@langchain/google-common 0.2.18 → 1.0.0

package/dist/embeddings.js

@@ -1,184 +1,136 @@
-import { Embeddings } from "@langchain/core/embeddings";
-import { chunkArray } from "@langchain/core/utils/chunk_array";
-import { getEnvironmentVariable } from "@langchain/core/utils/env";
 import { GoogleAIConnection } from "./connection.js";
 import { ApiKeyGoogleAuth } from "./auth.js";
-class EmbeddingsConnection extends GoogleAIConnection {
-    constructor(fields, caller, client, streaming) {
-        super(fields, caller, client, streaming);
-        Object.defineProperty(this, "convertSystemMessageToHumanContent", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-    }
-    buildUrlMethodAiStudio() {
-        return "embedContent";
-    }
-    buildUrlMethodVertex() {
-        return "predict";
-    }
-    async buildUrlMethod() {
-        switch (this.platform) {
-            case "gcp":
-                return this.buildUrlMethodVertex();
-            case "gai":
-                return this.buildUrlMethodAiStudio();
-            default:
-                throw new Error(`Unknown platform when building method: ${this.platform}`);
-        }
-    }
-    get modelPublisher() {
-        // All the embedding models are currently published by "google"
-        return "google";
-    }
-    formatDataAiStudio(input, parameters) {
-        const parts = input.map((instance) => ({
-            text: instance.content,
-        }));
-        const content = {
-            parts,
-        };
-        const outputDimensionality = parameters?.outputDimensionality;
-        const ret = {
-            content,
-            outputDimensionality,
-        };
-        // Remove undefined attributes
-        let key;
-        for (key in ret) {
-            if (ret[key] === undefined) {
-                delete ret[key];
-            }
-        }
-        return ret;
-    }
-    formatDataVertex(input, parameters) {
-        return {
-            instances: input,
-            parameters,
-        };
-    }
-    async formatData(input, parameters) {
-        switch (this.platform) {
-            case "gcp":
-                return this.formatDataVertex(input, parameters);
-            case "gai":
-                return this.formatDataAiStudio(input, parameters);
-            default:
-                throw new Error(`Unknown platform to format embeddings ${this.platform}`);
-        }
-    }
-}
+import { getEnvironmentVariable } from "@langchain/core/utils/env";
+import { Embeddings } from "@langchain/core/embeddings";
+import { chunkArray } from "@langchain/core/utils/chunk_array";
+
+//#region src/embeddings.ts
+var EmbeddingsConnection = class extends GoogleAIConnection {
+	convertSystemMessageToHumanContent;
+	constructor(fields, caller, client, streaming) {
+		super(fields, caller, client, streaming);
+	}
+	buildUrlMethodAiStudio() {
+		return "embedContent";
+	}
+	buildUrlMethodVertex() {
+		return "predict";
+	}
+	async buildUrlMethod() {
+		switch (this.platform) {
+			case "gcp": return this.buildUrlMethodVertex();
+			case "gai": return this.buildUrlMethodAiStudio();
+			default: throw new Error(`Unknown platform when building method: ${this.platform}`);
+		}
+	}
+	get modelPublisher() {
+		return "google";
+	}
+	formatDataAiStudio(input, parameters) {
+		const parts = input.map((instance) => ({ text: instance.content }));
+		const content = { parts };
+		const outputDimensionality = parameters?.outputDimensionality;
+		const ret = {
+			content,
+			outputDimensionality
+		};
+		let key;
+		for (key in ret) if (ret[key] === void 0) delete ret[key];
+		return ret;
+	}
+	formatDataVertex(input, parameters) {
+		return {
+			instances: input,
+			parameters
+		};
+	}
+	async formatData(input, parameters) {
+		switch (this.platform) {
+			case "gcp": return this.formatDataVertex(input, parameters);
+			case "gai": return this.formatDataAiStudio(input, parameters);
+			default: throw new Error(`Unknown platform to format embeddings ${this.platform}`);
+		}
+	}
+};
 /**
- * Enables calls to Google APIs for generating
- * text embeddings.
