chromadb 2.2.1 → 2.3.0

package/dist/cjs/chromadb.cjs

@@ -6836,7 +6836,7 @@ __export(src_exports, {
 module.exports = __toCommonJS(src_exports);
 
 // ../chromadb-core/package.json
-var version = "2.2.1";
+var version = "2.3.0";
 
 // ../chromadb-core/src/generated/runtime.ts
 var import_isomorphic_fetch = require("isomorphic-fetch");
@@ -7568,6 +7568,67 @@ var ApiApiFetchParamCreator = function(configuration) {
         options: localVarRequestOptions
       };
     },
+    /**
+     * @summary Forks an existing collection.
+     * @param {string} tenant <p>Tenant ID</p>
+     * @param {string} database <p>Database name</p>
+     * @param {string} collectionId <p>UUID of the collection to update</p>
+     * @param {Api.ForkCollectionPayload} request
+     * @param {RequestInit} [options] Override http request option.
+     * @throws {RequiredError}
+     */
+    forkCollection(tenant, database, collectionId, request, options = {}) {
+      if (tenant === null || tenant === void 0) {
+        throw new RequiredError(
+          "tenant",
+          "Required parameter tenant was null or undefined when calling forkCollection."
+        );
+      }
+      if (database === null || database === void 0) {
+        throw new RequiredError(
+          "database",
+          "Required parameter database was null or undefined when calling forkCollection."
+        );
+      }
+      if (collectionId === null || collectionId === void 0) {
+        throw new RequiredError(
+          "collectionId",
+          "Required parameter collectionId was null or undefined when calling forkCollection."
+        );
+      }
+      if (request === null || request === void 0) {
+        throw new RequiredError(
+          "request",
+          "Required parameter request was null or undefined when calling forkCollection."
+        );
+      }
+      let localVarPath = `/api/v2/tenants/{tenant}/databases/{database}/collections/{collection_id}/fork`.replace("{tenant}", encodeURIComponent(String(tenant))).replace("{database}", encodeURIComponent(String(database))).replace("{collection_id}", encodeURIComponent(String(collectionId)));
+      const localVarPathQueryStart = localVarPath.indexOf("?");
+      const localVarRequestOptions = Object.assign(
+        { method: "POST" },
+        options
+      );
+      const localVarHeaderParameter = options.headers ? new Headers(options.headers) : new Headers();
+      const localVarQueryParameter = new URLSearchParams(
+        localVarPathQueryStart !== -1 ? localVarPath.substring(localVarPathQueryStart + 1) : ""
+      );
+      if (localVarPathQueryStart !== -1) {
+        localVarPath = localVarPath.substring(0, localVarPathQueryStart);
+      }
+      localVarHeaderParameter.set("Content-Type", "application/json");
+      localVarRequestOptions.headers = localVarHeaderParameter;
+      if (request !== void 0) {
+        localVarRequestOptions.body = JSON.stringify(request || {});
+      }
+      const localVarQueryParameterString = localVarQueryParameter.toString();
+      if (localVarQueryParameterString) {
+        localVarPath += "?" + localVarQueryParameterString;
+      }
+      return {
+        url: localVarPath,
+        options: localVarRequestOptions
+      };
+    },
     /**
      * @summary Retrieves a collection by ID or name.
      * @param {string} tenant <p>Tenant ID</p>
@@ -8605,6 +8666,54 @@ var ApiApiFp = function(configuration) {
         });
       };
     },
+    /**
+     * @summary Forks an existing collection.
+     * @param {string} tenant <p>Tenant ID</p>
+     * @param {string} database <p>Database name</p>
+     * @param {string} collectionId <p>UUID of the collection to update</p>
+     * @param {Api.ForkCollectionPayload} request
+     * @param {RequestInit} [options] Override http request option.
