Source Code for Package Bio.Medline

  1  # Copyright 1999 by Jeffrey Chang.  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  This module provides code to work with Medline. 
  8   
  9  Classes: 
 10  Record           A dictionary holding Medline data. 
 11   
 12  Functions: 
 13  read             Reads one Medline record 
 14  parse            Allows you to iterate over a bunch of Medline records 
 15  """ 
 16   
 17   
 18 -class Record(dict): 
 19      """A dictionary holding information from a Medline record. 
 20      All data are stored under the mnemonic appearing in the Medline 
 21      file. These mnemonics have the following interpretations: 
 22   
 23      Mnemonic  Description 
 24      AB        Abstract 
 25      CI        Copyright Information 
 26      AD        Affiliation 
 27      IRAD      Investigator Affiliation 
 28      AID       Article Identifier 
 29      AU        Author 
 30      FAU       Full Author 
 31      CN        Corporate Author 
 32      DCOM      Date Completed 
 33      DA        Date Created 
 34      LR        Date Last Revised 
 35      DEP       Date of Electronic Publication 
 36      DP        Date of Publication 
 37      EDAT      Entrez Date 
 38      GS        Gene Symbol 
 39      GN        General Note 
 40      GR        Grant Number 
 41      IR        Investigator Name 
 42      FIR       Full Investigator Name 
 43      IS        ISSN 
 44      IP        Issue 
 45      TA        Journal Title Abbreviation 
 46      JT        Journal Title 
 47      LA        Language 
 48      LID       Location Identifier 
 49      MID       Manuscript Identifier 
 50      MHDA      MeSH Date 
 51      MH        MeSH Terms 
 52      JID       NLM Unique ID 
 53      RF        Number of References 
 54      OAB       Other Abstract 
 55      OCI       Other Copyright Information 
 56      OID       Other ID 
 57      OT        Other Term 
 58      OTO       Other Term Owner 
 59      OWN       Owner 
 60      PG        Pagination 
 61      PS        Personal Name as Subject 
 62      FPS       Full Personal Name as Subject 
 63      PL        Place of Publication 
 64      PHST      Publication History Status 
 65      PST       Publication Status 
 66      PT        Publication Type 
 67      PUBM      Publishing Model 
 68      PMC       PubMed Central Identifier 
 69      PMID      PubMed Unique Identifier 
 70      RN        Registry Number/EC Number 
 71      NM        Substance Name 
 72      SI        Secondary Source ID 
 73      SO        Source 
 74      SFM       Space Flight Mission 
 75      STAT      Status 
 76      SB        Subset 
 77      TI        Title 
 78      TT        Transliterated Title 
 79      VI        Volume 
 80      CON       Comment on 
 81      CIN       Comment in 
 82      EIN       Erratum in 
 83      EFR       Erratum for 
 84      CRI       Corrected and Republished in 
 85      CRF       Corrected and Republished from 
 86      PRIN      Partial retraction in 
 87      PROF      Partial retraction of 
 88      RPI       Republished in 
 89      RPF       Republished from 
 90      RIN       Retraction in 
 91      ROF       Retraction of 
 92      UIN       Update in 
 93      UOF       Update of 
 94      SPIN      Summary for patients in 
 95      ORI       Original report in 
 96      """ 
 97   
 98   
 99 -def parse(handle): 
100      """Read Medline records one by one from the handle. 
101   
102      The handle is either is a Medline file, a file-like object, or a list 
103      of lines describing one or more Medline records. 
104   
105      Typical usage: 
106   
107          from Bio import Medline 
108          handle = open("mymedlinefile") 
109          records = Medline.parse(handle) 
110          for record in record: 
111              print record['TI'] 
112   
113      """ 
114      # These keys point to string values 
115      textkeys = ("ID", "PMID", "SO", "RF", "NI", "JC", "TA", "IS", "CY", "TT", 
116                  "CA", "IP", "VI", "DP", "YR", "PG", "LID", "DA", "LR", "OWN", 
117                  "STAT", "DCOM", "PUBM", "DEP", "PL", "JID", "SB", "PMC", 
118                  "EDAT", "MHDA", "PST", "AB", "AD", "EA", "TI", "JT") 
119      handle = iter(handle) 
120      # First skip blank lines 
121      for line in handle: 
122          line = line.rstrip() 
123          if line: 
124              break 
125      else: 
126          return 
127      record = Record() 
128      finished = False 
129      while not finished: 
130          if line[:6] == "      ":  # continuation line 
131              record[key].append(line[6:]) 
132          elif line: 
133              key = line[:4].rstrip() 
134              if not key in record: 
135                  record[key] = [] 
136              record[key].append(line[6:]) 
137          try: 
138              line = handle.next() 
139          except StopIteration: 
140              finished = True 
141          else: 
142              line = line.rstrip() 
143              if line: 
144                  continue 
145          # Join each list of strings into one string. 
146          for key in textkeys: 
147              if key in record: 
148                  record[key] = " ".join(record[key]) 
149          if record: 
150              yield record 
151          record = Record() 
152   
153   
154 -def read(handle): 
155      """Read a single Medline record from the handle. 
156   
157      The handle is either is a Medline file, a file-like object, or a list 
158      of lines describing a Medline record. 
159   
160      Typical usage: 
161   
162          from Bio import Medline 
163          handle = open("mymedlinefile") 
164          record = Medline.read(handle) 
165          print record['TI'] 
166   
167      """ 
168      records = parse(handle) 
169      return records.next() 
170