def search(data):
"""
Search OMDB for an id based on a name/year.
:param data: {name, year}
:return: id
"""
name = data['name']
year = data['year']
# if we managed to parse the year from the name
# include it, since it'll narrow results
if year:
year_query = '&y={}'.format(year.replace('(', '').replace(')', ''))
else:
year_query = ''
try:
result = requests.get(TMDB_SEARCH_URL + name + year_query).json()
except:
log.critical('There was a problem accessing the IMDB API page.')
return None
if 'results' in result:
for movie in result['results']:
ratio = difflib.SequenceMatcher(
None, pynab.ids.clean_name(name),
pynab.ids.clean_name(movie['title'])).ratio()
if ratio > 0.8 and year in movie['release_date']:
temp = requests.get('https://api.themoviedb.org/3/movie/{}'.
format(movie['id']) + '?api_key=' +
API_KEY).json()
return temp['imdb_id']
return None
def search(data):
"""
Search OMDB for an id based on a name/year.
:param data: {name, year}
:return: id
"""
name = data['name']
year = data['year']
# if we managed to parse the year from the name
# include it, since it'll narrow results
if year:
year_query = '&y={}'.format(year.replace('(', '').replace(')', ''))
else:
year_query = ''
try:
result = requests.get(OMDB_SEARCH_URL + name + year_query).json()
except:
log.critical('There was a problem accessing the IMDB API page.')
return None
if 'Search' in result:
for movie in result['Search']:
# doublecheck, but the api should've searched properly
ratio = difflib.SequenceMatcher(None, pynab.ids.clean_name(name), pynab.ids.clean_name(movie['Title'])).ratio()
if ratio > 0.8 and year == movie['Year'] and movie['Type'] == 'movie':
return movie['imdbID']
return None
def search(data):
"""
Search OMDB for an id based on a name/year.
:param data: {name, year}
:return: id
"""
name = data['name']
year = data['year']
# if we managed to parse the year from the name
# include it, since it'll narrow results
if year:
year_query = '&y={}'.format(year.replace('(', '').replace(')', ''))
else:
year_query = ''
try:
result = requests.get(OMDB_SEARCH_URL + name + year_query).json()
except:
log.critical('There was a problem accessing the IMDB API page.')
return None
if 'Search' in result:
for movie in result['Search']:
# doublecheck, but the api should've searched properly
ratio = difflib.SequenceMatcher(
None, pynab.ids.clean_name(name),
pynab.ids.clean_name(movie['Title'])).ratio()
if ratio > 0.8 and year == movie['Year'] and movie[
'Type'] == 'movie':
return movie['imdbID']
return None
def search(name, year):
"""Search OMDB for a movie and return the IMDB ID."""
# if we managed to parse the year from the name
# include it, since it'll narrow results
if year:
year_query = '&y={}'.format(year.replace('(', '').replace(')', ''))
else:
year_query = ''
data = {}
try:
r = requests.get(OMDB_SEARCH_URL + name + year_query)
data = r.json()
except:
log.critical('There was a problem accessing the IMDB API page.')
return None
if 'Search' in data:
for movie in data['Search']:
# doublecheck, but the api should've searched properly
ratio = difflib.SequenceMatcher(None, clean_name(name), clean_name(movie['Title'])).ratio()
if ratio > 0.8 and year == movie['Year'] and movie['Type'] == 'movie':
return movie
def search_lxml(show, content):
"""Search TVRage online API for show data."""
try:
tree = etree.fromstring(content)
except:
log.critical('Problem parsing XML with lxml')
return None
matches = defaultdict(list)
# parse show names in the same order as returned by tvrage, first one is usually the good one
for xml_show in XPATH_SHOW(tree):
for name in extract_names(xml_show):
ratio = int(difflib.SequenceMatcher(None, show['clean_name'], clean_name(name)).ratio() * 100)
if ratio == 100:
return xmltodict.parse(etree.tostring(xml_show))['show']
matches[ratio].append(xml_show)
# if no 100% is found, check highest ratio matches
for ratio, xml_matches in sorted(matches.items(), reverse=True):
for xml_match in xml_matches:
if ratio >= 80:
return xmltodict.parse(etree.tostring(xml_match))['show']
elif 80 > ratio > 60:
if 'country' in show and show['country'] and XPATH_COUNTRY(xml_match):
if str.lower(show['country']) == str.lower(XPATH_COUNTRY(xml_match)[0]):
return xmltodict.parse(etree.tostring(xml_match))['show']
def process_movies():
try:
return pynab.ids.process(
'movie',
interfaces=config.postprocess.get('process_movies'),
limit=500)
except Exception as e:
log.critical(traceback.format_exc())
raise Exception
def daemonize(pidfile):
try:
import traceback
from daemonize import Daemonize
daemon = Daemonize(app='pynab', pid=pidfile, action=main)
daemon.start()
except SystemExit:
raise
except:
log.critical(traceback.format_exc())
def get_nzb_details(nzb):
"""Returns a JSON-like Python dict of NZB contents, including extra information
such as a list of any nfos/rars that the NZB references."""
