def get_indexer_nzb_link(indexer_name, indexerguid, title, searchid, mode, log_api_access):
"""
Build a link that leads to the actual NZB of the indexer using the given informations. We log this as indexer API access and NZB download because this is only called
when the NZB will be actually downloaded later (by us or a downloader)
:return: str
"""
for p in indexers.enabled_indexers:
if p.name.strip() == indexer_name.strip():
link = p.get_nzb_link(indexerguid, title)
# Log to database
indexer = Indexer.get(fn.lower(Indexer.name) == indexer_name.lower())
papiaccess = IndexerApiAccess(indexer=p.indexer, type="nzb", url=link, response_successful=None, indexer_search=searchid) if log_api_access else None
try:
papiaccess.username = request.authorization.username if request.authorization is not None else None
except RuntimeError:
pass
papiaccess.save()
pnzbdl = IndexerNzbDownload(indexer=indexer, indexer_search=searchid, api_access=papiaccess, mode=mode, title=title, guid=indexerguid)
pnzbdl.save()
return link, papiaccess, pnzbdl
else:
logger.error("Did not find indexer with name %s" % indexer_name)
return None, None, None
def get_url_with_papi_access(self, url, type, cookies=None, timeout=None, saveToDb=True):
papiaccess = IndexerApiAccess(indexer=self.indexer, type=type, url=url, time=arrow.utcnow().datetime)
try:
papiaccess.username = request.authorization.username if request.authorization is not None else None
except RuntimeError:
# Is thrown when we're searching which is run in a thread. When downloading NFOs or whatever this will work
pass
indexerStatus = None
try:
time_before = arrow.utcnow()
response = self.get(url, cookies=cookies, timeout=timeout)
response.raise_for_status()
time_after = arrow.utcnow()
papiaccess.response_time = (time_after - time_before).seconds * 1000 + ((time_after - time_before).microseconds / 1000)
papiaccess.response_successful = True
self.debug("HTTP request to indexer completed in %dms" % papiaccess.response_time)
indexerStatus = self.handle_indexer_success(doSaveIndexerStatus=saveToDb)
except RequestException as e:
self.error("Error while connecting to URL %s: %s" % (url, str(e)))
papiaccess.error = "Connection failed: %s" % removeSensitiveData(str(e))
response = None
indexerStatus = self.handle_indexer_failure("Connection failed: %s" % removeSensitiveData(str(e)), saveIndexerStatus=saveToDb)
finally:
if saveToDb:
self.saveIndexerStatus(papiaccess)
return response, papiaccess, indexerStatus
def get_indexer_nzb_link(searchResultId, mode, log_api_access, internal=False):
"""
Build a link that leads to the actual NZB of the indexer using the given informations. We log this as indexer API access and NZB download because this is only called
when the NZB will be actually downloaded later (by us or a downloader)
:return: str
"""
searchResult = SearchResult.get(SearchResult.id == searchResultId)
indexerName = searchResult.indexer.name
indexer = indexers.getIndexerByName(indexerName)
link = searchResult.link
# Log to database
papiaccess = IndexerApiAccess(
indexer=indexer.indexer,
type="nzb",
url=link,
response_successful=None) if log_api_access else None
try:
papiaccess.username = request.authorization.username if request.authorization is not None else None
except RuntimeError:
pass
papiaccess.save()
pnzbdl = IndexerNzbDownload(searchResult=searchResult,
apiAccess=papiaccess,
mode=mode,
title=searchResult.title,
internal=internal)
pnzbdl.save()
return link, papiaccess, pnzbdl
def get_avg_indexer_response_times(after, before):
result = []
response_times = []
for p in Indexer.select().order_by(Indexer.name):
try:
indexer = getIndexerByName(p.name)
if not indexer.settings.enabled:
logger.debug("Skipping download stats for %s" % p.name)
continue
except IndexerNotFoundException:
logger.error("Unable to find indexer %s in configuration" % p.name)
continue
where = (IndexerApiAccess.response_successful) & (
IndexerApiAccess.indexer == p
) & (IndexerApiAccess.time > after) & (IndexerApiAccess.time < before)
avg_response_time = IndexerApiAccess().select(
fn.AVG(IndexerApiAccess.response_time)).where(where).tuples()[0][0]
if avg_response_time:
response_times.append({
"name": p.name,
"avgResponseTime": int(avg_response_time)
})
where = (IndexerApiAccess.response_successful) & (
IndexerApiAccess.response_time is not None) & (
IndexerApiAccess.time > after) & (IndexerApiAccess.time < before)
avg_response_time = IndexerApiAccess().select(
fn.AVG(IndexerApiAccess.response_time)).where(where).tuples()[0][0]
for i in response_times:
delta = i["avgResponseTime"] - avg_response_time
i["delta"] = delta
result.append(i)
result = sorted(result, key=lambda x: x["name"])
result = sorted(result, key=lambda x: x["avgResponseTime"])
return result
def get_url_with_papi_access(self, url, type, cookies=None, timeout=None, saveToDb=True):
papiaccess = IndexerApiAccess(indexer=self.indexer, type=type, url=url, time=arrow.utcnow().datetime)
try:
papiaccess.username = request.authorization.username if request.authorization is not None else None
except RuntimeError:
#Is thrown when we're searching which is run in a thread. When downloading NFOs or whatever this will work
pass
indexerStatus = None
try:
time_before = arrow.utcnow()
response = self.get(url, cookies=cookies, timeout=timeout)
response.raise_for_status()
time_after = arrow.utcnow()
