(=O)CC3=CC=CC=C3)C(=O)O)C 2 Ampicilline CC1(C(N2C(S1)C(C2=O)NC(=O)C(C3=CC=CC=C3)N)C(=O... It is also possible to load an SDF file can be load into a dataframe. >>> sdfFile = os.path.join(RDConfig.RDDataDir,'NCI/first_200.props.sdf') >>> frame = PandasTools.LoadSDF(sdfFile,smilesName='SMILES',molColName='Molecule', ... includeFingerprints=True) >>> frame.info # doctest: +SKIP Int64Index: 200 entries, 0 to 199 Data columns: AMW 200 non-null values CLOGP 200 non-null values CP 200 non-null values CR 200 non-null values DAYLIGHT.FPG 200 non-null values DAYLIGHT_CLOGP 200 non-null values FP 200 non-null values ID 200 non-null values ISM 200 non-null values LIPINSKI_VIOLATIONS 200 non-null values NUM_HACCEPTORS 200 non-null values NUM_HDONORS 200 non-null values NUM_HETEROATOMS 200 non-null values NUM_LIPINSKIHACCEPTORS 200 non-null values NUM_LIPINSKIHDONORS 200 non-null values NUM_RINGS 200 non-null values NUM_ROTATABLEBONDS 200 non-null values P1 30 non-null values SMILES 200 non-null values Molecule 200 non-null values dtypes: object(20)> The standard ForwardSDMolSupplier keywords are also available: >>> sdfFile = os.path.join(RDConfig.RDDataDir,'NCI/first_200.props.sdf') >>> frame = PandasTools.LoadSDF(sdfFile, smilesName='SMILES', molColName='Molecule', ... includeFingerprints=True, removeHs=False, strictParsing=True) Conversion to html is quite easy: >>> htm = frame.to_html() # doctest: ... >>> str(htm[:36]) '' In order to support rendering the molecules as images in the HTML export of the dataframe, we use a custom formatter for columns containing RDKit molecules, and also disable escaping of HTML where needed. ) b64encodeN)Chem) DataStructs)AllChem)Draw)SDWriter)rdchem)MurckoScaffoldZIPythonConsole)InteractiveRenderer drawOptions)BytesIO)minidomTZpng)rcentercCsv|dks|dkrdSt|drHt|ds4t|d|_t|j|jsHdS||}g|_|rntrjt||_dSdSdS)a*Allows for substructure check using the >= operator (X has substructure Y -> X >= Y) by monkey-patching the __ge__ function This has the effect that the pandas/numpy rowfilter can be used for substructure filtering (filtered = dframe[dframe['RDKitColumn'] >= SubstructureMolecule]) NF _substructfpT) hasattr_fingerprinterrrZAllProbeBitsMatchZGetSubstructMatch __sssAtomshighlightSubstructureslist)xymatchr5lib/python3.7/site-packages/rdkit/Chem/PandasTools.py_molges     rcCs(ttds ttddtdt|S)z*Deprecated, use PrintAsImageString insteadZ warning_shownTzEPrintAsBase64PNGString is deprecated - use PrintAsImageString instead)rPrintAsBase64PNGStringsetattrlogwarningPrintAsImageString)rrrrrs   rc Cstrt|dr|j}ng}tdk}trRt|rRddtD}tj|||dS|rt j |t|ddt d}t |}|dd }d td d td |jd <dtd dtd d|jd<d|jd<|St j|t|dddt d}dtd dtd dt|SdS)z;Returns the molecules as base64 encoded PNG image or as SVGrsvgcSsg|]}td|qS))max).0srrr sz&PrintAsImageString..)useSVGT)kekulizer rz0 0  Zviewboxz max-width: z px; height: zpx;Zstylezrdkit/moleculez data-content)Z returnPNGr)r z
Mol
N)rrrmolRepresentationlowerr Z isEnabledmolSizeZgenerateHTMLBodyrZ _moltoSVGr r Z parseStringZgetElementsByTagNameZ attributesZtoxmlZ _moltoimg _get_image)rZhighlightAtomsr'sizer!datarrrr s$    r r+) PandasPatcherzDFailed to patch pandas - PandasTools will have limited functionalitycCs2yt|Wntk r,tdYnXdS)aNChanges the default dataframe rendering to not escape HTML characters, thus allowing rendered images in all dataframes. IMPORTANT: THIS IS A GLOBAL CHANGE THAT WILL AFFECT TO COMPLETE PYTHON SESSION. If you want to change the rendering only for a single dataframe use the "ChangeMoleculeRendering" method instead. z szLoadSDF.._Name)isomericSmilesz2No valid smiles could be generated for