Merge pull request #2 from ResearchIT/master

Changes needed for server rollout
Pranavkhade · Jun 11, 2019 · 167f0f6 · 167f0f6
2 parents 129c005 + d4244cc
commit 167f0f6
Show file tree

Hide file tree

Showing 6 changed files with 147 additions and 65 deletions.
diff --git a/.gitignore b/.gitignore
@@ -0,0 +1 @@
+*.pyc
diff --git a/Dockerfile b/Dockerfile
@@ -0,0 +1,21 @@
+FROM centos:7
+
+# System dependencies
+RUN yum -y install git which python-virtualenv gcc gcc-c++
+
+# SSH deployment key
+RUN mkdir ~/.ssh
+COPY ppackman-key ~/.ssh/
+RUN chmod 400 ~/.ssh/ppackman-key
+
+# Setup virtualenv
+RUN virtualenv ppackman-env
+RUN source ppackman-env/bin/activate
+
+# Clone the source
+RUN git clone git@github.com:ResearchIT/PPACKMAN.git
+
+# Install python requirements
+WORKDIR PPACKMAN
+RUN pip install -r requirements.txt
+
diff --git a/Molecule.pyc b/Molecule.pyc
diff --git a/PACKMAN.py b/PACKMAN.py
@@ -11,6 +11,7 @@
 import os
 import sys
 import functools
+import urllib
 
 import networkx as nx
 
@@ -32,7 +33,7 @@
 '''
 
 
-def HingePredict(atoms,Alpha=float('Inf'),method='AlphaShape',GenerateKirchoff=False,filename='Output.pdb',MinimumHingeLength=5,nclusters=4):
+def HingePredict(atoms, outputfile, Alpha=float('Inf'),method='AlphaShape',GenerateKirchoff=False,filename='Output.pdb',MinimumHingeLength=5,nclusters=4):
     """
     """
     if(method=='AlphaShape'):
@@ -74,13 +75,13 @@ def GetStats(atoms,SelectedHingeResidues,filename='Output'):
 
             #Example\tGroup\tNumberofelements\tMin\tMax\tMean\tMode\tMedian\tStandardDeviation\n
 
-            sys.stdout.write('\nSTATISTICS\n\t\tN\tMin\tMax\tMean\tMode\tMedian\tSTDDev\n')
-            sys.stdout.write('Total   '+'\t'+str(len(all_atoms_bfactor))+'\t'+str(numpy.min(all_atoms_bfactor))+'\t'+str(numpy.max(all_atoms_bfactor))+'\t'+str(numpy.mean(all_atoms_bfactor))+'\t'+str(stats.mode(all_atoms_bfactor)[0][0])+'\t'+str(numpy.median(all_atoms_bfactor))+'\t'+str(numpy.std(all_atoms_bfactor))+'\n')
-            sys.stdout.write('Hinge   '+'\t'+str(len(hinge_atoms_bfactor))+'\t'+str(numpy.min(hinge_atoms_bfactor))+'\t'+str(numpy.max(hinge_atoms_bfactor))+'\t'+str(numpy.mean(hinge_atoms_bfactor))+'\t'+str(stats.mode(hinge_atoms_bfactor)[0][0])+'\t'+str(numpy.median(hinge_atoms_bfactor))+'\t'+str(numpy.std(hinge_atoms_bfactor))+'\n')
-            sys.stdout.write('NonHinge'+'\t'+str(len(non_hinge_atoms_bfactor))+'\t'+str(numpy.min(non_hinge_atoms_bfactor))+'\t'+str(numpy.max(non_hinge_atoms_bfactor))+'\t'+str(numpy.mean(non_hinge_atoms_bfactor))+'\t'+str(stats.mode(non_hinge_atoms_bfactor)[0][0])+'\t'+str(numpy.median(non_hinge_atoms_bfactor))+'\t'+str(numpy.std(non_hinge_atoms_bfactor))+'\n')
+            outputfile.write('\nSTATISTICS\n\t\tN\tMin\tMax\tMean\tMode\tMedian\tSTDDev\n')
+            outputfile.write('Total   '+'\t'+str(len(all_atoms_bfactor))+'\t'+str(numpy.min(all_atoms_bfactor))+'\t'+str(numpy.max(all_atoms_bfactor))+'\t'+str(numpy.mean(all_atoms_bfactor))+'\t'+str(stats.mode(all_atoms_bfactor)[0][0])+'\t'+str(numpy.median(all_atoms_bfactor))+'\t'+str(numpy.std(all_atoms_bfactor))+'\n')
+            outputfile.write('Hinge   '+'\t'+str(len(hinge_atoms_bfactor))+'\t'+str(numpy.min(hinge_atoms_bfactor))+'\t'+str(numpy.max(hinge_atoms_bfactor))+'\t'+str(numpy.mean(hinge_atoms_bfactor))+'\t'+str(stats.mode(hinge_atoms_bfactor)[0][0])+'\t'+str(numpy.median(hinge_atoms_bfactor))+'\t'+str(numpy.std(hinge_atoms_bfactor))+'\n')
+            outputfile.write('NonHinge'+'\t'+str(len(non_hinge_atoms_bfactor))+'\t'+str(numpy.min(non_hinge_atoms_bfactor))+'\t'+str(numpy.max(non_hinge_atoms_bfactor))+'\t'+str(numpy.mean(non_hinge_atoms_bfactor))+'\t'+str(stats.mode(non_hinge_atoms_bfactor)[0][0])+'\t'+str(numpy.median(non_hinge_atoms_bfactor))+'\t'+str(numpy.std(non_hinge_atoms_bfactor))+'\n')
 