papiaccess.response_time = (time_after - time_before).seconds * 1000 + ((time_after - time_before).microseconds / 1000)
papiaccess.response_successful = True
indexerStatus = self.handle_indexer_success(saveIndexerStatus=saveToDb)
except RequestException as e:
self.logger.error("Error while connecting to URL %s: %s" % (url, str(e)))
papiaccess.error = "Connection failed: %s" % removeSensitiveData(str(e))
response = None
indexerStatus = self.handle_indexer_failure("Connection failed: %s" % removeSensitiveData(str(e)), saveIndexerStatus=saveToDb)
finally:
if saveToDb:
papiaccess.save()
return response, papiaccess, indexerStatus
def checkHitOrDownloadLimit(p):
if p.settings.hitLimit > 0 or p.settings.downloadLimit > 0:
if p.settings.hitLimitResetTime:
comparisonTime = arrow.utcnow().replace(hour=p.settings.hitLimitResetTime, minute=0, second=0)
if comparisonTime > arrow.utcnow():
comparisonTime = arrow.get(comparisonTime.datetime - datetime.timedelta(days=1)) # Arrow is too dumb to properly subtract 1 day (throws an error on every first of the month)
else:
# Use rolling time window
comparisonTime = arrow.get(arrow.utcnow().datetime - datetime.timedelta(days=1))
if p.settings.hitLimit > 0:
apiHitsQuery = IndexerApiAccess().select().where((IndexerApiAccess.indexer == p.indexer) & (IndexerApiAccess.time > comparisonTime) & IndexerApiAccess.response_successful)
apiHits = apiHitsQuery.count()
if apiHits >= p.settings.hitLimit:
if p.settings.hitLimitResetTime:
logger.info("Did not pick %s because its API hit limit of %d was reached. Will pick again after %02d:00" % (p, p.settings.hitLimit, p.settings.hitLimitResetTime))
else:
try:
firstHitTimeInWindow = arrow.get(list(apiHitsQuery.order_by(IndexerApiAccess.time.desc()).offset(p.settings.hitLimit-1).dicts())[0]["time"]).to("local")
nextHitAfter = arrow.get(firstHitTimeInWindow + datetime.timedelta(days=1))
logger.info("Did not pick %s because its API hit limit of %d was reached. Next possible hit at %s" % (p, p.settings.hitLimit, nextHitAfter.format('YYYY-MM-DD HH:mm')))
except IndexerApiAccess.DoesNotExist:
logger.info("Did not pick %s because its API hit limit of %d was reached" % (p, p.settings.hitLimit))
return False, "API limit reached"
else:
logger.debug("%s has had %d of a maximum of %d API hits since %02d:%02d" % (p, apiHits, p.settings.hitLimit, comparisonTime.hour, comparisonTime.minute))
if p.settings.downloadLimit > 0:
downloadsQuery = IndexerNzbDownload().select(IndexerApiAccess, IndexerNzbDownload).join(IndexerApiAccess).where((IndexerApiAccess.indexer == p.indexer) & (IndexerApiAccess.time > comparisonTime))
downloads = downloadsQuery.count()
if downloads >= p.settings.downloadLimit:
if p.settings.hitLimitResetTime:
logger.info("Did not pick %s because its download limit of %d was reached. Will pick again after %02d:00" % (p, p.settings.downloadLimit, p.settings.hitLimitResetTime))
else:
try:
firstHitTimeInWindow = arrow.get(list(downloadsQuery.order_by(IndexerApiAccess.time.desc()).offset(p.settings.downloadLimit-1).limit(1).dicts())[0]["time"]).to("local")
nextHitAfter = arrow.get(firstHitTimeInWindow + datetime.timedelta(days=1))
logger.info("Did not pick %s because its download limit of %d was reached. Next possible hit at %s" % (p, p.settings.downloadLimit, nextHitAfter.format('YYYY-MM-DD HH:mm')))
except IndexerApiAccess.DoesNotExist:
logger.info("Did not pick %s because its download limit of %d was reached" % (p, p.settings.downloadLimit))
return False, "Download limit reached"
else:
logger.debug("%s has had %d of a maximum of %d downloads since %02d:%02d" % (p, downloads, p.settings.downloadLimit, comparisonTime.hour, comparisonTime.minute))
return True, None
def checkHitOrDownloadLimit(p):
if p.settings.hitLimit > 0 or p.settings.downloadLimit > 0:
if p.settings.hitLimitResetTime:
comparisonTime = arrow.utcnow().replace(hour=p.settings.hitLimitResetTime, minute=0, second=0)
if comparisonTime > arrow.utcnow():
comparisonTime = arrow.get(comparisonTime.datetime - datetime.timedelta(days=1)) # Arrow is too dumb to properly subtract 1 day (throws an error on every first of the month)
else:
# Use rolling time window
comparisonTime = arrow.get(arrow.utcnow().datetime - datetime.timedelta(days=1))
if p.settings.hitLimit > 0:
apiHitsQuery = IndexerApiAccess().select().where((IndexerApiAccess.indexer == p.indexer) & (IndexerApiAccess.time > comparisonTime) & IndexerApiAccess.response_successful)
apiHits = apiHitsQuery.count()
if apiHits >= p.settings.hitLimit:
if p.settings.hitLimitResetTime:
logger.info("Did not pick %s because its API hit limit of %d was reached. Will pick again after %02d:00" % (p, p.settings.hitLimit, p.settings.hitLimitResetTime))
else:
try:
firstHitTimeInWindow = arrow.get(list(apiHitsQuery.order_by(IndexerApiAccess.time.desc()).offset(p.settings.hitLimit - 1).dicts())[0]["time"]).to("local")
nextHitAfter = arrow.get(firstHitTimeInWindow + datetime.timedelta(days=1))
logger.info("Did not pick %s because its API hit limit of %d was reached. Next possible hit at %s" % (p, p.settings.hitLimit, nextHitAfter.format('YYYY-MM-DD HH:mm')))
except IndexerApiAccess.DoesNotExist:
logger.info("Did not pick %s because its API hit limit of %d was reached" % (p, p.settings.hitLimit))
return False, "API limit reached"
else:
logger.debug("%s has had %d of a maximum of %d API hits since %02d:%02d" % (p, apiHits, p.settings.hitLimit, comparisonTime.hour, comparisonTime.minute))
if p.settings.downloadLimit > 0:
downloadsQuery = IndexerNzbDownload().select(IndexerApiAccess, IndexerNzbDownload).join(IndexerApiAccess).where((IndexerApiAccess.indexer == p.indexer) & (IndexerApiAccess.time > comparisonTime))
downloads = downloadsQuery.count()
if downloads >= p.settings.downloadLimit:
if p.settings.hitLimitResetTime:
logger.info("Did not pick %s because its download limit of %d was reached. Will pick again after %02d:00" % (p, p.settings.downloadLimit, p.settings.hitLimitResetTime))
else:
try:
firstHitTimeInWindow = arrow.get(list(downloadsQuery.order_by(IndexerApiAccess.time.desc()).offset(p.settings.downloadLimit - 1).limit(1).dicts())[0]["time"]).to("local")
nextHitAfter = arrow.get(firstHitTimeInWindow + datetime.timedelta(days=1))
logger.info("Did not pick %s because its download limit of %d was reached. Next possible hit at %s" % (p, p.settings.downloadLimit, nextHitAfter.format('YYYY-MM-DD HH:mm')))
except IndexerApiAccess.DoesNotExist:
logger.info("Did not pick %s because its download limit of %d was reached" % (p, p.settings.downloadLimit))
return False, "Download limit reached"
else:
logger.debug("%s has had %d of a maximum of %d downloads since %02d:%02d" % (p, downloads, p.settings.downloadLimit, comparisonTime.hour, comparisonTime.minute))
return True, None
def get_indexer_nzb_link(searchResultId, mode, log_api_access, internal=False):
"""
Build a link that leads to the actual NZB of the indexer using the given informations. We log this as indexer API access and NZB download because this is only called
when the NZB will be actually downloaded later (by us or a downloader)
:return: str
"""
searchResult = SearchResult.get(SearchResult.id == searchResultId)
indexerName = searchResult.indexer.name
indexer = indexers.getIndexerByName(indexerName)
link = searchResult.link
# Log to database
papiaccess = IndexerApiAccess(indexer=indexer.indexer, type="nzb", url=link, response_successful=None) if log_api_access else None
try:
papiaccess.username = request.authorization.username if request.authorization is not None else None
except RuntimeError:
pass
papiaccess.save()
pnzbdl = IndexerNzbDownload(searchResult=searchResult, apiAccess=papiaccess, mode=mode, title=searchResult.title, internal=internal)
pnzbdl.save()
return link, papiaccess, pnzbdl
def get_nzb_link(indexer_name, guid, title, searchid):
"""
Build a link that leads to the actual NZB of the indexer using the given informations. We log this as indexer API access and NZB download because this is only called
when the NZB will be actually downloaded later (by us or a downloader)
:return: str
"""
for p in indexers.enabled_indexers:
if p.name == indexer_name:
link = p.get_nzb_link(guid, title)
# Log to database
indexer = Indexer.get(fn.lower(Indexer.name) == indexer_name.lower())
papiaccess = IndexerApiAccess(indexer=p.indexer, type="nzb", url=link, response_successful=None, indexer_search=indexer)
papiaccess.save()
pnzbdl = IndexerNzbDownload(indexer=indexer, indexer_search=searchid, api_access=papiaccess, mode="redirect")
pnzbdl.save()
return link
else:
logger.error("Did not find indexer with name %s" % indexer_name)
return None
def testIndexersApiLimits(self):
config.settings.searching.generate_queries = []
self.newznab1.hitLimit = 3
self.newznab1.hitLimitResetTime = None
config.settings.indexers = [self.newznab1]
read_indexers_from_config()
search_request = SearchRequest()
indexers = search.pick_indexers(search_request)
self.assertEqual(1, len(indexers))
dbsearch = Search(internal=True, time=arrow.utcnow().datetime)
dbsearch.save()
indexer = Indexer().get(name="newznab1")
#Two accesses one and 12 hours ago
IndexerApiAccess(indexer=indexer, search=dbsearch, time=arrow.utcnow().replace(hours=-1).datetime, type="search", url="", response_successful=True).save()
IndexerApiAccess(indexer=indexer, search=dbsearch, time=arrow.utcnow().replace(hours=-12).datetime, type="search", url="", response_successful=True).save()
self.assertEqual(1, len(search.pick_indexers(search_request)))
#Another one 20 hours ago, so limit should be reached
IndexerApiAccess(indexer=indexer, search=dbsearch, time=arrow.utcnow().replace(hours=-20).datetime, type="search", url="", response_successful=True).save()
self.assertEqual(0, len(search.pick_indexers(search_request)))
def get_indexer_nzb_link(indexer_name, indexerguid, title, searchid, mode,
log_api_access):
"""
Build a link that leads to the actual NZB of the indexer using the given informations. We log this as indexer API access and NZB download because this is only called
when the NZB will be actually downloaded later (by us or a downloader)
:return: str
"""
for p in indexers.enabled_indexers:
if p.name.strip() == indexer_name.strip():
link = p.get_nzb_link(indexerguid, title)
# Log to database
indexer = Indexer.get(
fn.lower(Indexer.name) == indexer_name.lower())
papiaccess = IndexerApiAccess(
indexer=p.indexer,
type="nzb",
url=link,
response_successful=None,
indexer_search=searchid) if log_api_access else None
try:
papiaccess.username = request.authorization.username if request.authorization is not None else None
except RuntimeError:
pass
papiaccess.save()
pnzbdl = IndexerNzbDownload(indexer=indexer,
indexer_search=searchid,
