print "[Error]:", e finally: cur.close() conn.close() offset = 0.0 print "input download from db:" + str(len(self.pkg_dlcount_dict)) print str(self.max_dlcount) + "," + str(self.min_dlcount) + "," + str( self.log_max_dlcount) + "," + str( self.log_min_dlcount) + "," + str( (offset + self.log_min_dlcount) / (offset + self.log_max_dlcount)) if __name__ == '__main__': if len(sys.argv) != 4: print "<mergedSim><inFile><outFile>" sys.exit() inFile = sys.argv[2] outFile = sys.argv[3] g = GetFromFile(inFile) pkg_list = g.get_pkg_list() pkg_titlename_dict = g.get_pkg_titlename_dict() mergedSim = WriteTool.load_nested_dict(sys.argv[1]) b = BoostProcessor() boostedSim = b.boost(mergedSim, 0) WriteTool.write_nested_dict(boostedSim, outFile, pkg_titlename_dict, pkg_list)
other = line.strip().strip('\002').strip() sim_dict[pkg].append(other) else: print "error in line:" + str(line_num) print "load dict okay,len is " + str(len(sim_dict)) return sim_dict if __name__ == '__main__': if len(sys.argv) != 4: print "<discountedSim><inFile><outFile>" sys.exit() #discountedSim = WriteTool.load_nested_dict(sys.argv[1]) #discountedSim_list = WriteTool.nested_dict2list(discountedSim) g = GetFromFile(sys.argv[2]) pkg_list = g.get_pkg_list() pkg_titlename_dict = g.get_pkg_titlename_dict() sim_dict = load() of = open(sys.argv[3], 'w') for pkg in pkg_list: if pkg not in sim_dict: continue of.write(pkg + '\001') for t in sim_dict[pkg]: of.write(t + '\002') of.write('\n') of.close ''' of = open('data/pkg_list','w')
cleaned_game_list = [] TC = TitleClean() for game in game_list: temp = list(game) temp[0] = TC.clean(temp[0].encode('utf-8')) cleaned_game_list.append(list(temp)) print "in get_titleCleaned_game_list" for game in game_list: if game[0].encode('utf-8') == '捕鱼达人': print ",".join(game) print "after cleaned" for game in cleaned_game_list: if game[0].encode('utf-8') == '捕鱼达人': print ",".join(game) return cleaned_game_list if __name__ == '__main__': if len(sys.argv) != 2: print "<outFile>" sys.exit() g = GetFromFile() titlename_pkg_dict = g.get_titlename_pkg_dict() cleanedTitlename_pkg_dict = get_cleanedTitlename_pkg_dict(titlename_pkg_dict) game_list = inputData() cleaned_game_list = get_titleCleaned_game_list(game_list) #alianed_game_list = align(game_list,titlename_pkg_dict) alianed_game_list = align(cleaned_game_list,cleanedTitlename_pkg_dict) write(alianed_game_list,sys.argv[1])