             p_value = permutation_test(hinge_atoms_bfactor, non_hinge_atoms_bfactor,method='approximate',num_rounds=10000,seed=0)
-            sys.stdout.write('\np-value:\t'+str(p_value)+'\n')
+            outputfile.write('\np-value:\t'+str(p_value)+'\n')
             return p_value
 
 
@@ -166,12 +167,12 @@ def cross(a, b):
                 # This evaluates a * x3 + b * y3 + c * z3 which equals d
                 d = numpy.dot(cp, p3)
                 a,b,c,d=a/d,b/d,c/d,d/d
-                #print('The equation is {0}x + {1}y + {2}z = {3}'.format(a, b, c, d))
+                #outputfile.write('The equation is {0}x + {1}y + {2}z = {3}'.format(a, b, c, d))
 
                 #Plane in PyMol
-                sys.stdout.write('import plane\n')
-                sys.stdout.write('plane.make_plane_points(name=\'HingePlane'+ str(HingePlane) +'\', l1='+str(list(p4))+', l2='+str(list(p2))+', l3='+str(list(p3))+', center=False, makepseudo=False)'+'\n')
-                sys.stdout.write('set cgo_transparency, 0.35, HingePlane'+str(HingePlane)+'\n')
+                outputfile.write('import plane\n')
+                outputfile.write('plane.make_plane_points(name=\'HingePlane'+ str(HingePlane) +'\', l1='+str(list(p4))+', l2='+str(list(p2))+', l3='+str(list(p3))+', center=False, makepseudo=False)'+'\n')
+                outputfile.write('set cgo_transparency, 0.35, HingePlane'+str(HingePlane)+'\n')
                 return a,b,c,d
 
             a,b,c,d=FitPlane(hinge_points)
@@ -235,16 +236,16 @@ def cross(a, b):
                 PredictedHinges.append(SortedHingeResidues[HingeResiduesID.index(Hinge[0]):HingeResiduesID.index(Hinge[-1])+1])
 
         #Print part
-        sys.stdout.write('Filename= '+str(filename)+'\t| AlphaValue= '+str(Alpha)+'\t| MinimumHingeLength= '+str(MinimumHingeLength)+'\t| EccentricityClusters= '+str(nclusters)+ '\n')
-        sys.stdout.write('Hindge Residues(Predicted):\n')
+        outputfile.write('Filename= '+str(filename)+'\t| AlphaValue= '+str(Alpha)+'\t| MinimumHingeLength= '+str(MinimumHingeLength)+'\t| EccentricityClusters= '+str(nclusters)+ '\n')
+        outputfile.write('Hindge Residues(Predicted):\n')
         for numi,i in enumerate(PredictedHinges):
-            #sys.stdout.write('\nHinge #'+str(numi+1)+'\nResidues: '+','.join([j.get_name()+'-'+str(j.get_id()) for j in i])+'\n')
-            sys.stdout.write('\nHinge #'+str(numi+1)+'\nResidues: '+i[0].get_name()+'-'+str(i[0].get_id()) +' to '+i[-1].get_name()+'-'+str(i[-1].get_id())+'\n')
+            #outputfile.write('\nHinge #'+str(numi+1)+'\nResidues: '+','.join([j.get_name()+'-'+str(j.get_id()) for j in i])+'\n')
+            outputfile.write('\nHinge #'+str(numi+1)+'\nResidues: '+i[0].get_name()+'-'+str(i[0].get_id()) +' to '+i[-1].get_name()+'-'+str(i[-1].get_id())+'\n')
             GetStats(atoms,i,filename=filename)
-            #sys.stdout.write('\nPymol Terminal Commands for Visualizing:\ncolor blue, resi '+'+'.join(str(j.get_id()) for j in i)+'\n')
-            sys.stdout.write('\nPymol Terminal Commands for Visualizing:\ncolor blue, resi '+str(i[0].get_id())+':'+str(i[-1].get_id())+'\n')
+            #outputfile.write('\nPymol Terminal Commands for Visualizing:\ncolor blue, resi '+'+'.join(str(j.get_id()) for j in i)+'\n')
+            outputfile.write('\nPymol Terminal Commands for Visualizing:\ncolor blue, resi '+str(i[0].get_id())+':'+str(i[-1].get_id())+'\n')
             GetLeastSquarePlane(atoms,i,numi+1)
-            sys.stdout.write("#--------------------------------------------------#\n")
+            outputfile.write("#--------------------------------------------------#\n")
         #RotateHingeBonds(atoms,HingeResidues)
 