api_access=papiaccess,
mode=mode,
title=title,
guid=indexerguid)
pnzbdl.save()
return link, papiaccess, pnzbdl
else:
logger.error("Did not find indexer with name %s" % indexer_name)
return None, None, None
def get_indexer_response_times():
result = []
for p in Indexer.select().order_by(Indexer.name):
result.append({
"key":
p.name,
"values": [{
"responseTime": x.response_time,
"date": x.time.timestamp
} for x in IndexerApiAccess().select(
IndexerApiAccess.response_time, IndexerApiAccess.time).where(
(IndexerApiAccess.response_successful)
& (IndexerApiAccess.indexer == p)).join(Indexer).limit(1)]
})
return result
def get_avg_indexer_response_times():
result = []
response_times = []
for p in Indexer.select().order_by(Indexer.name):
avg_response_time = IndexerApiAccess().select(
fn.AVG(IndexerApiAccess.response_time)).where(
(IndexerApiAccess.response_successful)
& (IndexerApiAccess.indexer == p)).tuples()[0][0]
if avg_response_time:
response_times.append({
"name": p.name,
"avgResponseTime": avg_response_time
})
avg_response_time = IndexerApiAccess().select(
fn.AVG(IndexerApiAccess.response_time)).where(
(IndexerApiAccess.response_successful)
& (IndexerApiAccess.response_time is not None)).tuples()[0][0]
for i in response_times:
delta = i["avgResponseTime"] - avg_response_time
i["delta"] = delta
result.append(i)
return result
def download_nzb_and_log(indexer_name, provider_guid, title, searchid):
"""
Gets the NZB link from the indexer using the guid, downloads it and logs the download
:param indexer_name: name of the indexer
:param provider_guid: guid to build link
:param title: the title to build the link
:param searchid: the id of the IndexerSearch entry so we can link the download to a search
:return: IndexerNzbDownloadResult
"""
for p in indexers.enabled_indexers:
if p.name == indexer_name:
link = p.get_nzb_link(provider_guid, title)
indexer = Indexer.get(fn.lower(Indexer.name) == indexer_name.lower())
psearch = IndexerSearch.get((IndexerSearch.indexer == indexer) & (IndexerSearch.search == searchid))
papiaccess = IndexerApiAccess(indexer=p.indexer, type="nzb", url=link, indexer_search=psearch)
papiaccess.save()
internallink, guid = get_nzb_link_and_guid(indexer_name, provider_guid, searchid, title)
pnzbdl = IndexerNzbDownload(indexer=indexer, indexer_search=searchid, api_access=papiaccess, mode="serve", title=title, guid=internallink)
pnzbdl.save()
try:
r = p.get(link, timeout=10)
r.raise_for_status()
papiaccess.response_successful = True
papiaccess.response_time = r.elapsed.microseconds / 1000
return IndexerNzbDownloadResult(content=r.content, headers=r.headers)
except RequestException as e:
logger.error("Error while connecting to URL %s: %s" % (link, str(e)))
papiaccess.error = str(e)
return None
finally:
papiaccess.save()
else:
return "Unable to find NZB link"
def get_url_with_papi_access(self, url, type, cookies=None, timeout=None):
papiaccess = IndexerApiAccess(indexer=self.indexer, type=type, url=url, time=arrow.utcnow().datetime)
try:
time_before = arrow.utcnow()
response = self.get(url, cookies=cookies, timeout=timeout)
response.raise_for_status()
time_after = arrow.utcnow()
papiaccess.response_time = (time_after - time_before).seconds * 1000 + ((time_after - time_before).microseconds / 1000)
papiaccess.response_successful = True
self.handle_indexer_success()
except RequestException as e:
self.logger.error("Error while connecting to URL %s: %s" % (url, str(e)))
papiaccess.error = "Connection failed: %s" % str(e)
response = None
self.handle_indexer_failure("Connection failed: %s" % str(e))
finally:
papiaccess.save()
return response, papiaccess
def pick_indexers(search_request, internal=True):
# type: (nzbhydra.search.SearchRequest, bool) -> List[nzbhydra.search_modules.SearchModule]
query_supplied = True if search_request.query else False
queryCanBeGenerated = None #Store if we can generate a query from IDs. Initiall true but when we need this the first time and query generation fails we set it to false
picked_indexers = []
selected_indexers = search_request.indexers.split(
"|") if search_request.indexers is not None else None
for p in indexers.enabled_indexers:
if not p.settings.enabled:
logger.debug("Did not pick %s because it is disabled" % p)
continue
if internal and p.settings.accessType == "external":
logger.debug(
"Did not pick %s because it is only enabled for external searches"
% p)
continue
if not internal and p.settings.accessType == "internal":
logger.debug(
"Did not pick %s because it is only enabled for internal searches"
% p)
continue
if selected_indexers and p.name not in selected_indexers:
logger.debug(
"Did not pick %s because it was not selected by the user" % p)
continue
try:
status = p.indexer.status.get()
if status.disabled_until > arrow.utcnow(
) and not config.settings.searching.ignoreTemporarilyDisabled:
logger.info(
"Did not pick %s because it is disabled temporarily due to an error: %s"
% (p, status.reason))
continue
except IndexerStatus.DoesNotExist:
pass
if p.settings.hitLimit > 0:
if p.settings.hitLimitResetTime:
hitLimitResetTime = arrow.get(p.settings.hitLimitResetTime)
comparisonTime = arrow.now().replace(
hour=hitLimitResetTime.hour,
minute=hitLimitResetTime.minute,
second=0)
if comparisonTime > arrow.now():
comparisonTime = arrow.get(
comparisonTime.datetime - timedelta(days=1)
