dh=0T%SrSSKrSSKrSSKrSSKJrJr SSKJrJ r J r J r J r J r JrJr SSKJr SSKJr SSKJr SSKJr SS KJr SS KJr S rS r"S S\5r"SS\5r"SS\5r "SS\5r!\\ \!S.r"\ \#\\4\$S'"SS\%5r&"SS\5r'g)zWrapper around scikit-learn NearestNeighbors implementation. The vector store can be persisted in json, bson or parquet format. N)ABCabstractmethod)AnyDictIterableListLiteralOptionalTupleType)uuid4)Document) Embeddings) guard_import) VectorStore)maximal_marginal_relevancec\rSrSrSrS\SS4Sjr\\S\4Sj55r \S\ SS4S j5r \S\ 4S j5r S r g) BaseSerializerz Base class for serializing data. persist_pathreturnNcXlgNr)selfrs `/var/www/html/shao/venv/lib/python3.13/site-packages/langchain_community/vectorstores/sklearn.py__init__BaseSerializer.__init__s(cg)z>The file extension suggested by this serializer (without dot).Nclss r extensionBaseSerializer.extensionr!datacg)z"Saves the data to the persist_pathNr#rr)s rsaveBaseSerializer.save#r(r!cg)z$Loads the data from the persist_pathNr#rs rloadBaseSerializer.load'r(r!r)__name__ __module__ __qualname____firstlineno____doc__strr classmethodrr&rr,r0__static_attributes__r#r!rrrs{*)S)T)M#MM11113c33r!rcP\rSrSrSr\S\4Sj5rS\SS4Sjr S\4Sjr S r g) JsonSerializer,zKSerialize data in JSON using the json package from python standard library.rcg)Njsonr#r$s rr&JsonSerializer.extension/r!r)Nc[URS5n[R"X5 SSS5 g!,(df  g=f)Nw)openrr>dumprr)fps rr,JsonSerializer.save3s, $##S )R IId * ) )s 7 Ac[URS5n[R"U5sSSS5 $!,(df  g=f)Nr)rCrr>r0rrFs rr0JsonSerializer.load7s+ $##S )R99R=* ) )s 7 Ar#) r2r3r4r5r6r8r7r&rr,r0r9r#r!rr;r;,s=U#   !c!r!r;cp^\rSrSrSrS\SS4U4Sjjr\S\4Sj5rS\ SS4S jr S\ 4S jr S r U=r $) BsonSerializer<z>Serialize data in Binary JSON using the `bson` python package.rrNcD>[TU]U5 [S5UlgNbson)superrrrQrr __class__s rrBsonSerializer.__init__?s & ( r!cgrPr#r$s rr&BsonSerializer.extensionCr@r!r)c[URS5nURURR U55 SSS5 g!,(df  g=f)Nwb)rCrwriterQdumpsrEs rr,BsonSerializer.saveGs9 $##T *b HHTYY__T* ++ * *s +A  Ac[URS5nURRUR 55sSSS5 $!,(df  g=f)Nrb)rCrrQloadsreadrJs rr0BsonSerializer.loadKs6 $##T *b99??2779-+ * *s )A  A)rQr2r3r4r5r6r7rr8r&rr,r0r9 __classcell__rTs@rrMrM<sXH)S)T)#,,,.c..r!rMcp^\rSrSrSrS\SS4U4Sjjr\S\4Sj5rS\ SS4S jr S\ 4S jr S r U=r $) ParquetSerializerPzFSerialize data in `Apache Parquet` format using the `pyarrow` package.rrNc>[TU]U5 [S5Ul[S5Ul[S5Ulg)Npandaspyarrowzpyarrow.parquet)rRrrpdpapqrSs rrParquetSerializer.