+     * @throws {RequiredError}
+     */
+    forkCollection(tenant, database, collectionId, request, options) {
+      const localVarFetchArgs = ApiApiFetchParamCreator(
+        configuration
+      ).forkCollection(tenant, database, collectionId, request, options);
+      return (fetch2 = defaultFetch, basePath = BASE_PATH) => {
+        return fetch2(
+          basePath + localVarFetchArgs.url,
+          localVarFetchArgs.options
+        ).then((response) => {
+          const contentType = response.headers.get("Content-Type");
+          const mimeType = contentType ? contentType.replace(/;.*/, "") : void 0;
+          if (response.status === 200) {
+            if (mimeType === "application/json") {
+              return response.json();
+            }
+            throw response;
+          }
+          if (response.status === 401) {
+            if (mimeType === "application/json") {
+              throw response;
+            }
+            throw response;
+          }
+          if (response.status === 404) {
+            if (mimeType === "application/json") {
+              throw response;
+            }
+            throw response;
+          }
+          if (response.status === 500) {
+            if (mimeType === "application/json") {
+              throw response;
+            }
+            throw response;
+          }
+          throw response;
+        });
+      };
+    },
     /**
      * @summary Retrieves a collection by ID or name.
      * @param {string} tenant <p>Tenant ID</p>
@@ -9271,6 +9380,24 @@ var ApiApi = class extends BaseAPI {
       options
     )(this.fetch, this.basePath);
   }
+  /**
+   * @summary Forks an existing collection.
+   * @param {string} tenant <p>Tenant ID</p>
+   * @param {string} database <p>Database name</p>
+   * @param {string} collectionId <p>UUID of the collection to update</p>
+   * @param {Api.ForkCollectionPayload} request
+   * @param {RequestInit} [options] Override http request option.
+   * @throws {RequiredError}
+   */
+  forkCollection(tenant, database, collectionId, request, options) {
+    return ApiApiFp(this.configuration).forkCollection(
+      tenant,
+      database,
+      collectionId,
+      request,
+      options
+    )(this.fetch, this.basePath);
+  }
   /**
    * @summary Retrieves a collection by ID or name.
    * @param {string} tenant <p>Tenant ID</p>
@@ -9435,7 +9562,7 @@ var ApiApi = class extends BaseAPI {
 
 // ../chromadb-core/src/generated/models.ts
 var Api;
-((Api2) => {
+((Api3) => {
   let EmbeddingFunctionConfiguration;
   ((EmbeddingFunctionConfiguration2) => {
     let ObjectValue;
@@ -9452,13 +9579,13 @@ var Api;
         TypeEnum2["Known"] = "known";
       })(TypeEnum = AllofValue2.TypeEnum || (AllofValue2.TypeEnum = {}));
     })(AllofValue = EmbeddingFunctionConfiguration2.AllofValue || (EmbeddingFunctionConfiguration2.AllofValue = {}));
-  })(EmbeddingFunctionConfiguration = Api2.EmbeddingFunctionConfiguration || (Api2.EmbeddingFunctionConfiguration = {}));
+  })(EmbeddingFunctionConfiguration = Api3.EmbeddingFunctionConfiguration || (Api3.EmbeddingFunctionConfiguration = {}));
   let HnswSpace;
   ((HnswSpace2) => {
     HnswSpace2["L2"] = "l2";
     HnswSpace2["Cosine"] = "cosine";
     HnswSpace2["Ip"] = "ip";
-  })(HnswSpace = Api2.HnswSpace || (Api2.HnswSpace = {}));
+  })(HnswSpace = Api3.HnswSpace || (Api3.HnswSpace = {}));
   let Include;
   ((Include2) => {
     Include2["Distances"] = "distances";
@@ -9466,7 +9593,7 @@ var Api;
     Include2["Embeddings"] = "embeddings";
     Include2["Metadatas"] = "metadatas";
     Include2["Uris"] = "uris";
-  })(Include = Api2.Include || (Api2.Include = {}));
+  })(Include = Api3.Include || (Api3.Include = {}));
 })(Api || (Api = {}));
 
 // ../chromadb-core/src/generated/configuration.ts
@@ -9569,17 +9696,135 @@ function createErrorByType(type, message) {
   }
 }
 
+// ../chromadb-core/src/CollectionConfiguration.ts
+var InvalidConfigurationError = class extends Error {
+  constructor(message) {
+    super(message);
+    this.name = "InvalidConfigurationError";
+  }
+};
+function serializeEmbeddingFunction(ef) {
+  let efConfig = null;
+  if (ef === null || ef === void 0) {
+    efConfig = { type: "legacy" };
+  } else {
+    try {
+      if (ef.getConfig && ef.name) {
+        if (ef.validateConfig) {
+          ef.validateConfig(ef.getConfig());
+        }
+        efConfig = {
+          name: ef.name,
+          type: "known",
+          config: ef.getConfig()
+        };
+      } else {
+        console.warn(
+          "Could not serialize embedding function - missing getConfig or name method."