) #Arrow is too dumb to properly subtract 1 day (throws an error on every first of the month)
else:
comparisonTime = arrow.now().replace(hour=0,
minute=0,
second=0)
apiHits = IndexerApiAccess().select().where(
(IndexerApiAccess.indexer == p.indexer)
& (IndexerApiAccess.time > comparisonTime)
& IndexerApiAccess.response_successful).count()
if apiHits > p.settings.hitLimit:
logger.info(
"Did not pick %s because its API hit limit of %d was reached"
% (p, p.settings.hitLimit))
continue
else:
logger.debug(
"%s has had %d of a maximum of %d API hits since %02d:%02d"
% (p, apiHits, p.settings.hitLimit, comparisonTime.hour,
comparisonTime.minute))
if (query_supplied or search_request.identifier_key
is not None) and not p.supports_queries:
logger.debug(
"Did not pick %s because a query was supplied but the indexer does not support queries"
% p)
continue
# Here on we check if we could supply the indexer with generated/retrieved data like the title of a series
if not query_supplied and p.needs_queries and search_request.identifier_key is None:
logger.debug(
"Did not pick %s because no query was supplied but the indexer needs queries"
% p)
continue
# If we can theoretically do that we must try to actually get the title, otherwise the indexer won't be able to search
allow_query_generation = (
config.InternalExternalSelection.internal
in config.settings.searching.generate_queries
and internal) or (config.InternalExternalSelection.external
in config.settings.searching.generate_queries
and not internal)
if search_request.identifier_key is not None and not canUseIdKey(
p, search_request.identifier_key):
if not (allow_query_generation and p.generate_queries):
logger.debug(
"Did not pick %s because search will be done by an identifier and the indexer or system wide settings don't allow query generation"
% p)
continue
else:
if queryCanBeGenerated is None:
try:
title = infos.convertId(
search_request.identifier_key, "title",
search_request.identifier_value)
if title:
search_request.title = title
queryCanBeGenerated = True
else:
queryCanBeGenerated = False
except:
queryCanBeGenerated = False
logger.debug(
"Unable to get title for supplied ID. Indexers that don't support the ID will be skipped"
)
if not queryCanBeGenerated:
logger.debug(
"Did not pick %s because search will be done by an identifier and retrieval of the title for query generation failed"
% p)
continue
logger.debug("Picked %s" % p)
picked_indexers.append(p)
return picked_indexers
def execute_queries(self, queries, searchRequest):
if len(queries) == 0:
return QueriesExecutionResult(didsearch=False, results=[], indexerSearchEntry=None, indexerApiAccessEntry=None, indexerStatus=None, total=0, loaded_results=0, total_known=True, has_more=False)
results = []
executed_queries = set()
psearch = IndexerSearch(indexer=self.indexer)
papiaccess = IndexerApiAccess()
indexerStatus = None
#psearch.save()
total_results = 0
total_known = False
has_more = False
while len(queries) > 0:
query = queries.pop()
if query in executed_queries:
# To make sure that in case an offset is reported wrong or we have a bug we don't get stuck in an endless loop
continue
try:
request, papiaccess, indexerStatus = self.get_url_with_papi_access(query, "search", saveToDb=False)
papiaccess.indexer_search = psearch
executed_queries.add(query)
#papiaccess.save()
if request is not None:
self.check_auth(request.text)
self.debug("Successfully loaded URL %s" % request.url)
try:
parsed_results = self.process_query_result(request.content, searchRequest)
results.extend(parsed_results.entries) # Retrieve the processed results
queries.extend(parsed_results.queries) # Add queries that were added as a result of the parsing, e.g. when the next result page should also be loaded
total_results += parsed_results.total
total_known = parsed_results.total_known
has_more = parsed_results.has_more
papiaccess.response_successful = True
self.handle_indexer_success(False)
except IndexerResultParsingException as e:
self.error("Error while processing search results from indexer %s" % e)
except Exception:
self.exception("Error while processing search results from indexer %s" % self)
raise IndexerResultParsingException("Error while parsing the results from indexer", self)
except IndexerAuthException as e:
self.error("Unable to authorize with %s: %s" % (e.search_module, e.message))
papiaccess.error = "Authorization error :%s" % e.message
self.handle_indexer_failure(reason="Authentication failed", disable_permanently=True)
papiaccess.response_successful = False
except IndexerAccessException as e:
self.error("Unable to access %s: %s" % (e.search_module, e.message))
papiaccess.error = "Access error: %s" % e.message
self.handle_indexer_failure(reason="Access failed")
papiaccess.response_successful = False
except IndexerResultParsingException as e:
papiaccess.exception = "Access error: %s" % e.message
self.handle_indexer_failure(reason="Parsing results failed")
papiaccess.response_successful = False
except Exception as e:
self.exception("An error error occurred while searching: %s", e)
if papiaccess is not None:
papiaccess.error = "Unknown error :%s" % e
papiaccess.response_successful = False
finally:
if papiaccess is not None:
#papiaccess.save()
psearch.successful = papiaccess.response_successful
else:
self.error("Unable to save API response to database")
psearch.resultsCount = total_results