try:
# using the html parser here instead of the straight lxml might be slower
# but some of the nzbs spewed forth by newznab are broken and contain
# non-xml entities, ie. ²
# this breaks the normal lxml parser
tree = html.fromstring(gzip.decompress(nzb.data))
except Exception as e:
log.critical('nzbs: problem parsing XML with lxml: {}'.format(e))
return None
nfos = []
sfvs = []
rars = []
pars = []
zips = []
rar_count = 0
par_count = 0
for file_subject in XPATH_FILE(tree):
if rar_part_regex.search(file_subject):
rar_count += 1
if nfo_regex.search(
file_subject) and not metadata_regex.search(file_subject):
nfos.append(filexml_to_dict(file_subject.getparent()))
if sfv_regex.search(file_subject):
sfvs.append(filexml_to_dict(file_subject.getparent()))
if rar_regex.search(
file_subject) and not metadata_regex.search(file_subject):
rars.append(filexml_to_dict(file_subject.getparent()))
if par2_regex.search(file_subject):
par_count += 1
if not par_vol_regex.search(file_subject):
pars.append(filexml_to_dict(file_subject.getparent()))
if zip_regex.search(
file_subject) and not metadata_regex.search(file_subject):
zips.append(filexml_to_dict(file_subject.getparent()))
return {
'nfos': nfos,
'sfvs': sfvs,
'rars': rars,
'pars': pars,
'zips': zips,
'rar_count': rar_count,
'par_count': par_count,
}
def daemonize(pidfile):
try:
import traceback
from daemonize import Daemonize
fds = []
if log_descriptor:
fds = [log_descriptor]
daemon = Daemonize(app='pynab', pid=pidfile, action=main, keep_fds=fds)
daemon.start()
except SystemExit:
raise
except:
log.critical(traceback.format_exc())
def get_nzb_details(nzb):
"""Returns a JSON-like Python dict of NZB contents, including extra information
such as a list of any nfos/rars that the NZB references."""
try:
# using the html parser here instead of the straight lxml might be slower
# but some of the nzbs spewed forth by newznab are broken and contain
# non-xml entities, ie. ²
# this breaks the normal lxml parser
tree = html.fromstring(gzip.decompress(nzb.data))
except Exception as e:
log.critical('nzbs: problem parsing XML with lxml: {}'.format(e))
return None
nfos = []
sfvs = []
rars = []
pars = []
zips = []
rar_count = 0
par_count = 0
for file_subject in XPATH_FILE(tree):
if rar_part_regex.search(file_subject):
rar_count += 1
if nfo_regex.search(file_subject) and not metadata_regex.search(file_subject):
nfos.append(filexml_to_dict(file_subject.getparent()))
if sfv_regex.search(file_subject):
sfvs.append(filexml_to_dict(file_subject.getparent()))
if rar_regex.search(file_subject) and not metadata_regex.search(file_subject):
rars.append(filexml_to_dict(file_subject.getparent()))
if par2_regex.search(file_subject):
par_count += 1
if not par_vol_regex.search(file_subject):
pars.append(filexml_to_dict(file_subject.getparent()))
if zip_regex.search(file_subject) and not metadata_regex.search(file_subject):
zips.append(filexml_to_dict(file_subject.getparent()))
return {
'nfos': nfos,
'sfvs': sfvs,
'rars': rars,
'pars': pars,
'zips': zips,
'rar_count': rar_count,
'par_count': par_count,
}
def get_size(nzb):
"""Returns the size of a release (in bytes) as given by the NZB, compressed."""