__init__Ss5 &x(y)01r!cg)Nparquetr#r$s rr&ParquetSerializer.extensionYsr!r)c~URRU5nURRR U5n[ R RUR5(av[UR5S-n[ R"URU5 URRX0R5 [ R"U5 gURRX0R5 g![a'n[ R"X@R5 UeSnAff=f)Nz-backup)rk DataFramerlTable from_pandasospathexistsrr7renamerm write_tableremove Exception)rr)dftable backup_pathexcs rr,ParquetSerializer.save]s WW  t $ ))"- 77>>$++ , ,d//09rQrpSERIALIZER_MAPc\rSrSrSrSrg)SKLearnVectorStoreExceptionzz'Exception raised by SKLearnVectorStore.r#N)r2r3r4r5r6r9r#r!rrrzs1r!rcZ\rSrSrSrSSSS.S\S\\S \S S \S \ S S4 Sjjr \ S \4Sj5r S&Sjr S&SjrS'S\\S\\\S\\\S \ S \\4 SjjrS&Sjr\S.S\\S\S \ S \\\\44Sjjr\S.S\S\S \ S \\\\44Sjjr\4S\S\S \ S \\4Sjjr\4S\S\S \ S \\\\44Sjjr\\S4S\\S\S \S!\S \ S \\4 S"jjr\\S4S\S\S \S!\S \ S \\4 S#jjr\ S(S\\S\S\\\S\\\S\\S \ S S4S$jj5r!S%r"g))SKLearnVectorStorezUSimple in-memory vector store based on the `scikit-learn` library `NearestNeighbors`.Nr>cosine)r serializermetric embeddingrrrrkwargsrc [S5n[SSS9nX`lUR"SSU0UD6UlSUlXlX lSUlUR b[UnU"UR S9Ul/Ul /Ul /Ul /Ul UR/5UlUR b@[R R#UR 5(aUR%5 ggg) Nnumpyzsklearn.neighborsz scikit-learn)pip_namerFrr#)r_npNearestNeighbors _neighbors_neighbors_fitted_embedding_function _persist_path _serializerr _embeddings_texts _metadatas_idsasarray_embeddings_nprvrwisfile_load) rrrrrrnpsklearn_neighborsserializer_clss rrSKLearnVectorStore.__init__s' "()<~V+<<UFUfU!&#, )59    )+J7N-4;M;MND /1!# &(! $&::b>    )bggnnT=O=O.P.P JJL/Q )r!cUR$r)rr/s r embeddingsSKLearnVectorStore.embeddingss'''r!cURc [S5eURURURUR S.nURR U5 g)NzFYou must specify a persist_path on creation to persist the collection.)idstexts metadatasr)rrrrrrr,r+s rpersistSKLearnVectorStore.persists[    #-X 99[[**   d#r!cURc [S5eURR5nUSUlUSUlUSUlUSUlUR5 g)NzCYou must specify a persist_path on creation to load the collection.rrrr)rrr0rrrr_update_neighborsr+s rrSKLearnVectorStore._loadso    #-U $$& -7m {+K   r!rrrc [U5nU=(d# UVs/sHn[[55PM snnURR U5 UR R UR RU55 URR U=(d 0/[U5-5 URR U5 UR5 U$s snfr) listr7r rextendrrembed_documentsrlenrr)rrrrrr_rs r add_textsSKLearnVectorStore.add_textsse4V4Vs57|V4 6"  8 8 H H PQ y@bTCK-?A    5sC%c[UR5S:Xa [S5eURR UR5UlUR RUR 5 SUlg)Nr(No data was added to SKLearnVectorStore.T) rrrrrrrfitrr/s rr$SKLearnVectorStore._update_neighborssd t A %-: #hh..t/?/?@ D//0!%r!)kquery_embeddingrc UR(d [S5eURRU/US9upE[ [ USUS55$)z_Search k embeddings similar to the query embedding. Returns a list of (index, distance) tuples.r) n_neighborsr)rrr kneighborsrzip)rrrr neigh_dists neigh_idxss r#_similarity_index_search_with_score6SKLearnVectorStore._similarity_index_search_with_scores_ %%-: #'//"<"<  1#=#  C 1 {1~677r!queryc URRU5nUR"U4SU0UD6nUVVs/sH<upg[URUSUR U0UR UES9U4PM> snn$s snnf)Nrid page_contentmetadata)r embed_queryrrrrr)rrrrr indices_distsidxdists rsimilarity_search_with_score/SKLearnVectorStore.similarity_search_with_scores22>>uE@@    $* +  +  !%S!1"DIIcNKdooc6JK  +   sAA<c dUR"U4SU0UD6nUVVs/sHupVUPM snn$s snnf)Nr)r)rrrr docs_scoresdocrs rsimilarity_search$SKLearnVectorStore.similarity_searchs877MMfM "-.