#psearch.save()
return QueriesExecutionResult(didsearch= True, results=results, indexerSearchEntry=psearch, indexerApiAccessEntry=papiaccess, indexerStatus=indexerStatus, total=total_results, loaded_results=len(results), total_known=total_known, has_more=has_more)
def testThatDatabaseValuesAreStored(self):
with self.app.test_request_context('/'):
with responses.RequestsMock(assert_all_requests_are_fired=False) as rsps:
newznabItems = [
[mockbuilder.buildNewznabItem(title="title1", pubdate=arrow.get(1000).format("ddd, DD MMM YYYY HH:mm:ss Z"), size=1000, indexer_name="newznab1")],
[mockbuilder.buildNewznabItem(title="title2", pubdate=arrow.get(1000).format("ddd, DD MMM YYYY HH:mm:ss Z"), size=1000, indexer_name="newznab2")]
]
self.prepareSearchMocks(rsps, indexerCount=len(newznabItems), newznabItems=newznabItems)
# Make the second access unsuccessful
rsps._urls.pop(1)
rsps.add(responses.GET, r".*",
body="an error message", status=500,
content_type='application/x-html')
searchRequest = SearchRequest(type="search", query="aquery", category="acategory", identifier_key="imdbid", identifier_value="animdbid", season=1, episode=2, indexers="newznab1|newznab2")
result = search.search(searchRequest)
results = result["results"]
self.assertEqual(1, len(results))
dbSearch = Search().get()
self.assertEqual(True, dbSearch.internal)
self.assertEqual("aquery", dbSearch.query)
self.assertEqual("All", dbSearch.category)
self.assertEqual("imdbid", dbSearch.identifier_key)
self.assertEqual("animdbid", dbSearch.identifier_value)
self.assertEqual("1", dbSearch.season)
self.assertEqual("2", dbSearch.episode)
self.assertEqual("search", dbSearch.type)
self.assertEqual(18, dbSearch.time.hour)
indexerSearch1 = IndexerSearch.get(IndexerSearch.indexer == Indexer.get(Indexer.name == "newznab1"))
self.assertEqual(indexerSearch1.search, dbSearch)
self.assertEqual(18, indexerSearch1.time.hour)
indexerSearch2 = IndexerSearch.get(IndexerSearch.indexer == Indexer.get(Indexer.name == "newznab2"))
self.assertEqual(indexerSearch2.search, dbSearch)
self.assertEqual(18, indexerSearch2.time.hour)
calledUrls = sorted([x.request.url for x in rsps.calls])
indexerApiAccess1 = IndexerApiAccess.get(IndexerApiAccess.indexer == Indexer.get(Indexer.name == "newznab1"))
self.assertEqual(indexerSearch1, indexerApiAccess1.indexer_search)
self.assertEqual(18, indexerApiAccess1.time.hour)
self.assertEqual("search", indexerApiAccess1.type)
self.assertEqual(calledUrls[0], indexerApiAccess1.url)
self.assertTrue(indexerApiAccess1.response_successful)
self.assertEqual(0, indexerApiAccess1.response_time)
self.assertIsNone(indexerApiAccess1.error)
indexerApiAccess2 = IndexerApiAccess.get(IndexerApiAccess.indexer == Indexer.get(Indexer.name == "newznab2"))
self.assertEqual(indexerSearch2, indexerApiAccess2.indexer_search)
self.assertEqual(18, indexerApiAccess2.time.hour)
self.assertEqual("search", indexerApiAccess2.type)
self.assertEqual(calledUrls[1], indexerApiAccess2.url)
self.assertFalse(indexerApiAccess2.response_successful)
self.assertIsNone(indexerApiAccess2.response_time)
self.assertTrue("Connection refused" in indexerApiAccess2.error)
indexerStatus2 = IndexerStatus.get(IndexerStatus.indexer == Indexer.get(Indexer.name == "newznab2"))
self.assertEqual(1, indexerStatus2.level)
self.assertTrue("Connection refused" in indexerStatus2.reason)
def execute_queries(self, queries, searchRequest):
if len(queries) == 0:
return QueriesExecutionResult(didsearch=False,
results=[],
indexerSearchEntry=None,
indexerApiAccessEntry=None,
indexerStatus=None,
total=0,
loaded_results=0,
total_known=True,
has_more=False,
rejected=self.getRejectedCountDict())
results = []
executed_queries = set()
psearch = IndexerSearch(indexer=self.indexer)
papiaccess = IndexerApiAccess()
indexerStatus = None
total_results = 0
total_known = False
has_more = False
rejected = self.getRejectedCountDict()
while len(queries) > 0:
query = queries.pop()
if query in executed_queries:
# To make sure that in case an offset is reported wrong or we have a bug we don't get stuck in an endless loop
continue
try:
request, papiaccess, indexerStatus = self.get_url_with_papi_access(
query, "search", saveToDb=False)
papiaccess.indexer_search = psearch
executed_queries.add(query)
if request is not None:
if request.text == "":
raise IndexerResultParsingException(
"Indexer returned an empty page", self)
self.check_auth(request.text)
self.debug("Successfully loaded URL %s" % request.url)
try:
parsed_results = self.process_query_result(
request.content, searchRequest)
results.extend(parsed_results.entries
) # Retrieve the processed results
queries.extend(
parsed_results.queries
) # Add queries that were added as a result of the parsing, e.g. when the next result page should also be loaded
total_results += parsed_results.total
total_known = parsed_results.total_known
has_more = parsed_results.has_more
rejected = parsed_results.rejected
papiaccess.response_successful = True
indexerStatus = self.handle_indexer_success(False)
except Exception:
self.exception(
"Error while processing search results from indexer %s"
% self)
raise IndexerResultParsingException(
"Error while parsing the results from indexer",
self)
except IndexerAuthException as e:
papiaccess.error = "Authorization error :%s" % e.message
self.error(papiaccess.error)
indexerStatus = self.handle_indexer_failure(