try:
# using the html parser here instead of the straight lxml might be slower
# but some of the nzbs spewed forth by newznab are broken and contain
# non-xml entities, ie. ²
# this breaks the normal lxml parser
tree = html.fromstring(gzip.decompress(nzb.data))
except Exception as e:
log.critical('nzbs: problem parsing XML with lxml: {}'.format(e))
return None
size = 0
for bytes in XPATH_BYTES(tree):
try:
size += int(bytes)
except:
# too bad, there was a problem
pass
return size
def get_size(nzb):
"""Returns the size of a release (in bytes) as given by the NZB, compressed."""
try:
# using the html parser here instead of the straight lxml might be slower
# but some of the nzbs spewed forth by newznab are broken and contain
# non-xml entities, ie. ²
# this breaks the normal lxml parser
tree = html.fromstring(gzip.decompress(nzb.data))
except Exception as e:
log.critical('nzbs: problem parsing XML with lxml: {}'.format(e))
return None
size = 0
for bytes in XPATH_BYTES(tree):
try:
size += int(bytes)
except:
# too bad, there was a problem
pass
return size
def process_requests():
try:
return pynab.requests.process(500)
except Exception as e:
log.critical(traceback.format_exc())
raise Exception
def process_movies():
try:
return pynab.ids.process("movie", interfaces=config.postprocess.get("process_movies"), limit=500)
except Exception as e:
log.critical(traceback.format_exc())
raise Exception
def process_requests():
try:
return pynab.requests.process(500)
except Exception as e:
log.critical(traceback.format_exc())
raise Exception
def process_imdb():
try:
return pynab.imdb.process(500)
except Exception as e:
log.critical(traceback.format_exc())
raise Exception
def save_all(parts):
"""Save a set of parts to the DB, in a batch if possible."""
if parts:
start = time.time()
group_name = list(parts.values())[0]['group_name']
with db_session() as db:
# this is a little tricky. parts have no uniqueness at all.
# no uniqid and the posted dates can change since it's based off the first
# segment that we see in that part, which is different for each scan.
# what we do is get the next-closest thing (subject+author+group) and
# order it by oldest first, so when it's building the dict the newest parts
# end on top (which are the most likely to be being saved to).
# realistically, it shouldn't be a big problem - parts aren't stored in the db
# for very long anyway, and they're only a problem while there. saving 500 million
# segments to the db is probably not a great idea anyway.
existing_parts = dict(
((part.hash, part) for part in
db.query(Part.id, Part.hash).filter(Part.hash.in_(parts.keys())).filter(
Part.group_name == group_name).order_by(Part.posted.asc()).all()
)
)
part_inserts = []
for hash, part in parts.items():
existing_part = existing_parts.get(hash, None)
if not existing_part:
segments = part.pop('segments')
part_inserts.append(part)
part['segments'] = segments
if part_inserts:
ordering = ['hash', 'subject', 'group_name', 'posted', 'posted_by', 'total_segments', 'xref']
s = io.StringIO()
for part in part_inserts:
for item in ordering:
if item == 'posted':
s.write('"' + part[item].replace(tzinfo=None).strftime('%Y-%m-%d %H:%M:%S').replace('"',
'\\"') + '",')
elif item == 'xref':
# leave off the comma
s.write('"' + part[item].encode('utf-8', 'replace').decode('utf-8').replace('"', '\\"') + '"')
else:
s.write('"' + str(part[item]).encode('utf-8', 'replace').decode().replace('"', '\\"') + '",')
s.write("\n")
s.seek(0)
if not copy_file(engine, s, ordering, Part):
return False
s.close()
db.close()
with db_session() as db:
existing_parts = dict(
((part.hash, part) for part in
db.query(Part)
.options(
subqueryload('segments'),
Load(Part).load_only(Part.id, Part.hash),
Load(Segment).load_only(Segment.id, Segment.segment)
)
.filter(Part.hash.in_(parts.keys()))
.filter(Part.group_name == group_name)
.order_by(Part.posted.asc())
.all()
)
)
segment_inserts = []
for hash, part in parts.items():
existing_part = existing_parts.get(hash, None)
if existing_part:
segments = dict(((s.segment, s) for s in existing_part.segments))
for segment_number, segment in part['segments'].items():
if int(segment_number) not in segments:
segment['part_id'] = existing_part.id
segment_inserts.append(segment)
else:
# we hit a duplicate message for a part
# kinda wish people would stop reposting shit constantly
pass
else:
log.critical(
'parts: part didn\'t exist when we went to save it. backfilling with dead_binary_age not set to 0?')
