Add embl reader

src/obitools/format/sequence/embl.py

@@ -0,0 +1,2 @@
+from obitools.seqdb.embl.parser import emblIterator,emblParser
+

src/obitools/obischemas/kb/__init__.py

@@ -0,0 +1,55 @@
+"""
+    kb package is devoted to manage access to postgresql database from python
+    script
+"""
+
+
+class Connection(object):
+
+    def __init__(self):
+        raise RuntimeError('pyROM.KB.Connection is an abstract class')
+
+    def cursor(self):
+        raise RuntimeError('pyROM.KB.Connection.cursor is an abstract function')
+
+    def commit(self):
+        raise RuntimeError('pyROM.KB.Connection.commit is an abstract function')
+
+    def rollback(self):
+        raise RuntimeError('pyROM.KB.Connection.rollback is an abstract function')
+        
+    def __call__(self,query):
+        return self.cursor().execute(query)
+    
+
+class Cursor(object):
+
+    def __init__(self,db):
+        raise RuntimeError('pyROM.KB.Cursor is an abstract class')
+
+    def execute(self,query):
+        raise RuntimeError('pyROM.KB.Cursor.execute is an abstract function')
+        
+    __call__=execute
+    
+    
+_current_connection = None  # Static variable used to store connection to KB
+
+def getConnection(*args,**kargs):
+    """
+        return a connection to the database.
+        When call from database backend no argument are needed.
+        All connection returned by this function 
+    """
+    global _current_connection
+    
+    if _current_connection==None or args or kargs :
+        try:
+            from obischemas.kb import backend
+            _current_connection = backend.Connection()
+        except ImportError:
+            from obischemas.kb import extern
+            _current_connection = extern.Connection(*args,**kargs)
+    return _current_connection
+    
+ 

src/obitools/obischemas/kb/extern.py

@@ -0,0 +1,78 @@
+"""
+Module : KB.extern
+Author : Eric Coissac
+Date   : 03/05/2004
+
+Module wrapping psycopg interface module to allow connection
+to a postgresql databases with the same interface from
+backend and external script.
+
+This module define a class usable from external script 
+"""
+
+
+import psycopg2
+import sys
+from obischemas import kb
+
+class Connection(kb.Connection):
+
+    def __init__(self,*connectParam,**kconnectParam):
+        if connectParam:
+            self.connectParam=={'dsn':connectParam}
+        else:
+            self.connectParam=kconnectParam 
+        print self.connectParam
+        self.db = psycopg2.connect(**(self.connectParam))
+
+    def restart(self):
+	ok=1
+	while (ok and ok < 1000):
+	  try:	
+	    self.db = psycopg2.connect(**self.connectParam)
+	  except:
+            ok+=1
+	  else:
+            ok=0
+
+	    
+    def cursor(self):
+        curs = Cursor(self.db)
+        if hasattr(self,'autocommit') and self.autocommit:
+            curs.autocommit = self.autocommit
+        return curs
+    
+    def commit(self):
+        self.db.commit()
+
+    def rollback(self):
+        if hasattr(self,'db'):
+            self.db.rollback()
+
+    def __del__(self):
+        if hasattr(self,'db'):
+            self.rollback()
+        
+class Cursor(kb.Cursor):
+
+    def __init__(self,db):
+	self.db   = db
+        self.curs = db.cursor()
+        
+    def execute(self,query):
+        try:
+            self.curs.execute(query)
+            if hasattr(self,'autocommit') and self.autocommit:
+                self.db.commit()
+        except psycopg2.ProgrammingError,e:
+            print >>sys.stderr,"===> %s" % query
+            raise e
+        except psycopg2.IntegrityError,e:
+            print >>sys.stderr,"---> %s" % query
+            raise e
+        try:
+           label = [x[0] for x in self.curs.description]
+           return [dict(map(None,label,y))
+                   for y in self.curs.fetchall()]
+        except TypeError:
+           return []

src/obitools/seqdb/embl/__init__.py

@@ -0,0 +1,13 @@
+from obitools.seqdb import AnnotatedNucSequence, AnnotatedAASequence
+from obitools.location import locationGenerator,extractExternalRefs
+
+
+
+class EmblSequence(AnnotatedNucSequence):
+    '''
+    Class used to represent a nucleic sequence issued from EMBL.
+    '''
+    
+    
+    
+

src/obitools/seqdb/embl/parser.py

@@ -0,0 +1,50 @@
+import re
+import sys
+
+from obitools.seqdb import embl
+from obitools.seqdb import nucEntryIterator
+
+_featureMatcher = re.compile('(^FT  .*\n)+', re.M)
+_cleanFT       = re.compile('^FT',re.M)
+
+_headerMatcher = re.compile('^ID.+(?=\nFH  )', re.DOTALL)
+_seqMatcher    = re.compile('(^    ).+(?=//\n)', re.DOTALL + re.M)
+_cleanSeq      = re.compile('[ \n0-9]+')
+_acMatcher     = re.compile('(?<=^AC   ).+',re.M)
+_deMatcher     = re.compile('(^DE   .+\n)+',re.M)
+_cleanDe       = re.compile('(^|\n)DE +')
+
+def __emblparser(text):
+    try:
+        header = _headerMatcher.search(text).group()
+
+        ft     = _featureMatcher.search(text).group()
+        ft     = _cleanFT.sub('  ',ft)
+        
+        seq    = _seqMatcher.search(text).group()
+        seq    = _cleanSeq.sub('',seq).upper()
+        
+        acs    = _acMatcher.search(text).group()
+        acs    = acs.split()
+        ac     = acs[0]
+        acs    = acs[1:]
+        
+        de     = _deMatcher.search(header).group()
+        de     = _cleanDe.sub(' ',de).strip().strip('.')
+    except AttributeError,e:
+        print >>sys.stderr,'======================================================='
+        print >>sys.stderr,text
+        print >>sys.stderr,'======================================================='
+        raise e
+    
+    return (ac,seq,de,header,ft,acs)
+
+def emblParser(text):
+    return embl.EmblSequence(*__emblparser(text))
+    
+    
+def emblIterator(file):
+    for e in nucEntryIterator(file):
+        yield emblParser(e)
+    
+