         if(GenerateKirchoff):
@@ -259,13 +260,10 @@ def cross(a, b):
 ##################################################################################################
 '''
 
-def WriteOBJ(atoms,faces,filename='Output.obj'):
+def WriteOBJ(atoms,faces, fh):
     """
     """
-    filename='AlphaShape-'+filename.replace('pdb','obj')
     NewIDs={i.get_id():numi+1 for numi,i in enumerate(atoms)}
-    open(filename,'w').write('')
-    fh=open(filename,'a')
     fh.write('mtllib master.mtl\ng\n')
     fh.write('usemtl atoms\n')
     for i in atoms:
@@ -411,15 +409,23 @@ def IO():
     """
     parser=argparse.ArgumentParser(description='PACKMAN: PACKing and Motion ANalysis. (https://github.com/Pranavkhade/PACKMAN)')
 
-    required = parser.add_mutually_exclusive_group(required=True)
-    required.add_argument('-pdbid','--pdbid',nargs=2,metavar='PDB_ID', type=str, help='(1) PDB ID of the input file (2) Location and Name by which you wish to save the downloaded file')
-    required.add_argument('-filename','--filename',nargs=1,metavar='FILENAME', type=str, help='Path and filename of the PDB file')
+    parser.add_argument('-pdbid','--pdbid', metavar='PDB_ID', type=str, help='If provided, the PBD with this ID will be downloaded and saved to FILENAME.')
+
+    parser.add_argument('alpha', metavar='AlphaValue', help='Recommended: 2.8 for closed; 4.5 for open form, Please refer to the paper for more details')
+    parser.add_argument('filename', metavar='FILENAME', help='Path and filename of the PDB file.')
 
-    parser.add_argument('alpha',metavar='AlphaValue',help='Recommended: Start from 2 and keep increasing the parameter value till the hinges become redundant compared to the previous alpha values (Typically around 5), Please refer to the paper for more details')
     parser.add_argument('--e_clusters',metavar='NumberOfEccentricityClusters',type=int,default=4,help='Recommended: 4, Please refer to the paper for more details')
     parser.add_argument('--minhnglen',metavar='MinimumHingeLength',type=int,default=5,help='Recommended: 5, Please refer to the paper for more details')
     parser.add_argument("--chain", help='Enter The Chain ID')
-    parser.add_argument('--generateobj',choices=['yes', 'no'], default='no', help='Select yes if you wish to generate the .obj file')
+    parser.add_argument('--generateobj', type=argparse.FileType('wb', 0), help='Path and filename to save the .obj file at. Ignored unless --chain is provided.')
+
+    # web server parameters
+    web_server_group = parser.add_argument_group('Web server parameters', 'Used by the web form')
+    web_server_group.add_argument('--outputfile', type=argparse.FileType('wb', 0), default=sys.stdout, help='Path and filename write output to')
+    web_server_group.add_argument('--logfile', type=argparse.FileType('wb', 0), default=sys.stderr, help='Path and filename write log messages to')
+    web_server_group.add_argument('--callbackurl', type=str, help='Optional callback url if this script was called from Drupal.')
+    web_server_group.add_argument('--nodeid', type=int, help='Optional node id if this script was called from Drupal.')
+
     args=parser.parse_args()
     return args
 
