skrecsys |
RecommenderMixin, is_recommender, supports_partial_fit |
skrecsys.base |
ConditionMixin, FeaturesMixin, RankerMixin, Not, AllOf, AnyOf, is_condition, is_features, is_ranker, serves_unknown_users |
skrecsys.compose |
Switch, Cascade, KnownUser, MinInteractions, QueryIn, Not, AllOf, AnyOf, JoinStaticFeatures, JoinDynamicFeatures, GeneratorScores, InteractionCounts, RecommenderScores, SegmentPopularity, ProfileAffinity, ConcatFeatures, PointwiseRanker, GroupRanker, BlendRanker, AugmentedRanker, ReciprocalRankFusion, ReciprocalRankRanker, Backfill, ReservedSlots, ItemListRecommender |
skrecsys.exceptions |
InsufficientDataError |
skrecsys.typing |
Protocols for estimators and composite components (Recommender, FittedRecommender, Ranker, Features, Condition, ...), RandomStateLike, RankingMetric, clone_as, and the override/TypeIs backports |
skrecsys.integrations.catboost |
CatBoostRanker; third-party integration, requires skrecsys[catboost] |
skrecsys.integrations.xgboost |
XGBRanker; third-party integration, requires skrecsys[xgboost] |
skrecsys.integrations.lightgbm |
LGBMRanker; third-party integration, requires skrecsys[lightgbm] |
skrecsys.integrations.batching |
MicroBatcher, AsyncRecommender; asyncio micro-batching of concurrent recommend calls, no extra |
skrecsys.metrics |
precision_at_k, recall_at_k, ndcg_at_k, average_precision_at_k, reciprocal_rank_at_k, hit_rate_at_k, make_recommender_scorer, evaluate_recommender |
skrecsys.metrics |
catalog_coverage_at_k, user_coverage_at_k, mean_popularity_at_k, novelty_at_k, item_popularity |
skrecsys.datasets |
fetch_movielens_100k, fetch_movielens_1m, fetch_amazon_books, get_data_home, clear_data_home |
skrecsys.model_selection |
WarmStartKFold, ColdStartSplit, LatestInteractionsSplit |
skrecsys.tune |
AutoTune, Study, Trial, Float, Int, Categorical, search_space |
skrecsys.recommendation |
MostPopularRecommender, ItemKNNRecommender, AlternatingLeastSquares, BM25Recommender, EASE, RP3Beta, SLIMElasticNet, BayesianPersonalizedRanking |
skrecsys.nn |
SimpleX, XSimGCL, HSTU (sequential), Mamba4Rec (sequential); requires skrecsys[nn] |
skrecsys.inspection |
trace, Trace, QueryTrace and its steps, explain, Explanation, Retrieval, RankerDecision, LeafReasons, and the Level, Sample and Status types |
skrecsys.indexing |
HNSW, QuantizedFlatIndex, VectorIndex, VectorIndexMixin, available_indexes |