+        );
+        efConfig = { type: "legacy" };
+      }
+    } catch (e) {
+      console.warn(
+        "Error processing embedding function for serialization, falling back to legacy:",
+        e,
+        "DeprecationWarning"
+      );
+      efConfig = { type: "legacy" };
+    }
+  }
+  return efConfig;
+}
+function loadApiCollectionConfigurationFromCreateCollectionConfiguration(config) {
+  return createCollectionConfigurationToJson(
+    config
+  );
+}
+function createCollectionConfigurationToJson(config) {
+  if (config.hnsw && config.spann) {
+    throw new InvalidConfigurationError(
+      "Cannot specify both 'hnsw' and 'spann' configurations during creation."
+    );
+  }
+  let hnswConfig = config.hnsw;
+  let spannConfig = config.spann;
+  let ef = config.embedding_function;
+  let efConfig = serializeEmbeddingFunction(ef);
+  if (hnswConfig && typeof hnswConfig !== "object") {
+    throw new Error(
+      "Invalid HNSW config provided in CreateCollectionConfiguration"
+    );
+  }
+  if (spannConfig && typeof spannConfig !== "object") {
+    throw new Error(
+      "Invalid SPANN config provided in CreateCollectionConfiguration"
+    );
+  }
+  return {
+    hnsw: hnswConfig,
+    spann: spannConfig,
+    embedding_function: efConfig
+  };
+}
+function updateCollectionConfigurationToJson(config) {
+  if (config.hnsw && config.spann) {
+    throw new InvalidConfigurationError(
+      "Cannot specify both 'hnsw' and 'spann' configurations during update."
+    );
+  }
+  let hnswConfig = config.hnsw;
+  let spannConfig = config.spann;
+  let ef = config.embedding_function;
+  let efConfig = void 0;
+  if (hnswConfig) {
+    if (typeof hnswConfig !== "object") {
+      throw new Error(
+        "Invalid HNSW config provided in UpdateCollectionConfiguration"
+      );
+    }
+  }
+  if (spannConfig) {
+    if (typeof spannConfig !== "object") {
+      throw new Error(
+        "Invalid SPANN config provided in UpdateCollectionConfiguration"
+      );
+    }
+  }
+  if (ef !== void 0) {
+    efConfig = serializeEmbeddingFunction(ef);
+  }
+  const result = {};
+  if (hnswConfig !== void 0)
+    result.hnsw = hnswConfig;
+  if (spannConfig !== void 0)
+    result.spann = spannConfig;
+  if (efConfig !== void 0)
+    result.embedding_function = efConfig;
+  if (Object.keys(result).length === 0) {
+    throw new InvalidConfigurationError(
+      "No valid configuration fields provided for update."