@@ -434,36 +440,50 @@ def main():
     """
     args=IO()
 
-
-    #Mutually Exclusive Opetion to load the file
     if(args.pdbid is not None):
-        Molecule.DownloadPDB(args.pdbid[0],args.pdbid[1])
-        mol=Molecule.LoadPDB(args.pdbid[1])
-    elif(args.filename is not None):
-        mol=Molecule.LoadPDB(args.filename[0])
-
-    if(args.chain):
-        Backbone=[item for sublist in mol[0][args.chain].get_backbone() for item in sublist]
-        SelectedTesselations,PredictedHinges=HingePredict(Backbone,Alpha=float(args.alpha),filename=args.filename,nclusters=args.e_clusters,MinimumHingeLength=args.minhnglen)
-        if(args.generateobj=='yes'):WriteOBJ(Backbone,SelectedTesselations,'CHAIN-'+args.chain+'-'+args.filename[0])
-    else:
-        for i in mol[0].get_chains():
-            Backbone=[item for sublist in mol[0].get_backbone() for item in sublist]
-            SelectedTesselations,PredictedHinges=HingePredict(Backbone,Alpha=float(args.alpha),filename=args.filename,nclusters=args.e_clusters,MinimumHingeLength=args.minhnglen)
-            if(args.generateobj=='yes'):WriteOBJ(Backbone,SelectedTesselations,'CHAIN-'+i.get_id()+'-'+args.filename)
+        Molecule.DownloadPDB(args.pdbid, args.filename)
+
+    mol = Molecule.LoadPDB(args.filename)
+
+    try:
+
+        if(args.chain):
+            Backbone = [item for sublist in mol[0][args.chain].get_backbone() for item in sublist]
+            SelectedTesselations, PredictedHinges = HingePredict(Backbone, args.outputfile, Alpha=float(args.alpha), filename=args.filename,nclusters=args.e_clusters,MinimumHingeLength=args.minhnglen)
+            if(args.generateobj is not None):
+                WriteOBJ(Backbone, SelectedTesselations, args.generateobj)
+        else:
+            for i in mol[0].get_chains():
+                Backbone = [item for sublist in mol[0].get_backbone() for item in sublist]
+                SelectedTesselations, PredictedHinges = HingePredict(Backbone, args.outputfile, Alpha=float(args.alpha), filename=args.filename,nclusters=args.e_clusters,MinimumHingeLength=args.minhnglen)
+
+        if args.nodeid is not None:
+            urllib.urlopen(args.callbackurl + '/' + str(args.nodeid) + "/0")
+
+    except Exception:
+
+        if args.nodeid is not None:
+            urllib.urlopen(args.callbackurl + '/' + str(args.nodeid) + "/1")
 
-    print_footnotes()
-    sys.stdout.flush()
+    finally:
+
+        print_footnotes(args.outputfile)
+        args.outputfile.flush()
+        args.logfile.flush()
+
+        if(args.generateobj is not None):
+            args.generateobj.flush()
+
     return True
 
 '''
 ##################################################################################################
 #                                             End                                                #
 ##################################################################################################
 '''
-def print_footnotes():
-    sys.stdout.write('Footnotes:\n\nSTATISTICS Section Legend:\nN: Number of residues\nMin: Minimum B-factor value\nMax: Maximum B-factor value\nMean: Mean B-factor value\nMode: Mode B-factor value\nMedian: Median B-factor value\nSTDDev: Standard Deviation of the B-factor values\n')
-    sys.stdout.write('\nRMSF Plane Visualization:\nDownload plane.py file from https://pymolwiki.org/index.php/Plane_Wizard and place it in the PyMol working directory if the RMSF plane needs to be visualized.\n')
+def print_footnotes(outputfile):
+    outputfile.write('Footnotes:\n\nSTATISTICS Section Legend:\nN: Number of residues\nMin: Minimum B-factor value\nMax: Maximum B-factor value\nMean: Mean B-factor value\nMode: Mode B-factor value\nMedian: Median B-factor value\nSTDDev: Standard Deviation of the B-factor values\n')
+    outputfile.write('\nRMSF Plane Visualization:\nDownload plane.py file from https://pymolwiki.org/index.php/Plane_Wizard and place it in the PyMol working directory if the RMSF plane needs to be visualized.\n')
     return True
 
