Merge pull request #8 from dbmi-bgm/utf8_encode

Utf8 encode

README.md

@@ -11,7 +11,7 @@ For more details, see granite [*documentation*](https://granite-suite.readthedoc
 ## Availability and requirements
 A ready-to-use docker image is available to download.
 
-    docker pull b3rse/granite:v0.1.9
+    docker pull b3rse/granite:v0.1.10
 
 To run locally, install the following libraries:
 

docs/conf.py

@@ -22,7 +22,7 @@
 author = 'Michele Berselli'
 
 # The full version, including alpha/beta/rc tags
-release = '0.1.9'
+release = '0.1.10'
 
 
 # -- General configuration ---------------------------------------------------

docs/install.md

@@ -2,7 +2,7 @@
 
 A ready-to-use docker image is available to download.
 
-    docker pull b3rse/granite:v0.1.9
+    docker pull b3rse/granite:v0.1.10
 
 To run locally, Python 3.6+ is required together with the following libraries:
 

granite/_version.py

@@ -1,4 +1,4 @@
 """Version information."""
 
 # The following line *must* be the last in the module, exactly as formatted:
-__version__ = "0.1.9"
+__version__ = "0.1.10"

granite/blackList.py

@@ -57,7 +57,7 @@ def main(args):
     #end if
 
     # Buffers
-    fo = open(args['outputfile'], 'w')
+    fo = open(args['outputfile'], 'w', encoding='utf-8')
 
     # Loading big if specified
     if is_bigfile: big_dict = load_big(args['bigfile'])

granite/cleanVCF.py

@@ -52,7 +52,7 @@ def main(args):
     is_verbose = True if args['verbose'] else False
 
     # Buffers
-    fo = open(args['outputfile'], 'w')
+    fo = open(args['outputfile'], 'w', encoding='utf-8')
 
     # Creating Vcf object
     vcf_obj = vcf_parser.Vcf(args['inputfile'])

granite/comHet.py

@@ -542,7 +542,7 @@ def main(args, test=False):
     is_verbose = True if args['verbose'] else False
 
     # Buffers
-    fo = open(args['outputfile'], 'w')
+    fo = open(args['outputfile'], 'w', encoding='utf-8')
 
     # Creating Vcf object
     vcf_obj = vcf_parser.Vcf(args['inputfile'])

granite/geneList.py

@@ -47,7 +47,7 @@ def main(args):
     #end with
 
     # Buffers
-    fo = open(args['outputfile'], 'w')
+    fo = open(args['outputfile'], 'w', encoding='utf-8')
 
     # Creating Vcf object
     vcf_obj = vcf_parser.Vcf(args['inputfile'])

granite/lib/pedigree_parser.py

@@ -16,7 +16,7 @@
 #
 #################################################################
 import sys, os
-
+import json
 
 #################################################################
 #
@@ -118,12 +118,12 @@ def has_parents(self):
 
     #end class Member
 
-    def __init__(self, pedigree):
+    def __init__(self, pedigree, format='json'):
         ''' initialize Pedigree object,
         pedigree information must be provided as json '''
         self.members = {} # dictionary of Member objects by name
         self.samples = {} # dictionary to map sample to name
-        self.parse_pedigree(pedigree)
+        self.parse_pedigree(pedigree, format)
     #end def __init__
 
     def add_member(self, member):
@@ -217,15 +217,36 @@ def get_family(self, sample):
         return family
     #end def get_family
 
-    def parse_pedigree(self, pedigree):
-        ''' read pedigree information to build Pedigree object,
-        pedigree information must be provided as json '''
+    def read_json(self, pedigree):
+        ''' read pedigree as json '''
+        # Reading pedigree
+        if os.path.isfile(pedigree):
+            with open(pedigree) as fi:
+                pedigree_json = json.load(fi)
+            #end with
+        else:
+            try: pedigree_json = json.loads(pedigree)
+            except Exception:
+                raise ValueError('\nERROR in parsing arguments, pedigree must be either a json file or a string representing a json\n')
+            #end try
+        #end if
+        return pedigree_json
+    #end def read_json
+
+    def parse_pedigree(self, pedigree, format):
+        ''' use pedigree information to build Pedigree object,
+        pedigree information must be provided in specified format '''
+        if format == 'json':
+            pedigree_json = self.read_json(pedigree)
+        else:
+            raise ValueError('\nERROR in parsing pedigree, specified pedigree format is not supported\n')
+        #end if
         # Creating Member objects
-        for member in pedigree:
+        for member in pedigree_json:
             self.add_member(member)
         #end for
         # Adding relations
-        for member in pedigree:
+        for member in pedigree_json:
             name = member['individual']
             if len(member['parents']) > 2:
                 raise ValueError('\nERROR in pedigree structure, {0} has more than two parents\n'