+    );
+  }
+  return result;
+}
+function loadApiUpdateCollectionConfigurationFromUpdateCollectionConfiguration(config) {
+  return updateCollectionConfigurationToJson(
+    config
+  );
+}
+
 // ../chromadb-core/src/Collection.ts
-var Collection = class {
+var Collection = class _Collection {
   /**
    * @ignore
    */
-  constructor(name, id, client, embeddingFunction, metadata) {
+  constructor(name, id, client, embeddingFunction, metadata, configuration) {
     this.name = name;
     this.id = id;
     this.metadata = metadata;
     this.client = client;
     this.embeddingFunction = embeddingFunction;
+    this.configuration = configuration;
   }
   /**
    * Add items to the collection
@@ -9712,7 +9957,14 @@ var Collection = class {
       },
       this.client.api.options
     );
-    return resp;
+    const finalResp = {
+      ...resp,
+      metadatas: resp.metadatas,
+      documents: resp.documents,
+      embeddings: resp.embeddings,
+      included: resp.include
+    };
+    return finalResp;
   }
   /**
    * Update items in the collection
@@ -9811,7 +10063,15 @@ var Collection = class {
       },
       this.client.api.options
     );
-    return resp;
+    const finalResp = {
+      ...resp,
+      metadatas: resp.metadatas,
+      documents: resp.documents,
+      embeddings: resp.embeddings,
+      distances: resp.distances,
+      included: resp.include
+    };
+    return finalResp;
   }
   /**
    * Modify the collection name or metadata
@@ -9830,16 +10090,24 @@ var Collection = class {
    */
   async modify({
     name,
-    metadata
+    metadata,
+    configuration
   }) {
     await this.client.init();
+    let updateCollectionConfiguration = void 0;
+    if (configuration) {
+      updateCollectionConfiguration = loadApiUpdateCollectionConfigurationFromUpdateCollectionConfiguration(
+        configuration
+      );
+    }
     const resp = await this.client.api.updateCollection(
       this.client.tenant,
       this.client.database,
       this.id,
       {
         new_name: name,
-        new_metadata: metadata
+        new_metadata: metadata,
+        new_configuration: updateCollectionConfiguration
       },
       this.client.api.options
     );
@@ -9914,6 +10182,42 @@ var Collection = class {
       this.client.api.options
     );
   }
+  /**
+   * Forks the collection into a new collection with a new name and configuration.
+   *
+   * @param {Object} params - The parameters for forking the collection.
+   * @param {string} params.newName - The name for the new forked collection.
+   *
+   * @returns {Promise<Collection>} A promise that resolves to the new forked Collection object.
+   * @throws {Error} If there is an issue forking the collection.
+   *
+   * @example
+   * ```typescript
+   * const newCollection = await collection.fork({
+   *   newName: "my_forked_collection",
+   * });
+   * ```
+   */
+  async fork({ newName }) {
+    await this.client.init();
+    const resp = await this.client.api.forkCollection(
+      this.client.tenant,
+      this.client.database,
+      this.id,
+      {
+        new_name: newName
+      },
+      this.client.api.options
+    );
+    const newCollection = new _Collection(
+      resp.name,
+      resp.id,
+      this.client,
+      this.embeddingFunction,
+      resp.metadata
+    );
+    return newCollection;
+  }
 };
 
 // ../chromadb-core/src/utils.ts
@@ -9998,6 +10302,9 @@ async function prepareRecordRequest(reqParams, embeddingFunction, update) {
       `ID's must be unique, found duplicates for: ${duplicateIds}`
     );
   }
+  if (embeddingsArray && embeddingsArray.some((embedding) => embedding.length === 0)) {
+    throw new Error("got empty embedding at pos");
+  }
   return {
     ids,
     metadatas,
@@ -10011,7 +10318,8 @@ function wrapCollection(api, collection) {
     collection.id,
     api,
     collection.embeddingFunction,
-    collection.metadata
+    collection.metadata,
+    collection.configuration
   );
 }
 
@@ -10678,6 +10986,30 @@ var jina_default = {
     api_key_env_var: {
       type: "string",
       description: "Parameter api_key_env_var for the jina embedding function"
+    },
+    task: {
+      type: "string",
+      description: "Parameter task for the jina embedding function"
+    },
+    late_chunking: {
+      type: "boolean",
+      description: "Parameter late_chunking for the jina embedding function"
+    },
+    truncate: {
+      type: "boolean",
+      description: "Parameter truncate for the jina embedding function"
+    },
+    dimensions: {
+      type: "integer",
+      description: "Parameter dimensions for the jina embedding function"
+    },
+    embedding_type: {
+      type: "string",
+      description: "Parameter embedding_type for the jina embedding function"
+    },
+    normalized: {
+      type: "boolean",
+      description: "Parameter normalized for the jina embedding function"
     }
   },
   required: [
@@ -11029,7 +11361,9 @@ function getSchemaVersion(schemaName) {
   return schema.version || "1.0.0";
 }
 function getAvailableSchemas() {
-  return Object.keys(schemaMap).filter((name) => name !== "base_schema");
+  return Object.keys(schemaMap).filter(
+    (name) => name !== "base_schema"
+  );
 }
 function getSchemaInfo() {
   const schemaInfo = {};
@@ -11148,6 +11482,7 @@ var DefaultEmbeddingFunction = class _DefaultEmbeddingFunction {
 };
 
 // ../chromadb-core/src/ChromaClient.ts
+var import_console = require("console");
 var DEFAULT_TENANT2 = "default_tenant";
 var DEFAULT_DATABASE2 = "default_database";
 var ChromaClient = class {
@@ -11300,26 +11635,40 @@ var ChromaClient = class {
   async createCollection({
     name,
     metadata,
-    embeddingFunction = new DefaultEmbeddingFunction()
+    embeddingFunction = new DefaultEmbeddingFunction(),
+    configuration
   }) {
     await this.init();
+    let collectionConfiguration = void 0;
+    if (configuration) {
+      collectionConfiguration = loadApiCollectionConfigurationFromCreateCollectionConfiguration(
+        configuration
+      );
+    }
     const newCollection = await this.api.createCollection(
       this.tenant,
       this.database,
       {
         name,
-        // @ts-ignore: we need to generate the client libraries again
-        configuration: null,
-        //TODO: Configuration type in JavaScript
+        configuration: collectionConfiguration,
         metadata
       },
       this.api.options
     );
+    let config = {};
+    try {
+      config = newCollection.configuration_json;
+    } catch {
+      (0, import_console.warn)(
+        "Server does not respond with configuration_json. Please update server"
+      );
+    }
     return wrapCollection(this, {
       name: newCollection.name,
       id: newCollection.id,
       metadata: newCollection.metadata,
-      embeddingFunction
+      embeddingFunction,
+      configuration: config
     });
   }
   /**
@@ -11346,27 +11695,41 @@ var ChromaClient = class {
   async getOrCreateCollection({
     name,
     metadata,
-    embeddingFunction = new DefaultEmbeddingFunction()
+    embeddingFunction = new DefaultEmbeddingFunction(),
+    configuration
   }) {
     await this.init();
+    let collectionConfiguration = void 0;
+    if (configuration) {
+      collectionConfiguration = loadApiCollectionConfigurationFromCreateCollectionConfiguration(
+        configuration
+      );
+    }
     const newCollection = await this.api.createCollection(
       this.tenant,
       this.database,
       {
         name,
-        // @ts-ignore: we need to generate the client libraries again
-        configuration: null,
-        //TODO: Configuration type in JavaScript
+        configuration: collectionConfiguration,
         metadata,
         get_or_create: true
       },
       this.api.options
     );
+    let config = {};
+    try {
+      config = newCollection.configuration_json;
+    } catch {
+      (0, import_console.warn)(
