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 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 search(internal, search_request):
for k in list(pseudo_cache.keys()):
if pseudo_cache[k]["last_access"].replace(minutes=+5) < arrow.utcnow():
pseudo_cache.pop(k)
limit = search_request.limit # todo use actual configured limit
external_offset = int(search_request.offset)
search_hash = search_request.search_hash
if search_hash not in pseudo_cache.keys() or search_request.offset == 0: #If it's a new search (which starts with offset 0) do it again instead of using the cached results
logger.debug("Didn't find this query in cache or want to do a new search")
cache_entry = {"results": [], "indexer_infos": {}, "total": 0, "last_access": arrow.utcnow(), "offset": 0}
indexers_to_call, with_query_generation = pick_indexers(query_supplied=True if search_request.query is not None and search_request.query != "" else False, identifier_key=search_request.identifier_key, internal=internal, selected_indexers=search_request.indexers)
for p in indexers_to_call:
cache_entry["indexer_infos"][p] = {"has_more": True, "search_request": search_request, "total_included": False}
dbsearch = Search(internal=internal, query=search_request.query, category=search_request.category, identifier_key=search_request.identifier_key, identifier_value=search_request.identifier_value, season=search_request.season, episode=search_request.episode, type=search_request.type)
dbsearch.save()
cache_entry["dbsearch"] = dbsearch
if with_query_generation and search_request.identifier_key and search_request.title is None:
search_request.title = infos.title_from_id(search_request.identifier_key, search_request.identifier_value)
pseudo_cache[search_hash] = cache_entry
else:
cache_entry = pseudo_cache[search_hash]
indexers_to_call = [indexer for indexer, info in cache_entry["indexer_infos"].items() if info["has_more"]]
dbsearch = cache_entry["dbsearch"]
logger.debug("Found search in cache")
logger.debug("Will search at indexers as long as we don't have enough results for the current offset+limit and any indexer has more results.")
while len(cache_entry["results"]) < external_offset + limit and len(indexers_to_call) > 0:
logger.debug("We want %d results but have only %d so far" % ((external_offset + limit), len(cache_entry["results"])))
logger.debug("%d indexers still have results" % len(indexers_to_call))
search_request.offset = cache_entry["offset"]
logger.debug("Searching indexers with offset %d" % search_request.offset)
result = search_and_handle_db(dbsearch, {x: search_request for x in indexers_to_call})
search_results = []
indexers_to_call = []
for indexer, queries_execution_result in result["results"].items():
search_results.extend(queries_execution_result.results)
logger.debug("%s returned %d results" % (indexer, len(queries_execution_result.results)))
cache_entry["indexer_infos"][indexer].update({"search_request": search_request, "has_more": queries_execution_result.has_more, "total": queries_execution_result.total, "total_known": queries_execution_result.total_known, "indexer_search": queries_execution_result.dbentry})
if queries_execution_result.has_more:
indexers_to_call.append(indexer)
logger.debug("%s still has more results so we could use it the next round" % indexer)
if queries_execution_result.total_known:
if not cache_entry["indexer_infos"][indexer]["total_included"]:
cache_entry["total"] += queries_execution_result.total
logger.debug("%s reports %d total results. We'll include in the total this time only" % (indexer, queries_execution_result.total))
cache_entry["indexer_infos"][indexer]["total_included"] = True
elif queries_execution_result.has_more:
logger.debug("%s doesn't report an exact number of results so let's just add another 100 to the total" % indexer)
cache_entry["total"] += 100
search_results = sorted(search_results, key=lambda x: x.epoch, reverse=True)
cache_entry["results"].extend(search_results)
cache_entry["offset"] += limit
#todo: perhaps move duplicate handling here. WOuld allow to recognize duplicates that were added, for example 100 were already loaded and then we get 101-200 und 100 and 101 are duplicates
#todo: then make configurable if we want to delete duplicates for api, internal, both, none. would also mean that we return 100 actually different results, otherwise in the worst case we could for example return 50 originals and 50 duplicates
if internal:
logger.debug("We have %d cached results and them all because we search internally" % len(cache_entry["results"]))
nzb_search_results = copy.deepcopy(cache_entry["results"][external_offset:])
else:
logger.debug("We have %d cached results and return %d-%d of %d total available accounting for the limit set for the API search" % (len(cache_entry["results"]), external_offset, external_offset + limit, cache_entry["total"]))
nzb_search_results = copy.deepcopy(cache_entry["results"][external_offset:(external_offset + limit)])
cache_entry["last_access"] = arrow.utcnow()
return {"results": nzb_search_results, "indexer_infos": cache_entry["indexer_infos"], "dbsearch": cache_entry["dbsearch"].id, "total": cache_entry["total"], "offset": external_offset}
