# Read configuration
inDir = config.getProperty("dir.parser.in")
minWordCount = int(config.getProperty("parser.minWordCount"))
print("Script Parser v.", parser_version)
print("================================")
print("Input directory:", inDir)
print("Minimum words per character: ", minWordCount)
if minWordCount < 150:
print("[WARN] Minimum word count is important for clearing out the noise, caused by" +
" wrongly-recognized character names etc. - Avoid adjusting this parameter to a" +
" value below 150!")
print()
mongoClient, mongoDb = mongohelper.getMongoClient()
if len(sys.argv)==1 or (len(sys.argv)==2 and sys.argv[1]=='-forceUpdate'):
scriptList = [f for f in listdir(inDir) if isfile(join(inDir, f))]
else:
scriptList = [f for f in listdir(inDir) if isfile(join(inDir, f)) and parseMovieId(f) in sys.argv[1:]]
print('[ 0%] Processing', len(scriptList), 'scripts')
countTotal = 0
countError = 0
countSkip = 0
# Iterate over the input files
for i in range(len(scriptList)):
countTotal += 1
# -*- coding: UTF-8 -*-
import sys
# Include custom libs
sys.path.append( '../../include/python' )
import serverutils.config as config
import serverutils.mongohelper as mongohelper
import pymongo
props = {'targetDbName': 'characterinsights_copy'}
for prop in sys.argv[1:]:
k,v = prop.split("=",1)
props[k]=v
# Connect DBs
mongoClientSource, mongoDbSource = mongohelper.getMongoClient()
mongoClientTarget, mongoDbTarget = mongohelper.getMongoClient(orMongoMode='local',orHost='localhost',orDbName=props['targetDbName'])
collections = ['inouttest','movies','personalitydimensionnormalizations','rawMovies','rawQuotes','results']
for collection in collections:
print(collection)
for doc in mongoDbSource[collection].find():
mongoDbTarget[collection].insert_one(doc)
print('.',end='',flush=True)
print("")
print("Done.")
#!/usr/bin/python3
dump_version = "0.10"
# Include custom libs
import sys
sys.path.append('../../include/python')
import serverutils.config as config
import serverutils.mongohelper as mongohelper
import pymongo
mongoClient, mongoDb = mongohelper.getMongoClient(silent=True)
if len(sys.argv) < 2:
print("Usage: dumpMovie.py <movieId>")
print("")
print("Listing movies:")
for movie in mongoDb.rawMovies.find().sort("_id", pymongo.ASCENDING):
print(movie['_id'])
else:
movieId = sys.argv[1]
movieCursor = mongoDb.rawMovies.find({'_id': movieId})
if movieCursor.count() == 0:
print("The movie with ID \"" + movieId +
"\" could not be found. Calling this tool" +
" without a movie ID will list all available movies.")
else:
movieData = movieCursor.next()
print(movieData['normalizedMovieId'])
# Remove any trailing slashes
while newHostname[-1:] == "/" and len(newHostname) >= 1:
newHostname = newHostname[:-1]
if len(newHostname) > 0 and (newHostname[0:4] == 'http'
and newHostname[8:].find("/") > 0):
if not quiet:
print("Invalid new hostname:", newHostname)
print("The hostname must not contain any slashes")
exit()
if not quiet:
print("New host is", newHostname)
# Connect DBs
mongoClient, mongoDb = mongohelper.getMongoClient(silent=quiet)
for movie in mongoDb['movies'].find():
movieName = movie['names']['resolved'] if len(
movie['names']['resolved']) > 0 else movie['names']['scriptUnified']
if not quiet:
print("Processing", movieName)
dataChanged = False
oldMoviePosterUrl = movie['picture']['path']
newMoviePosterUrl = replaceUrl(oldMoviePosterUrl, newHostname)
if oldMoviePosterUrl != newMoviePosterUrl:
movie['picture']['path'] = newMoviePosterUrl
dataChanged = True