granite/lib/vcf_parser.py

@@ -1,4 +1,5 @@
 #!/usr/bin/env python
+# -*- coding: utf-8 -*-
 
 #################################################################
 #
@@ -311,14 +312,15 @@ def get_genotype_value(self, ID_genotype, tag_to_get, sep=':'):
     def read_vcf(inputfile):
         ''' read vcf file, gzipped or ungzipped,
         return a generator '''
-        if inputfile.endswith('.gz'):
+        if inputfile.endswith('.gz') or \
+           inputfile.endswith('.bgz'):
             with gzip.open(inputfile, 'rb') as fz:
                 for byteline in fz:
-                    yield byteline.decode()
+                    yield byteline.decode('utf-8')
                 #end for
             #end with
         else:
-            with open(inputfile) as fi:
+            with open(inputfile, encoding='utf-8') as fi:
                 for line in fi:
                     yield line
                 #end for

granite/novoCaller.py

@@ -819,7 +819,7 @@ def main(args, test=False):
     is_verbose = True if args['verbose'] else False
 
     # Buffers
-    fo = open(args['outputfile'], 'w')
+    fo = open(args['outputfile'], 'w', encoding='utf-8')
 
     # Creating Vcf object
     vcf_obj = vcf_parser.Vcf(args['inputfile'])

granite/qcVCF.py

@@ -298,20 +298,8 @@ def main(args):
     # Get list of sample IDs to use
     ID_list = args['samples'] # list of sample IDs
 
-    # Loading pedigree
-    if os.path.isfile(args['pedigree']):
-        with open(args['pedigree']) as fi:
-            pedigree = json.load(fi)
-        #end with
-    else:
-        try: pedigree = json.loads(args['pedigree'])
-        except Exception:
-            sys.exit('\nERROR in parsing arguments: pedigree must be either a json file or a string representing a json\n')
-        #end try
-    #end if
-
     # Creating Pedigree object
-    pedigree_obj = pedigree_parser.Pedigree(pedigree)
+    pedigree_obj = pedigree_parser.Pedigree(args['pedigree'])
 
     # Initializing stat_dict
     for ID in ID_list:

granite/toPED.py

@@ -35,20 +35,8 @@ def main(args):
     encode_gender = {'M': 1, 'F': 2}
     familyID = args['family'] if args['family'] else 'FAM'
 
-    # Loading pedigree
-    if os.path.isfile(args['pedigree']):
-        with open(args['pedigree']) as fi:
-            pedigree = json.load(fi)
-        #end with
-    else:
-        try: pedigree = json.loads(args['pedigree'])
-        except Exception:
-            sys.exit('\nERROR in parsing arguments: pedigree must be either a json file or a string representing a json\n')
-        #end try
-    #end if
-
     # Creating Pedigree object
-    pedigree_obj = pedigree_parser.Pedigree(pedigree)
+    pedigree_obj = pedigree_parser.Pedigree(args['pedigree'])
 
     # Buffers
     fo = open(args['outputfile'], 'w')

granite/validateVCF.py

@@ -715,23 +715,6 @@ def main(args):
     # Creating Vcf object
     vcf_obj = vcf_parser.Vcf(args['inputfile'])
 
-    # Get pedigree / pedigrees information
-    pedigree_list = []
-    for pedigree in args['pedigree']:
-        # Loading pedigree
-        if os.path.isfile(pedigree):
-            with open(pedigree) as fi:
-                pedigree_list.append(json.load(fi))
-            #end with
-        else:
-            try: pedigree_list.append(json.loads(pedigree))
-            except Exception:
-                sys.exit('\nERROR in parsing arguments: {0} must be either a json file or a string representing a json\n'
-                        .format(pedigree))
-            #end try
-        #end if
-    #end for
-
     # Check novoPP
     if sample_novo:
         try: novotag, _ = vcf_obj.header.check_tag_definition('novoPP')
@@ -742,7 +725,7 @@ def main(args):
 
     # Creating Pedigree object / objects
     pedigree_obj_list = []
-    for pedigree in pedigree_list:
+    for pedigree in args['pedigree']:
         pedigree_obj_list.append(pedigree_parser.Pedigree(pedigree))
     #end for
 

granite/whiteList.py

@@ -55,7 +55,7 @@ def main(args):
     is_verbose = True if args['verbose'] else False
 
     # Buffers
-    fo = open(args['outputfile'], 'w')
+    fo = open(args['outputfile'], 'w', encoding='utf-8')
 
     # Creating Vcf object
     vcf_obj = vcf_parser.Vcf(args['inputfile'])