bin/extract-tooltip.py

   1 #!/usr/bin/env python
   2 import sys
   3 import os
   4 import re
   5 import urlparse
   6
   7 def usage():
   8     message = """ usage: {program} inDir outDir
   9 inDir: directory containing .ht files
  10 outDir: target for the new files"""
  11     print(message.format(program = os.path.basename(sys.argv[0])))
  12
  13 def parseFile(filename):
  14     file = open(filename, "r")
  15     data = file.readlines()
  16     data = [line.rstrip('\n') for line in data]
  17
  18     pairs = {}
  19     regEx = re.compile("^(\S+)\s(\S+)\s(\S+)\s((?:\s*\S*)+)$")
  20     old_line = None
  21     for line in data:
  22         if len(line) > 0:
  23             if(old_line != None):
  24                 print filename
  25                 #print("failed to parse line")
  26                 #print(old_line)
  27                 line = old_line + line
  28                 print line
  29                 old_line = None
  30             split_line = regEx.split(line)
  31             #print(split_line)
  32             #print(urlparse.unquote(split_line[2]))
  33             #print(split_line[4])
  34             if(old_line == None and split_line[4] == "" and split_line[3] != "0"):
  35                 print(line)
  36                 print(split_line)
  37                 old_line = line
  38             else:
  39                 pairs[urlparse.unquote(split_line[2])] = split_line[4]
  40                 assert(len(split_line) == 6)
  41     #print data
  42     #print(pairs)
  43     return pairs
  44
  45 def parseFiles(dir):
  46     strings = []
  47     for files in os.listdir(dir):
  48         if files.endswith(".ht"):
  49             string = parseFile(os.path.join(dir,files))
  50             print(files)
  51             #print string
  52             strings.append([files, string])
  53     return strings
  54
  55 def extractSharedEntries(strings):
  56     first_dict = strings[0][1]
  57     shared_dict = {}
  58     #print(first_dict)
  59     for key, value in first_dict.iteritems():
  60         # check that the entry in the same in all dics
  61         is_in_all_dicts = True
  62         for dict_file_pair in strings:
  63             dict = dict_file_pair[1]
  64             if not dict.has_key(key):
  65                 is_in_all_dicts = False
  66             elif not dict[key] == value:
  67                 print("Element with different values")
  68                 print(key)
  69                 is_in_all_dicts = False
  70         if is_in_all_dicts:
  71             shared_dict[key] = value
  72     #print(shared_dict)
  73     for dict_file_pair in strings:
  74         for key in shared_dict.iterkeys():
  75             dict_file_pair[1].pop(key)
  76
  77     strings.append(["shared.ht", shared_dict])
  78     return strings
  79
  80 def writeOutFiles(dir, strings):
  81     for string in strings:
  82         file_name_base = string[0]
  83         file_name_base = file_name_base.replace(".ht", ".properties")
  84         file_name = os.path.join(dir, file_name_base)
  85         file = open(file_name, "w")
  86         for key, value in string[1].iteritems():
  87             try:
  88                 file.write(key)
  89                 file.write("=")
  90                 file.write(value)
  91                 file.write("\n")
  92             except UnicodeDecodeError:
  93                 print key
  94                 print value
  95         file.close()
  96
  97 def main (args):
  98     if(len(args) != 3):
  99         usage()
 100         sys.exit(1)
 101
 102     strings = parseFiles(args[1])
 103     new_strings = extractSharedEntries(strings)
 104     writeOutFiles(args[2], new_strings)
 105
 106 if __name__ == "__main__":
 107     main(sys.argv)