Source Code for Module Bio.SCOP.Hie

  1  # Copyright 2001 by Gavin E. Crooks.  All rights reserved. 
  2  # This code is part of the Biopython distribution and governed by its 
  3  # license.  Please see the LICENSE file that should have been included 
  4  # as part of this package. 
  5   
  6   
  7  """Handle the SCOP HIErarchy files, which describe the SCOP hierarchy in 
  8  terms of SCOP unique identifiers (sunid). 
  9   
 10  The file format is described in the scop 
 11  "release notes.":http://scop.berkeley.edu/release-notes-1.55.html 
 12  The latest HIE file can be found 
 13  "elsewhere at SCOP.":http://scop.mrc-lmb.cam.ac.uk/scop/parse/ 
 14   
 15  "Release 1.55":http://scop.berkeley.edu/parse/dir.hie.scop.txt_1.55 (July 2001) 
 16  """ 
 17   
 18   
 19 -class Record(object): 
 20      """Holds information for one node in the SCOP hierarchy. 
 21   
 22      sunid      -- SCOP unique identifiers of this node 
 23   
 24      parent     -- Parents sunid 
 25   
 26      children   -- Sequence of childrens sunids 
 27      """ 
 28 -    def __init__(self, line=None): 
 29          self.sunid = '' 
 30          self.parent = '' 
 31          self.children = [] 
 32          if line: 
 33              self._process(line) 
 34   
 35 -    def _process(self, line): 
 36          """Parses HIE records. 
 37   
 38          Records consist of 3 tab deliminated fields; node's sunid, 
 39          parent's sunid, and a list of children's sunids. 
 40          """ 
 41          #For example :: 
 42          # 
 43          #0       -       46456,48724,51349,53931,56572,56835,56992,57942 
 44          #21953   49268   - 
 45          #49267   49266   49268,49269 
 46          line = line.rstrip()        # no trailing whitespace 
 47          columns = line.split('\t')   # separate the tab-delineated cols 
 48          if len(columns) != 3: 
 49              raise ValueError("I don't understand the format of %s" % line) 
 50   
 51          sunid, parent, children = columns 
 52   
 53          if sunid =='-': 
 54              self.sunid = '' 
 55          else: 
 56              self.sunid = int(sunid) 
 57   
 58          if parent=='-': 
 59              self.parent = '' 
 60          else: 
 61              self.parent = int(parent) 
 62   
 63          if children=='-': 
 64              self.children = () 
 65          else: 
 66              children = children.split(',') 
 67              self.children = map(int, children) 
 68   
 69 -    def __str__(self): 
 70          s = [] 
 71          s.append(str(self.sunid)) 
 72   
 73          if self.parent: 
 74              s.append(str(self.parent)) 
 75          else: 
 76              if self.sunid != 0: 
 77                  s.append('0') 
 78              else: 
 79                  s.append('-') 
 80   
 81          if self.children: 
 82              child_str = map(str, self.children) 
 83              s.append(",".join(child_str)) 
 84          else: 
 85              s.append('-') 
 86   
 87          return "\t".join(s) + "\n" 
 88   
 89   
 90 -def parse(handle): 
 91      """Iterates over a HIE file, returning a Hie record for each line 
 92      in the file. 
 93   
 94      Arguments: 
 95   
 96          handle -- file-like object. 
 97      """ 
 98      for line in handle: 
 99          if line.startswith('#'): 
100              continue 
101          yield Record(line) 
102