++...s ,c UR"U4SU0UD6n[U6upVUVs/sHnS[R"U5- PM nn[ [[ U5U55$s snf)Nr)rrmathexpr) rrrr docs_distsdocsdistsrscoress r(_similarity_search_with_relevance_scores;SKLearnVectorStore._similarity_search_with_relevance_scoressf66uLLVL :& 167!dhhtn$7CT F+,,8s"A&g?fetch_k lambda_multc UR"U4SU0UD6n[U6upxURU4n [URR XRR S9U UUS9n U V s/sHoU PM n n U V s/sH8n [URU SURU 0URU ES9PM: sn $s sn fs sn f)aReturn docs selected using the maximal marginal relevance. Maximal marginal relevance optimizes for similarity to query AND diversity among selected documents. Args: embedding: Embedding to look up documents similar to. k: Number of Documents to return. Defaults to 4. fetch_k: Number of Documents to fetch to pass to MMR algorithm. lambda_mult: Number between 0 and 1 that determines the degree of diversity among the results with 0 corresponding to maximum diversity and 1 to minimum diversity. Defaults to 0.5. Returns: List of Documents selected by maximal marginal relevance. r)dtype)rrrr) rrrrrarrayfloat32rrrr)rrrrrrrindicesrresult_embeddings mmr_selectedi mmr_indicesrs r'max_marginal_relevance_search_by_vector:SKLearnVectorStore.max_marginal_relevance_search_by_vector s,@@    $* -(  //91 HHNN9HH,<,>Cc URc [S5eURRU5nURXbX4S9nU$)aReturn docs selected using the maximal marginal relevance. Maximal marginal relevance optimizes for similarity to query AND diversity among selected documents. Args: query: Text to look up documents similar to. k: Number of Documents to return. Defaults to 4. fetch_k: Number of Documents to fetch to pass to MMR algorithm. lambda_mult: Number between 0 and 1 that determines the degree of diversity among the results with 0 corresponding to maximum diversity and 1 to minimum diversity. Defaults to 0.5. Returns: List of Documents selected by maximal marginal relevance. zCFor MMR search, you must specify an embedding function on creation.) lambda_mul)r ValueErrorrr)rrrrrrrrs rmax_marginal_relevance_search0SKLearnVectorStore.max_marginal_relevance_search5sY,  # # +U ,,88? ;; '<  r!c B[U4SU0UD6nURXUS9 U$)Nr)rr)rr)r%rrrrrrvss r from_textsSKLearnVectorStore.from_textsVs- O O O US 9 r!) rrrrrrrrrrr)rN)NN)NNN)#r2r3r4r5r6rr r7r rrpropertyrrrrrdictrr DEFAULT_Kfloatintr rrrrrDEFAULT_FETCH_Krrr8rr9r#r!rrrs'+9? !!sm ! 56 !  !! !F(J(( $ !+/#' }DJ'd3i    c  &9B 8#E{ 825 8MP 8 eCJ  8'0  # ;> eHeO$ % &$-// /8;/ h/$--- -8;- eHeO$ %-& ( ;(  (  (  (  (  h( Z&        hB +/#'&* Cy  DJ'  d3i  sm     r!r)(r6r>rrvabcrrtypingrrrrr r r r uuidr langchain_core.documentsrlangchain_core.embeddingsrlangchain_core.utilsrlangchain_core.vectorstoresr&langchain_community.vectorstores.utilsrrrrr;rMrfrr7__annotations__ RuntimeErrorrrr#r!rrs  #LLL-0-3M 3S3( !^ ! .^.( D DH   3S$~../ , bbr!