 

diff --git a/README.md b/README.md
@@ -47,42 +47,55 @@ pip install git+git://github.com/Pranavkhade/PACKMAN
 Description
 
 ```
-usage: PACKMAN.py [-h] (-pdbid PDB_ID PDB_ID | -filename FILENAME)
+usage: PACKMAN.py [-h] [-pdbid PDB_ID]
                   [--e_clusters NumberOfEccentricityClusters]
                   [--minhnglen MinimumHingeLength] [--chain CHAIN]
-                  [--generateobj {yes,no}]
-                  AlphaValue
+                  [--generateobj GENERATEOBJ] [--outputfile OUTPUTFILE]
+                  [--logfile LOGFILE] [--callbackurl CALLBACKURL]
+                  [--nodeid NODEID]
+                  AlphaValue FILENAME
 
 PACKMAN: PACKing and Motion ANalysis. (https://github.com/Pranavkhade/PACKMAN)
 
 positional arguments:
-  AlphaValue            Recommended: Start from 2 and keep increasing the
-                        parameter value till the hinges become redundant
-                        compared to the previous alpha values (Typically
-                        around 5), Please refer to the paper for more details
+  AlphaValue            Recommended: 2.8 for closed; 4.5 for open form, Please
+                        refer to the paper for more details
+  FILENAME              Path and filename of the PDB file.
 
 optional arguments:
   -h, --help            show this help message and exit
-  -pdbid PDB_ID PDB_ID, --pdbid PDB_ID PDB_ID
-                        (1) PDB ID of the input file (2) Location and Name by
-                        which you wish to save the downloaded file
-  -filename FILENAME, --filename FILENAME
-                        Path and filename of the PDB file
+  -pdbid PDB_ID, --pdbid PDB_ID
+                        If provided, the PBD with this ID will be downloaded
+                        and saved to FILENAME.
   --e_clusters NumberOfEccentricityClusters
                         Recommended: 4, Please refer to the paper for more
                         details
   --minhnglen MinimumHingeLength
                         Recommended: 5, Please refer to the paper for more
                         details
   --chain CHAIN         Enter The Chain ID
-  --generateobj {yes,no}
-                        Select yes if you wish to generate the .obj file
+  --generateobj GENERATEOBJ
+                        Path and filename to save the .obj file at. Ignored
+                        unless --chain is provided.
+
+Web server parameters:
+  Used by the web form
+
+  --outputfile OUTPUTFILE
+                        Path and filename write output to
+  --logfile LOGFILE     Path and filename write log messages to
+  --callbackurl CALLBACKURL
+                        Optional callback url if this script was called from
+                        Drupal.
+  --nodeid NODEID       Optional node id if this script was called from
+                        Drupal.
+
 ```
 ### Examples
 
-1. `python PACKMAN.py 2.8 --pdbid 1prw 1prw.pdb`
-First parameter 2.8 is the alpha value parameter which can be changed. for --pdbid, the first parameter is the PDB ID the user submits to download a corresponding PDB file and second is the name of the file you wish to save the downloaded PDB file.
+1. `python PPACKMAN.py --pdbid 1prw 2.8 1prw.pdb`
+for --pdbid, the parameter is the PDB ID the user submits to download a corresponding PDB file. First positional parameter 2.8 is the alpha value parameter which can be changed. and second is the name of the file you wish to save the downloaded PDB file.
 
 
-2. `python PACKMAN.py 2.8 --filename 1prw.pdb`
-First parameter 2.8 is the alpha value parameter which can be changed. for --filename, the parameter is location and name of the PDB file.
+2. `python PPACKMAN.py 2.8 1prw.pdb`
+First parameter 2.8 is the alpha value parameter which can be changed. second parameter is the parameter is location and name of the PDB file.
diff --git a/requirements.txt b/requirements.txt
@@ -0,0 +1,27 @@
+astroid==1.6.6
+backports.functools-lru-cache==1.5
+configparser==3.7.4
+cycler==0.10.0
+decorator==4.4.0
+enum34==1.1.6
+futures==3.2.0
+isort==4.3.18
+kiwisolver==1.1.0
+lazy-object-proxy==1.3.1
+matplotlib==2.2.4
+mccabe==0.6.1
+mlxtend==0.15.0.0
+networkx==2.2
+numpy==1.16.3
+pandas==0.24.2
+pylint==1.9.4
+pyparsing==2.4.0
+python-dateutil==2.8.0
+pytz==2019.1
+scikit-learn==0.20.3
+scipy==1.2.1
+singledispatch==3.4.0.3
+six==1.12.0
+sklearn==0.0
+subprocess32==3.5.3
+wrapt==1.11.1