return False
if segment_inserts:
ordering = ['segment', 'size', 'message_id', 'part_id']
s = io.StringIO()
for segment in segment_inserts:
for item in ordering:
if item == 'part_id':
# leave off the tab
s.write('"' + str(segment[item]).replace('"', '\\"') + '"')
else:
s.write('"' + str(segment[item]).encode('utf-8', 'replace').decode('utf-8').replace('"', '\\"') + '",')
s.write("\n")
s.seek(0)
if not copy_file(engine, s, ordering, Segment):
return False
s.close()
db.close()
end = time.time()
log.debug('parts: saved {} parts and {} segments in {:.2f}s'.format(
len(part_inserts),
len(segment_inserts),
end - start
))
del part_inserts[:]
del segment_inserts[:]
return True
def save_all(parts):
"""Save a set of parts to the DB, in a batch if possible."""
if parts:
start = time.time()
group_name = list(parts.values())[0]['group_name']
with db_session() as db:
# this is a little tricky. parts have no uniqueness at all.
# no uniqid and the posted dates can change since it's based off the first
# segment that we see in that part, which is different for each scan.
# what we do is get the next-closest thing (subject+author+group) and
# order it by oldest first, so when it's building the dict the newest parts
# end on top (which are the most likely to be being saved to).
# realistically, it shouldn't be a big problem - parts aren't stored in the db
# for very long anyway, and they're only a problem while there. saving 500 million
# segments to the db is probably not a great idea anyway.
existing_parts = dict(((part.hash, part) for part in db.query(
Part.id, Part.hash).filter(Part.hash.in_(parts.keys())).filter(
Part.group_name == group_name).order_by(
Part.posted.asc()).all()))
part_inserts = []
for hash, part in parts.items():
existing_part = existing_parts.get(hash, None)
if not existing_part:
segments = part.pop('segments')
part_inserts.append(part)
part['segments'] = segments
if part_inserts:
ordering = [
'hash', 'subject', 'group_name', 'posted', 'posted_by',
'total_segments', 'xref'
]
s = io.StringIO()
for part in part_inserts:
for item in ordering:
if item == 'posted':
s.write('"' + part[item].replace(
tzinfo=None).strftime(
'%Y-%m-%d %H:%M:%S').replace('"', '\\"') +
'",')
elif item == 'xref':
# leave off the comma
s.write('"' +
part[item].encode('utf-8', 'replace').
decode('utf-8').replace('"', '\\"') + '"')
else:
s.write('"' +
str(part[item]).encode('utf-8', 'replace').
decode().replace('"', '\\"') + '",')
s.write("\n")
s.seek(0)
if not copy_file(engine, s, ordering, Part):
return False
s.close()
db.close()
with db_session() as db:
existing_parts = dict(
((part.hash, part) for part in db.query(Part).options(
subqueryload('segments'),
Load(Part).load_only(Part.id, Part.hash),
Load(Segment).load_only(Segment.id, Segment.segment)).
filter(Part.hash.in_(parts.keys())).filter(
Part.group_name == group_name).order_by(
Part.posted.asc()).all()))
segment_inserts = []
for hash, part in parts.items():
existing_part = existing_parts.get(hash, None)
if existing_part:
segments = dict(
((s.segment, s) for s in existing_part.segments))
for segment_number, segment in part['segments'].items():
if int(segment_number) not in segments:
segment['part_id'] = existing_part.id
segment_inserts.append(segment)
else:
# we hit a duplicate message for a part
# kinda wish people would stop reposting shit constantly
pass
else:
log.critical(
'parts: part didn\'t exist when we went to save it. backfilling with dead_binary_age not set to 0?'
)
return False
if segment_inserts:
ordering = ['segment', 'size', 'message_id', 'part_id']
s = io.StringIO()
for segment in segment_inserts:
for item in ordering:
if item == 'part_id':
# leave off the tab
s.write('"' +
str(segment[item]).replace('"', '\\"') +
'"')
else:
s.write(
'"' +
str(segment[item]).encode('utf-8', 'replace').
decode('utf-8').replace('"', '\\"') + '",')
s.write("\n")
s.seek(0)
if not copy_file(engine, s, ordering, Segment):
return False
s.close()
db.close()
end = time.time()
log.debug('parts: saved {} parts and {} segments in {:.2f}s'.format(
len(part_inserts), len(segment_inserts), end - start))
del part_inserts[:]
del segment_inserts[:]
return True