reason="Authentication failed", disable_permanently=True)
papiaccess.response_successful = False
except IndexerAccessException as e:
papiaccess.error = "Access error: %s" % e.message
self.error(papiaccess.error)
indexerStatus = self.handle_indexer_failure(
reason="Access failed")
papiaccess.response_successful = False
except IndexerResultParsingException as e:
papiaccess.error = "Access error: %s" % e.message
self.error(papiaccess.error)
indexerStatus = self.handle_indexer_failure(
reason="Parsing results failed")
papiaccess.response_successful = False
except Exception as e:
self.exception("An error error occurred while searching: %s",
e)
if papiaccess is not None:
papiaccess.error = "Unknown error :%s" % e
papiaccess.response_successful = False
finally:
if papiaccess is not None:
psearch.successful = papiaccess.response_successful
else:
self.error("Unable to save API response to database")
psearch.resultsCount = total_results
return QueriesExecutionResult(didsearch=True,
results=results,
indexerSearchEntry=psearch,
indexerApiAccessEntry=papiaccess,
indexerStatus=indexerStatus,
total=total_results,
loaded_results=len(results),
total_known=total_known,
has_more=has_more,
rejected=rejected)
def pick_indexers(search_request):
# type: (nzbhydra.search.SearchRequest, bool) -> List[nzbhydra.search_modules.SearchModule]
query_supplied = True if search_request.query else False
queryCanBeGenerated = None # Store if we can generate a query from IDs. Initiall true but when we need this the first time and query generation fails we set it to false
picked_indexers = []
selected_indexers = search_request.indexers.split("|") if search_request.indexers is not None else None
notPickedReasons = {}
for p in indexers.enabled_indexers:
if not p.settings.enabled:
logger.debug("Did not pick %s because it is disabled" % p)
add_not_picked_indexer(notPickedReasons, "Disabled", p.name)
continue
if search_request.internal and p.settings.accessType == "external":
logger.debug("Did not pick %s because it is only enabled for external searches" % p)
add_not_picked_indexer(notPickedReasons, "Disabled for API searches", p.name)
continue
if not search_request.internal and p.settings.accessType == "internal":
logger.debug("Did not pick %s because it is only enabled for internal searches" % p)
add_not_picked_indexer(notPickedReasons, "Disabled for API searches", p.name)
continue
if selected_indexers and p.name not in selected_indexers:
logger.debug("Did not pick %s because it was not selected by the user" % p)
add_not_picked_indexer(notPickedReasons, "Not selected by user", p.name)
continue
try:
status = p.indexer.status.get()
if status.disabled_until and status.disabled_until > arrow.utcnow() and not config.settings.searching.ignoreTemporarilyDisabled:
logger.info("Did not pick %s because it is disabled temporarily due to an error: %s" % (p, status.reason))
add_not_picked_indexer(notPickedReasons, "Temporarily disabled", p.name)
continue
except IndexerStatus.DoesNotExist:
pass
if hasattr(p.settings, "categories") and len(p.settings.categories) > 0:
if search_request.category.category.name != "all" and search_request.category.category.name not in p.settings.categories:
logger.debug("Did not pick %s because it is not enabled for category %s" % (p, search_request.category.category.pretty))
add_not_picked_indexer(notPickedReasons, "Disabled for this category %s" % search_request.category.category.pretty, p.name)
continue
if p.settings.hitLimit > 0:
if p.settings.hitLimitResetTime:
comparisonTime = arrow.utcnow().replace(hour=p.settings.hitLimitResetTime, minute=0, second=0)
if comparisonTime > arrow.utcnow():
comparisonTime = arrow.get(comparisonTime.datetime - datetime.timedelta(days=1)) # Arrow is too dumb to properly subtract 1 day (throws an error on every first of the month)
else:
# Use rolling time window
comparisonTime = arrow.get(arrow.utcnow().datetime - datetime.timedelta(days=1))
apiHits = IndexerApiAccess().select().where((IndexerApiAccess.indexer == p.indexer) & (IndexerApiAccess.time > comparisonTime) & IndexerApiAccess.response_successful).count()
if apiHits >= p.settings.hitLimit:
if p.settings.hitLimitResetTime:
logger.info("Did not pick %s because its API hit limit of %d was reached. Will pick again after %02d:00" % (p, p.settings.hitLimit, p.settings.hitLimitResetTime))
else:
try:
firstHitTimeInWindow = IndexerApiAccess().select().where(IndexerApiAccess.indexer == p.indexer & IndexerApiAccess.response_successful).order_by(IndexerApiAccess.time.desc()).offset(p.settings.hitLimit).limit(1).get().time.datetime
nextHitAfter = arrow.get(firstHitTimeInWindow + datetime.timedelta(days=1))
logger.info("Did not pick %s because its API hit limit of %d was reached. Next possible hit at %s" % (p, p.settings.hitLimit, nextHitAfter.format('YYYY-MM-DD HH:mm')))
except IndexerApiAccess.DoesNotExist:
logger.info("Did not pick %s because its API hit limit of %d was reached" % (p, p.settings.hitLimit))
add_not_picked_indexer(notPickedReasons, "API limit reached", p.name)
continue
else:
logger.debug("%s has had %d of a maximum of %d API hits since %02d:%02d" % (p, apiHits, p.settings.hitLimit, comparisonTime.hour, comparisonTime.minute))
if (query_supplied or search_request.identifier_key is not None) and not p.supports_queries:
logger.debug("Did not pick %s because a query was supplied but the indexer does not support queries" % p)
add_not_picked_indexer(notPickedReasons, "Does not support queries", p.name)
continue
# Here on we check if we could supply the indexer with generated/retrieved data like the title of a series
if not query_supplied and p.needs_queries and search_request.identifier_key is None:
logger.debug("Did not pick %s because no query was supplied but the indexer needs queries" % p)
add_not_picked_indexer(notPickedReasons, "Query needed", p.name)
continue
# If we can theoretically do that we must try to actually get the title, otherwise the indexer won't be able to search
allow_query_generation = (config.InternalExternalSelection.internal in config.settings.searching.generate_queries and search_request.internal) or (config.InternalExternalSelection.external in config.settings.searching.generate_queries and not search_request.internal)
if search_request.identifier_key is not None and not canUseIdKey(p, search_request.identifier_key):
if not (allow_query_generation and p.generate_queries):
logger.debug("Did not pick %s because search will be done by an identifier and the indexer or system wide settings don't allow query generation" % p)
add_not_picked_indexer(notPickedReasons, "Does not support ID based searches", p.name)
continue
else:
if queryCanBeGenerated is None:
try:
title = infos.convertId(search_request.identifier_key, "title", search_request.identifier_value)
if title:
search_request.title = title
queryCanBeGenerated = True
else:
queryCanBeGenerated = False
except:
queryCanBeGenerated = False
logger.debug("Unable to get title for supplied ID. Indexers that don't support the ID will be skipped")
if not queryCanBeGenerated:
logger.debug("Did not pick %s because search will be done by an identifier and retrieval of the title for query generation failed" % p)
add_not_picked_indexer(notPickedReasons, "Does not support ID based searches", p.name)
continue
logger.debug("Picked %s" % p)
picked_indexers.append(p)
if len(picked_indexers) == 0:
warning = "No indexeres were selected for this search:"
for reason, notPickedIndexers in notPickedReasons.items():
warning += "\r\n%s: %s" % (reason, ", ".join(notPickedIndexers))
logger.warn(warning)
return picked_indexers
def testThatDatabaseValuesAreStored(self):
with self.app.test_request_context('/'):
with responses.RequestsMock(assert_all_requests_are_fired=False) as rsps:
newznabItems = [
[mockbuilder.buildNewznabItem(title="title1", pubdate=arrow.get(1000).format("ddd, DD MMM YYYY HH:mm:ss Z"), size=1000, indexer_name="newznab1")],
[mockbuilder.buildNewznabItem(title="title2", pubdate=arrow.get(1000).format("ddd, DD MMM YYYY HH:mm:ss Z"), size=1000, indexer_name="newznab2")]
]
self.prepareSearchMocks(rsps, indexerCount=len(newznabItems), newznabItems=newznabItems)
# Make the second access unsuccessful
rsps._urls.pop(1)
rsps.add(responses.GET, r".*",
body="an error message", status=500,
content_type='application/x-html')
searchRequest = SearchRequest(type="search", query="aquery", category="acategory", identifier_key="imdbid", identifier_value="animdbid", season=1, episode=2, indexers="newznab1|newznab2")
result = search.search(searchRequest)
results = result["results"]
self.assertEqual(1, len(results))
dbSearch = Search().get()
self.assertEqual(True, dbSearch.internal)
self.assertEqual("aquery", dbSearch.query)
self.assertEqual("All", dbSearch.category)
self.assertEqual("imdbid", dbSearch.identifier_key)
self.assertEqual("animdbid", dbSearch.identifier_value)
self.assertEqual("1", dbSearch.season)
self.assertEqual("2", dbSearch.episode)
self.assertEqual("search", dbSearch.type)
self.assertEqual(18, dbSearch.time.hour)
indexerSearch1 = IndexerSearch.get(IndexerSearch.indexer == Indexer.get(Indexer.name == "newznab1"))
self.assertEqual(indexerSearch1.search, dbSearch)
self.assertEqual(18, indexerSearch1.time.hour)
indexerSearch2 = IndexerSearch.get(IndexerSearch.indexer == Indexer.get(Indexer.name == "newznab2"))
self.assertEqual(indexerSearch2.search, dbSearch)
self.assertEqual(18, indexerSearch2.time.hour)
calledUrls = sorted([x.request.url for x in rsps.calls])
indexerApiAccess1 = IndexerApiAccess.get(IndexerApiAccess.indexer == Indexer.get(Indexer.name == "newznab1"))
self.assertEqual(indexerSearch1, indexerApiAccess1.indexer_search)
self.assertEqual(18, indexerApiAccess1.time.hour)
self.assertEqual("search", indexerApiAccess1.type)
self.assertEqual(calledUrls[0], indexerApiAccess1.url)
self.assertTrue(indexerApiAccess1.response_successful)
self.assertEqual(0, indexerApiAccess1.response_time)
self.assertIsNone(indexerApiAccess1.error)
indexerApiAccess2 = IndexerApiAccess.get(IndexerApiAccess.indexer == Indexer.get(Indexer.name == "newznab2"))
self.assertEqual(indexerSearch2, indexerApiAccess2.indexer_search)
self.assertEqual(18, indexerApiAccess2.time.hour)
self.assertEqual("search", indexerApiAccess2.type)
self.assertEqual(calledUrls[1], indexerApiAccess2.url)
self.assertFalse(indexerApiAccess2.response_successful)
self.assertIsNone(indexerApiAccess2.response_time)
self.assertTrue("Connection refused" in indexerApiAccess2.error)
indexerStatus2 = IndexerStatus.get(IndexerStatus.indexer == Indexer.get(Indexer.name == "newznab2"))
self.assertEqual(1, indexerStatus2.level)
self.assertTrue("Connection refused" in indexerStatus2.reason)