molecule %s)index) isinstancestrr-gzipopenclosebool enumeraterZForwardSDMolSupplierdictZ GetPropNamesZ ClearPropZHasPropr< MolToSmiles Exceptionrr_MolPlusFingerprintappendpd DataFrameChangeMoleculeRendering)filenameidName molColNameincludeFingerprintsrAZ smilesNameZ embedPropsr:r;rEfrGZrecordsindicesr9irowZpropdfr)r>rLoadSDFsN            r[c Csdgt|}|rtddlm}xP|D]D\}}|dkr>q,x0t|D]$\}} t| ||<|||qHWq,W|s| d|d=||d<t j ||d} | S)a returns a dataframe with the results of R-Group Decomposition >>> from rdkit import Chem >>> from rdkit.Chem import rdRGroupDecomposition >>> from rdkit.Chem import PandasTools >>> import pandas as pd >>> scaffold = Chem.MolFromSmiles('c1ccccn1') >>> mols = [Chem.MolFromSmiles(smi) for smi in 'c1c(F)cccn1 c1c(Cl)c(C)ccn1 c1c(O)cccn1 c1c(F)c(C)ccn1 c1cc(Cl)c(F)cn1'.split()] >>> groups,_ = rdRGroupDecomposition.RGroupDecompose([scaffold],mols,asSmiles=False,asRows=False) >>> df = PandasTools.RGroupDecompositionToFrame(groups,mols,include_core=True) >>> list(df.columns) ['Mol', 'Core', 'R1', 'R2'] >>> len(df) 5 >>> df.columns() # doctest: +SKIP RangeIndex: 5 entries, 0 to 4 Data columns (total 4 columns): Mol 5 non-null object Core 5 non-null object R1 5 non-null object R2 5 non-null object dtypes: object(4) memory usage: *...* Molr) rdDepictorZCore)columns) rkeys rdkit.Chemr]itemsrIrZRemoveHsCompute2DCoordsremoverOrP) groupsZmolsZ include_coreZredraw_sidechainscolsr]r=ZvlrXvframerrrRGroupDecompositionToFrame&s  rhcCst|dS)z!displayhook function for PNG dataascii)rdecode)rrrrr/Ssr/) pyAvalonToolscCstj||tjdS)N)ZisQueryZbitFlags)rkZ GetAvalonFPZ avalonSSSBits)rrrrrr^srcCstj|ddS)Ni)ZfpSize)rZPatternFingerprint)rrrrrrbscCs|dk rt|d|_|S)zoPrecomputes fingerprints and stores results in molecule objects to accelerate substructure matching NF)rr)mrrrrMfs rMSmilescCs<|s||tj||<n||dd||<t|dS)aConverts the molecules contains in "smilesCol" to RDKit molecules and appends them to the dataframe "frame" using the specified column name. If desired, a fingerprint can be computed and stored with the molecule objects to accelerate substructure matching cSstt|S)N)rMr MolFromSmiles)smilesrrryz*AddMoleculeColumnToFrame..N)maprrnrQ)rgZ smilesColmolColrUrrrAddMoleculeColumnToFrameos rtimagecCs@|dkr dSyt||Wntk r:tdYnXdS)agAllows to change the rendering of the molecules between image and string representations. This serves two purposes: First it allows to avoid the generation of images if this is not desired and, secondly, it allows to enable image rendering for newly created dataframe that already contains molecules, without having to rerun the time-consuming AddMoleculeColumnToFrame. Note: this behaviour is, because some pandas methods, e.g. head() returns a new dataframe instance that uses the default pandas rendering (thus not drawing images for molecules) instead of the monkey-patched one. Nz.r+ZRowIDr@z{:f}0.)rCrDr-rErFrGrrextendrxr_rcZSetPropsiterrowsrr\SetProprvrwtyperyformatrstripwrite)rZoutrTrSZ propertiesZ allNumericrGrEwriterrYr>pZ cell_valuer%r)rZrWriteSDFsH          rcs:tdkrddlm}|a|jfdddd|<dS)zD Removes salts from mols in pandas DataFrame's ROMol column Nr) SaltRemovercst|S)N) _saltRemoverZStripMol)r)rsrrrprqz&RemoveSaltsFromFrame..r+)axis)rr`rapply)rgrsrr)rsrRemoveSaltsFromFrames rr(cCstj||d}|dkr`x>t||dd||DD]\}}|d|||q4W|n$x||D]}||qjW|dS)zs Saves smi file. SMILES are generated from column with RDKit molecules. Column with names is optional. )rAr(css|]}t|VqdS)N)rD)r$crrrr?sz&SaveSMILESFromFrame..r@N)rZ SmilesWriterziprrrG)rgoutFilersZNamesColrAwrlnrrrSaveSMILESFromFrames&  ri,i,c Csddl}t|j}||t|j}||}|}|d|ddd} x"|D]} | d| | | d7} qVWd} x2| D]$\} } t }t j | ||d}|j|dd|j| |dd || dd d |id} x|D]} t|| d kr| | | t| | dd n|dt|| ks>dt|| krr| | tjks^| | tjkr|| | | | n$dt|| kr|| | | | | d7} qW| d7} qW||dS)a Saves pandas DataFrame as a xlsx file with embedded images. It maps numpy data types to excel cell types: int, float -> number datetime -> datetime object -> string (limited to 32k character - xlsx limitations) Cells with compound images are a bit larger than images due to excel. Column width weirdness explained (from xlsxwriter docs): The width corresponds to the column width value that is specified in Excel. It is approximately equal to the length of a string in the default font of Calibri 11. Unfortunately, there is no way to specify "AutoFit" for a column in the Excel file format. This feature is only available at runtime from within Excel. rNzA:Ag@r+)r0ZPNG)r)ZheightrV image_dataobjecti}floatintZdatetime) xlsxwriterrr^rcrJrxZWorkbookZ add_worksheetZ set_columnZ write_stringr~r rZ MolToImageZsaveZset_rowZ insert_imagerDrvnaninfZ write_numberZwrite_datetimerG)rgrrsr0rreZ dataTypesZworkbookZ worksheetZc2rr_rYrZimgrrrSaveXlsxFromFrames>        $   rcKsR|r<||jjkr&dd|jD|d<ndd||D|d<tjt||f|S)z6 Draw grid image of mols in pandas DataFrame. cSsg|] }t|qSr)rD)r$rrrrr&(sz$FrameToGridImage..ZlegendscSsg|] }t|qSr)rD)r$rrrrr&*s)rBnamerZMolsToGridImager)rgcolumnZ legendsColkwargsrrrFrameToGridImage"s  r Murcko_SMILEScs4|rfdd}n fdd}|j|dd||<dS)z Adds column with SMILES of Murcko scaffolds to pandas DataFrame. Generic set to true results in SMILES of generic framework. csttt|S)N)rrKr ZMakeScaffoldGenericGetScaffoldForMol)r)rsrrfunc6szAddMurckoToFrame..funccstt|S)N)rrKr r)r)rsrrr;sr+)rN)r)rgrsZ MurckoColZGenericrr)rsrAddMurckoToFrame.s rcCs$t|}t|t|||S)zC Aligns mol (RDKit mol object) to scaffold (SMILES string) )rrnrrbZ$GenerateDepictionMatching2DStructure)r>ZscaffoldrrrAlignMolAs   rcs |jfdddd|<dS)z@ Aligns molecules in molCol to scaffolds in scaffoldCol cst||S)N)r)r)rs scaffoldColrrrpOrqz!AlignToScaffold..r+)rN)r)rgrsrr)rsrrAlignToScaffoldKsrcCsBy tWntk r YnXdtkr>tjjtd<ttj_dS)z5 Monkey patch an RDKit method of Chem.Mol and pandas zChem.Mol.__ge__N)r2Z patchPandasr3_originalSettingsrr\__ge__rrrrrInstallPandasToolsTs  rcCsBy tWntk r YnXdtkr6tdtj_tdS)z0 Unpatch an RDKit method of Chem.Mol and pandas zChem.Mol.__ge__N)r2Z unpatchPandasr3rrr\rclearrrrrUninstallPandasTools_s  rcCs4ddl}|j|j|j|d\}}|r0t|dS)Nr)Z optionflagsverbose)doctestZtestmodELLIPSISZNORMALIZE_WHITESPACEsysexit)rrZfailedrrrr _runDoctestsns  r__main__cCsXddl}y tj}Wntk r.tjj}YnX|d|dd}tdd|DS)z# Get the pandas version as a tuple rNz[^0-9,.]r|css|]}t|VqdS)N)r)r$Zvirrrr?sz$_getPandasVersion..)rerO __version__AttributeErrorversionsplittuple)rrfrrr_getPandasVersions rc@sDeZdZeedkpedkdddZeedkdddZdS)TestCaseNzpandas/xlsxwriter not installedcCs:ddl}ddlm}|j|jd}t|}t|ddS)Nr)RDConfigzNCI/first_200.props.sdfzfoo.xlsx)osrdkitrpathjoinZ RDDataDirr[r)selfrrZsdfFilergrrrtestGithub1507s  zTestCase.testGithub1507zpandas not installedcCs:tddgddgd}t|dd|t|jdd S) z& problem with update to pandas v1.2.0 ZethanolZfuranZCCOZc1ccoc1)rroromoleculeN)rOrPrtZ assertEquallenr)rrZrrrtestGithub3701s zTestCase.testGithub3701) __name__ __module__ __qualname__unittestZskipIfrrOrrrrrrrs r)rz4pandas installation >=0.19 not found, skipping tests)T)r5r6FTNFTT)FF)rmr6F)Nru)r6NNF)r6)r6r(F)r6r)r6N)r6rF)r6r)N)H__doc__base64rrZloggingZnumpyrvrrrr`rrrrZrdkit.Chem.Scaffoldsr r r rZrdkit.Chem.Draw.IPythonConsole ImportErrorior Zxml.domr Z getLoggerrrrrr,r.Z molJustifyrrr r(r2rLrr4ZpandasrOr[rhr/Z rdkit.AvalonrkrrMrtrQrrrrrrrrrrrrrrrrZrunTestsr3mainrrrr|s                 2 -     9  :