+        "Server does not respond with configuration_json. Please update server"
+      );
+    }
     return wrapCollection(this, {
       name: newCollection.name,
       id: newCollection.id,
       metadata: newCollection.metadata,
-      embeddingFunction
+      embeddingFunction,
+      configuration: config
     });
   }
   /**
@@ -11471,11 +11834,20 @@ var ChromaClient = class {
       name,
       this.api.options
     );
+    let config = {};
+    try {
+      config = response.configuration_json;
+    } catch {
+      (0, import_console.warn)(
+        "Server does not respond with configuration_json. Please update server"
+      );
+    }
     return wrapCollection(this, {
       id: response.id,
       name: response.name,
       metadata: response.metadata,
-      embeddingFunction: embeddingFunction !== void 0 ? embeddingFunction : new DefaultEmbeddingFunction()
+      embeddingFunction: embeddingFunction !== void 0 ? embeddingFunction : new DefaultEmbeddingFunction(),
+      configuration: config
     });
   }
   /**
@@ -11822,7 +12194,13 @@ var JinaEmbeddingFunction = class _JinaEmbeddingFunction {
   constructor({
     jinaai_api_key,
     model_name = "jina-embeddings-v2-base-en",
-    api_key_env_var = "JINAAI_API_KEY"
+    api_key_env_var = "JINAAI_API_KEY",
+    task,
+    late_chunking,
+    truncate,
+    dimensions,
+    embedding_type,
+    normalized
   }) {
     this.name = "jina";
     const apiKey = jinaai_api_key ?? process.env[api_key_env_var];
@@ -11833,6 +12211,12 @@ var JinaEmbeddingFunction = class _JinaEmbeddingFunction {
     }
     this.model_name = model_name;
     this.api_key_env_var = api_key_env_var;
+    this.task = task;
+    this.late_chunking = late_chunking;
+    this.truncate = truncate;
+    this.dimensions = dimensions;
+    this.embedding_type = embedding_type;
+    this.normalized = normalized;
     this.api_url = "https://api.jina.ai/v1/embeddings";
     this.headers = {
       Authorization: `Bearer ${jinaai_api_key}`,
@@ -11841,14 +12225,33 @@ var JinaEmbeddingFunction = class _JinaEmbeddingFunction {
     };
   }
   async generate(texts) {
+    let json_body = {
+      input: texts,
+      model: this.model_name
+    };
+    if (this.task) {
+      json_body.task = this.task;
+    }
+    if (this.late_chunking) {
+      json_body.late_chunking = this.late_chunking;
+    }
+    if (this.truncate) {
+      json_body.truncate = this.truncate;
+    }
+    if (this.dimensions) {
+      json_body.dimensions = this.dimensions;
+    }
+    if (this.embedding_type) {
+      json_body.embedding_type = this.embedding_type;
+    }
+    if (this.normalized) {
+      json_body.normalized = this.normalized;
+    }
     try {
       const response = await fetch(this.api_url, {
         method: "POST",
         headers: this.headers,
-        body: JSON.stringify({
-          input: texts,
-          model: this.model_name
-        })
+        body: JSON.stringify(json_body)
       });
       const data = await response.json();
       if (!data || !data.data) {
@@ -11868,13 +12271,25 @@ var JinaEmbeddingFunction = class _JinaEmbeddingFunction {
   buildFromConfig(config) {
     return new _JinaEmbeddingFunction({
       model_name: config.model_name,
-      api_key_env_var: config.api_key_env_var
+      api_key_env_var: config.api_key_env_var,
+      task: config.task,
+      late_chunking: config.late_chunking,
+      truncate: config.truncate,
+      dimensions: config.dimensions,
+      embedding_type: config.embedding_type,
+      normalized: config.normalized
     });
   }
   getConfig() {
     return {
       api_key_env_var: this.api_key_env_var,
-      model_name: this.model_name
+      model_name: this.model_name,
+      task: this.task,
+      late_chunking: this.late_chunking,
+      truncate: this.truncate,
+      dimensions: this.dimensions,
+      embedding_type: this.embedding_type,
+      normalized: this.normalized
     };
   }
   validateConfig(config) {