def main(config):
"""
main occurrences processor
reads a whitelist and push a occurrences_worker() to a process pool
"""
whitelistpath = config['cooccurrences']["whitelist"]["path"]
logging.debug("loading whitelist from %s (id = %s)"%(whitelistpath, whitelistpath))
wlimport = Reader('whitelist://'+whitelistpath, dialect="excel", encoding="ascii")
wlimport.whitelist = whitelist.Whitelist( whitelistpath, whitelistpath )
newwl = wlimport.parse_file()
newwl['content']=[]
# cursor of Whitelist NGrams db
ngramgenerator = newwl.getNGram()
outputs = output.getConfiguredOutputs(config['cooccurrences'])
try:
while 1:
ngid, ng = ngramgenerator.next()
newwl['content'] += [ng]
outputs['exportwhitelistcsv'].save("%s,%s\n"%(ngid,ng['label']))
#raise StopIteration()
except StopIteration:
logging.debug('imported %d n-lemmes from the whitelist file %s'\
%(len(newwl['content']), whitelistpath))
input = mongodbhandler.MongoDB(config['cooccurrences']['input_db'])
#occspool = pool.Pool(processes=config['processes'])
for notice in input.notices.find(timeout=False):
#occspool.apply_async(worker, (config, notice, newwl))
worker(config, notice, newwl)
def apply(self, record):
"""
filters notices not match a regular expression
"""
whitelistpath = self.getRules()
extraction_fields = rules['fields']
wlimport = Reader('whitelist://'+whitelistpath, dialect="excel", encoding="ascii")
wlimport.whitelist = whitelist.Whitelist( whitelistpath, whitelistpath )
newwl = wlimport.parse_file()
for tag in extraction_fields:
if tag not in record: continue
if type(record[tag]) == str or type(record[tag]) == unicode:
if newwl.test(record[tag]):
return 1
elif type(record[tag]) == list:
for field in record[tag]:
if newwl.test(field):
return 1
# anyway : reject
raise NoticeRejected("notice did not match the whitelist")
return 0
def _importwhitelist(self):
"""
loads and cache all ngrams in the whitelist
"""
whitelistpath = self.config["whitelist"]["path"]
logging.debug("loading whitelist from %s (id = %s)"%(whitelistpath, whitelistpath))
wlimport = Reader('whitelist://'+whitelistpath, dialect="excel", encoding="ascii")
wlimport.whitelist = whitelist.Whitelist( whitelistpath, whitelistpath )
self.newwl = wlimport.parse_file()
try:
self.newwl['content']=[]
# cursor of Whitelist NGrams db
ngramgenerator = self.newwl.getNGram()
while 1:
ngid, ng = ngramgenerator.next()
self.newwl['content'] += [ng]
self.outputs['exportwhitelistcsv'].save("%s,%s\n"%(ngid,ng['label']))
#raise StopIteration()
except StopIteration:
logging.debug('imported %d n-lemmes from the whitelist file %s'\
%(len(self.newwl['content']), whitelistpath))
if len(self.newwl['content'])<2:
raise Exception("the whitelist contains only one element, aborting")
def _import_whitelist(
self,
whitelistpath,
dataset = None,
userstopwords = None,
dialect="excel",
encoding="utf_8"
):
"""
import one or a list of whitelits files
returns a whitelist object to be used as input of other methods
"""
whitelist_id = self._get_filepath_id(whitelistpath)
kwargs = {
'dialect': dialect,
'encoding': encoding
}
if whitelist_id is not None:
### whitelistpath EXISTS
self.logger.debug("loading whitelist from %s (%s)"%(whitelistpath, whitelist_id))
wlimport = Reader('whitelist://'+whitelistpath, **kwargs)
wlimport.whitelist = whitelist.Whitelist( whitelist_id, whitelist_id )
new_wl = wlimport.parse_file()
# NOT USED : TO CHECK
elif isinstance(dataset, corpora.Corpora):
self._load_config()
storage = self.get_storage(dataset, create=False, drop_tables=False)
if storage == self.STATUS_ERROR:
return self.STATUS_ERROR
# whitelistpath is a whitelist label into storage
self.logger.debug("loading whitelist %s from storage"%whitelist_id)
new_wl = whitelist.Whitelist( whitelist_id, whitelist_id )
new_wl.loadFromStorage(storage, dataset)
elif exists(whitelistpath):
### whitelist path is a real path but not in a correct format
whitelist_id = dataset
self.logger.debug("loading whitelist from %s (%s)"%(whitelistpath, whitelist_id))
wlimport = Reader('whitelist://'+whitelistpath, **kwargs)
wlimport.whitelist = whitelist.Whitelist( whitelist_id, whitelist_id )
new_wl = wlimport.parse_file()
else:
raise Exception("unable to find a whitelist at %s"%whitelistpath)
return new_wl
outpath = self._get_user_filepath(
dataset,
'cooccurrences',
"%s-matrix.csv"%(whitelist['label'])
)
exporter = Writer("coocmatrix://"+outpath)
whitelist_outpath = self._get_user_filepath(
dataset,
'cooccurrences',
"%s-terms.csv"%(whitelist['label'])
)
whitelist_exporter = Writer("basecsv://"+whitelist_outpath)
else:
exporter = None
archive = Reader( format + "://" + path, **self.config['datasets'] )
archive_walker = archive.walkArchive(periods)
try:
period_gen, period = archive_walker.next()
sc = indexer.ArchiveCounter(self.config['datasets'], storage)
walkCorpusGen = sc.walk_period(whitelist, period_gen, period)
try:
while 1:
yield walkCorpusGen.next()
except StopIteration:
pass
writeMatrixGen = sc.write_matrix(period, exporter, whitelist_exporter, minCooc)
try:
while 1:
yield writeMatrixGen.next()
except StopIteration, si:
