ZiuW"ddlmZmZmZmZmZmZmZmZddl m Z ddl m Z m Z mZmZmZmZmZmZmZmZmZmZmZmZmZmZmZmZmZddlm Z ddl!m"Z"ddl#Z#ddl$m%Z%erddl&m'Z'e#j(e)Z*erdd l+m,Z,Gd d e d Z-dS) ) TYPE_CHECKINGOptionalUnionListcastDictAnyTuple)CollectionCommon)URICollectionMetadata Embedding PyEmbeddingIncludeIndexingStatusMetadataDocumentImageWhereIDs GetResult QueryResultID OneOrMany ReadLevel WhereDocument SearchResultmaybe_cast_one_to_many)UpdateCollectionConfiguration)SearchN)Function)AttachedFunction) ServerAPIcTeZdZdefdZdefdZ d0deede e ee ee fde ee de eed e eed e eeddfd Zdddddddgfde eed e ed e ede ede ededefdZd1d edefdZddddddddgdf de e ee ee fde eede eede eede eeded e ede ededefdZ d2de ede ede eddfdZdeddfd Zej fd!ee!d"ede"fd#Z# d0deede e ee ee fde ee de eed e eed e eeddfd$Z$ d0deede e ee ee fde ee de eed e eed e eeddfd%Z% d2de e&d e ede eddfd&Z' d3d'e(ded(ed)e e)ee*fde+d*e,ff d+Z-dedd*fd,Z. d4ded.e,de,fd/Z/dS)5 CollectionreturncZ|j|j|j|jS)z/Return the number of records in the collection. collection_idtenantdatabase)_client_countidr*r+selfs [/root/projects/butler/venv/lib64/python3.11/site-packages/chromadb/api/models/Collection.pycountzCollection.count*s0|""';]#   cZ|j|j|j|jS)aGet the indexing status of this collection. Returns: IndexingStatus: An object containing: - num_indexed_ops: Number of user operations that have been indexed - num_unindexed_ops: Number of user operations pending indexing - total_ops: Total number of user operations in collection - op_indexing_progress: Proportion of user operations that have been indexed as a float between 0 and 1 r()r,_get_indexing_statusr.r*r+r/s r1get_indexing_statuszCollection.get_indexing_status2s2|00';]1   r3Nids embeddings metadatas documentsimagesurisc |||||||}|j|j|d|d|d|d|d|j|jdS) aAdd records to the collection. Args: ids: Record IDs to add. embeddings: Embeddings to add. If None, embeddings are computed. metadatas: Optional metadata for each record. documents: Optional documents for each record. images: Optional images for each record. uris: Optional URIs for loading images. Raises: ValueError: If embeddings and documents are both missing. ValueError: If embeddings and documents are both provided. ValueError: If lengths of provided fields do not match. ValueError: If an ID already exists. r7r8r9r:r;r<r7r8r9r:r<r)r7r8r9r:r<r*r+N)!_validate_and_prepare_add_requestr,_addr.r*r+)r0r7r8r9r:r;r< add_requests r1addzCollection.addBs><<! =   'E""<0!+.!+.V$;]  r3wherelimitoffsetwhere_documentincludec |||||}|j|j|d|d|d|d|||j|j }|||dS)aRetrieve records from the collection. If no filters are provided, returns records up to ``limit`` starting at ``offset``. Args: ids: If provided, only return records with these IDs. where: A Where filter used to filter based on metadata values. limit: Maximum number of results to return. offset: Number of results to skip before returning. where_document: A WhereDocument filter used to filter based on K.DOCUMENT. include: Fields to include in results. Can contain "embeddings", "metadatas", "documents", "uris". Defaults to "metadatas" and "documents". Returns: GetResult: Retrieved records and requested fields as a GetResult object. )r7rDrGrHr7rDrGrH) r)r7rDrGrHrErFr*r+responserH)!_validate_and_prepare_get_requestr,_getr.r*r+_transform_get_response) r0r7rDrErFrGrH get_request get_resultss r1getzCollection.getus2<<) =  l'''E"g&&'78 *;](   ++ +i*@,   r3 c||j|j||j|jS)zReturn the first ``limit`` records from the collection. Args: limit: Maximum number of records to return. Returns: GetResult: Retrieved records and requested fields. )r)nr*r+)_transform_peek_responser,_peekr.r*r+)r0rEs r1peekzCollection.peeksH,, L  "g{      r3)r9r: distancesquery_embeddings query_texts query_images query_uris n_resultsc $||||||||||  } |j|j| d| d| d| d| d| d|j|j } || | d S) aQuery for the K nearest neighbor records in the collection. This is a batch query API. Multiple queries can be performed at once by providing multiple embeddings, texts, or images. >>> query_1 = [0.1, 0.2, 0.3] >>> query_2 = [0.4, 0.5, 0.6] >>> results = collection.query( >>> query_embeddings=[query_1, query_2], >>> n_results=10, >>> ) If query_texts, query_images, or query_uris are provided, the collection's embedding function will be used to create embeddings before querying the API. The `ids`, `where`, `where_document`, and `include` parameters are applied to all queries. Args: query_embeddings: Raw embeddings to query for. query_texts: Documents to embed and query against. query_images: Images to embed and query against. query_uris: URIs to be loaded and embedded. ids: Optional subset of IDs to search within. n_results: Number of neighbors to return per query. where: Metadata filter. where_document: Document content filter. include: Fields to include in results. Can contain "embeddings", "metadatas", "documents", "uris", "distances". Defaults to "metadatas", "documents", "distances". Returns: QueryResult: Nearest neighbor results. Raises: ValueError: If no query input is provided. ValueError: If multiple query input types are provided. ) rYrZr[r\r7r]rDrGrHr7r8r]rDrGrH) r)r7rYr]rDrGrHr*r+rJ)#_validate_and_prepare_query_requestr,_queryr.r*r+_transform_query_response) r0rYrZr[r\r7r]rDrGrH query_request query_resultss r1queryzCollection.querysv@@-#%!)A    ++'e$*<8#K0(()9:!),;],   --"M),D.   r3namemetadata configurationc|||j|j||||j|j||||dS)zUpdate collection name, metadata, or configuration. Args: name: New collection name. metadata: New metadata for the collection. configuration: New configuration for the collection. )r.new_name new_metadatanew_configurationr*r+N)_validate_modify_requestr,_modifyr.r*r+"_update_model_after_modify_success)r0rerfrgs r1modifyzCollection.modify sq %%h/// w!+;]     //h NNNNNr3ric|j|j||j|j}t |j||j|jS)aFork the current collection under a new name. The returning collection should contain identical data to the current collection. This is an experimental API that only works for Hosted Chroma for now. Args: new_name: The name of the new collection. Returns: Collection: A new collection with the specified name and containing identical data to the current collection. )r)rir*r+)clientmodelembedding_function data_loader)r,_forkr.r*r+r%_embedding_function _data_loader)r0rirrs r1forkzCollection.fork+s_ ""';] #   <#7)     r3searches read_levelct|}|g}fd|D}jjt t t |jj|S)a Perform hybrid search on the collection. This is an experimental API that only works for distributed and hosted Chroma for now. Args: searches: A single Search object or a list of Search objects, each containing: - where: Where expression for filtering - rank: Ranking expression for hybrid search (defaults to Val(0.0)) - limit: Limit configuration for pagination (defaults to no limit) - select: Select configuration for keys to return (defaults to empty) read_level: Controls whether to read from the write-ahead log (WAL): - ReadLevel.INDEX_AND_WAL: Read from both the compacted index and WAL (default). All committed writes will be visible. - ReadLevel.INDEX_ONLY: Read only from the compacted index, skipping the WAL. Faster, but recent writes that haven't been compacted may not be visible. Returns: SearchResult: Column-major format response with: - ids: List of result IDs for each search payload - documents: Optional documents for each payload - embeddings: Optional embeddings for each payload - metadatas: Optional metadata for each payload - scores: Optional scores for each payload - select: List of selected keys for each payload Raises: NotImplementedError: For local/segment API implementations Examples: # Using builder pattern with Key constants from chromadb.execution.expression import ( Search, Key, K, Knn, Val ) # Note: K is an alias for Key, so K.DOCUMENT == Key.DOCUMENT search = (Search() .where((K("category") == "science") & (K("score") > 0.5)) .rank(Knn(query=[0.1, 0.2, 0.3]) * 0.8 + Val(0.5) * 0.2) .limit(10, offset=0) .select(K.DOCUMENT, K.SCORE, "title")) # Direct construction from chromadb.execution.expression import ( Search, Eq, And, Gt, Knn, Limit, Select, Key ) search = Search( where=And([Eq("category", "science"), Gt("score", 0.5)]), rank=Knn(query=[0.1, 0.2, 0.3]), limit=Limit(offset=0, limit=10), select=Select(keys={Key.DOCUMENT, Key.SCORE, "title"}) ) # Single search result = collection.search(search) # Multiple searches at once searches = [ Search().where(K("type") == "article").rank(Knn(query=[0.1, 0.2])), Search().where(K("type") == "paper").rank(Knn(query=[0.3, 0.4])) ] results = collection.search(searches) # Skip WAL for faster queries (may miss recent uncommitted writes) from chromadb.api.types import ReadLevel result = collection.search(search, read_level=ReadLevel.INDEX_ONLY) Nc:g|]}|S)_embed_search_string_queries).0searchr0s r1 z%Collection.search..s4   :@D - -f 5 5   r3)r)ryr*r+rz) rr,_searchr.rrr r*r+)r0ryrz searches_listembedded_searchess` r1rzCollection.searchEsP/x88  M    DQ   |##'$v,(9::;]! $   r3c |||||||}|j|j|d|d|d|d|d|j|jdS) a-Update existing records by ID. Records are provided in columnar format. If provided, the `embeddings`, `metadatas`, `documents`, and `uris` lists must be the same length. Entries in each list correspond to the same record. >>> ids = ["id1", "id2", "id3"] >>> embeddings = [[0.1, 0.2, 0.3], [0.4, 0.5, 0.6], [0.7, 0.8, 0.9]] >>> metadatas = [{"key": "value"}, {"key": "value"}, {"key": "value"}] >>> documents = ["document1", "document2", "document3"] >>> uris = ["uri1", "uri2", "uri3"] >>> collection.update(ids, embeddings, metadatas, documents, uris) If `embeddings` are not provided, the embeddings will be computed based on `documents` using the collection's embedding function. Args: ids: Record IDs to update. embeddings: Updated embeddings. If None, embeddings are computed. metadatas: Updated metadata. documents: Updated documents. images: Updated images. uris: Updated URIs for loading images. r>r7r8r9r:r<r?N)$_validate_and_prepare_update_requestr,_updater.r*r+)r0r7r8r9r:r;r<update_requests r1updatezCollection.updatesHBB! C   'u%%l3$[1$[1';]  r3c |||||||}|j|j|d|d|d|d|d|j|jdS) arCreate or update records by ID. Args: ids: Record IDs to upsert. embeddings: Embeddings to add or update. If None, embeddings are computed. metadatas: Metadata to add or update. documents: Documents to add or update. images: Images to add or update. uris: URIs for loading images. r>r7r8r9r:r<r?N)$_validate_and_prepare_upsert_requestr,_upsertr.r*r+)r0r7r8r9r:r;r<upsert_requests r1upsertzCollection.upserts0BB! C   'u%%l3$[1$[1';]  r3c||||}|j|j|d|d|d|j|jdS)aiDelete records by ID or filters. All documents that match the `ids` or `where` and `where_document` filters will be deleted. Args: ids: Record IDs to delete. where: Metadata filter. where_document: Document content filter. Raises: ValueError: If no IDs or filters are provided. r7rDrG)r)r7rDrGr*r+N)$_validate_and_prepare_delete_requestr,_deleter.r*r+)r0r7rDrGdelete_requests r1deletezCollection.deletess$BB    'u% ))*:;;]      r3functionoutput_collectionparamsr"c t|tr|jn|}|j|||j|||j|jS)aAttach a function to this collection. Args: function: A Function enum value (e.g., STATISTICS_FUNCTION, RECORD_COUNTER_FUNCTION) name: Unique name for this attached function output_collection: Name of the collection where function output will be stored params: Optional dictionary with function-specific parameters Returns: Tuple of (AttachedFunction, created) where created is True if newly created, False if already existed (idempotent request) Example: >>> from chromadb.api.functions import STATISTICS_FUNCTION >>> attached_fn = collection.attach_function( ... function=STATISTICS_FUNCTION, ... name="mycoll_stats_fn", ... output_collection="mycoll_stats", ... ) >>> if created: ... print("New function attached") ... else: ... print("Function already existed") ) function_idreinput_collection_idrrr*r+) isinstancer!valuer,attach_functionr.r*r+)r0rrerrrs r1rzCollection.attach_function!sX>)38X(F(FThnnH |++# $/;],   r3c\|j||j|j|jS)aGet an attached function by name for this collection. Args: name: Name of the attached function Returns: AttachedFunction: The attached function object Raises: NotFoundError: If the attached function doesn't exist )rerr*r+)r,get_attached_functionr.r*r+)r0res r1rz Collection.get_attached_functionKs5|11 $;] 2   r3Fdelete_output_collectionc^|j||j||j|jS)aDetach a function from this collection. Args: name: The name of the attached function delete_output_collection: Whether to also delete the output collection. Defaults to False. Returns: bool: True if successful Example: >>> success = collection.detach_function("my_function", delete_output_collection=True) )rer delete_outputr*r+)r,detach_functionr.r*r+)r0rers r1rzCollection.detach_function^s8"|++ $2;] ,   r3)NNNNN)rR)NNN)N)F)0__name__ __module__ __qualname__intr2rr6rrrrrrrrrr rCrrrrrQrWrrdstrr rrorxr INDEX_AND_WALr rrrrrrr!rr r boolrrrr}r3r1r%r%)s+ s     ^    0 3737-1)-1 1 r]1  )$+&(  1 Ih/01 Ih/01 5)*1 y~&1  1 1 1 1 j(,!%# $26'5- - im $- - } -  - !/ - -  - - - - ^  # y    2 5937/3'+!%26   U U " )$+&(  U i12U y/0U Ys^,U im $U U U !/U U ( )U U U U r#15AE OOsmO-.O => O  OOOO<       :!* 7W W F#W W   W W W W B 3737-1)-6 6 r]6  )$+&(  6 Ih/06 Ih/06 5)*6 y~&6  6 6 6 6 @ 3737-1)-* * r]*  )$+&(  * Ih/0* Ih/0* 5)** y~&*  * * * * \"!%26   c]  !/       H,0 ( ( ( (  ( c3h( ( !4' ( ( ( ( ( T # 2D    ,*/   #'         r3r%r#).typingrrrrrrr r $chromadb.api.models.CollectionCommonr chromadb.api.typesr r rrrrrrrrrrrrrrrrr%chromadb.api.collection_configurationr"chromadb.execution.expression.planr loggingchromadb.api.functionsr!$chromadb.api.models.AttachedFunctionr" getLoggerrlogger chromadb.apir#r%r}r3r1rsOOOOOOOOOOOOOOOOOOOOAAAAAA*POOOOO555555++++++FEEEEEE  8 $ $'&&&&&&L  L  L  L  L  !+.L  L  L  L  L  r3