- */
-export class BaseGoogleEmbeddings extends Embeddings {
-    constructor(fields) {
-        super(fields);
-        Object.defineProperty(this, "model", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "dimensions", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "connection", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        this.model = fields.model;
-        this.dimensions = fields.dimensions ?? fields.outputDimensionality;
-        this.connection = new EmbeddingsConnection({ ...fields, ...this }, this.caller, this.buildClient(fields), false);
-    }
-    buildApiKeyClient(apiKey) {
-        return new ApiKeyGoogleAuth(apiKey);
-    }
-    buildApiKey(fields) {
-        return fields?.apiKey ?? getEnvironmentVariable("GOOGLE_API_KEY");
-    }
-    buildClient(fields) {
-        const apiKey = this.buildApiKey(fields);
-        if (apiKey) {
-            return this.buildApiKeyClient(apiKey);
-        }
-        else {
-            return this.buildAbstractedClient(fields);
-        }
-    }
-    buildParameters() {
-        const ret = {
-            outputDimensionality: this.dimensions,
-        };
-        // Remove undefined attributes
-        let key;
-        for (key in ret) {
-            if (ret[key] === undefined) {
-                delete ret[key];
-            }
-        }
-        return ret;
-    }
-    vertexResponseToValues(response) {
-        const predictions = response?.data?.predictions ?? [];
-        return predictions.map((prediction) => prediction.embeddings.values);
-    }
-    aiStudioResponseToValues(response) {
-        const value = response?.data?.embedding?.values ?? [];
-        return [value];
-    }
-    responseToValues(response) {
-        switch (this.connection.platform) {
-            case "gcp":
-                return this.vertexResponseToValues(response);
-            case "gai":
-                return this.aiStudioResponseToValues(response);
-            default:
-                throw new Error(`Unknown response platform: ${this.connection.platform}`);
-        }
-    }
-    /**
-     * Takes an array of documents as input and returns a promise that
-     * resolves to a 2D array of embeddings for each document. It splits the
-     * documents into chunks and makes requests to the Google Vertex AI API to
-     * generate embeddings.
-     * @param documents An array of documents to be embedded.
-     * @returns A promise that resolves to a 2D array of embeddings for each document.
-     */
-    async embedDocuments(documents) {
-        // Vertex "text-" models could do up 5 documents at once,
-        // but the "gemini-embedding-001" can only do 1.
-        // AI Studio can only do a chunk size of 1.
-        // TODO: Make this configurable
-        const chunkSize = 1;
-        const instanceChunks = chunkArray(documents.map((document) => ({
-            content: document,
-        })), chunkSize);
-        const parameters = this.buildParameters();
-        const options = {};
-        const responses = await Promise.all(instanceChunks.map((instances) => this.connection.request(instances, parameters, options)));
-        const result = responses?.map((response) => this.responseToValues(response)).flat() ??
-            [];
-        return result;
-    }
-    /**
-     * Takes a document as input and returns a promise that resolves to an
-     * embedding for the document. It calls the embedDocuments method with the
-     * document as the input.
-     * @param document A document to be embedded.
-     * @returns A promise that resolves to an embedding for the document.
-     */
-    async embedQuery(document) {
-        const data = await this.embedDocuments([document]);
-        return data[0];
-    }
-}
+* Enables calls to Google APIs for generating
+* text embeddings.
+*/
+var BaseGoogleEmbeddings = class extends Embeddings {
+	model;
+	dimensions;
+	connection;
+	constructor(fields) {
+		super(fields);
+		this.model = fields.model;
+		this.dimensions = fields.dimensions ?? fields.outputDimensionality;
+		this.connection = new EmbeddingsConnection({
+			...fields,
+			...this
+		}, this.caller, this.buildClient(fields), false);
+	}
+	buildApiKeyClient(apiKey) {
+		return new ApiKeyGoogleAuth(apiKey);
+	}
+	buildApiKey(fields) {
+		return fields?.apiKey ?? getEnvironmentVariable("GOOGLE_API_KEY");
+	}
+	buildClient(fields) {
+		const apiKey = this.buildApiKey(fields);
+		if (apiKey) return this.buildApiKeyClient(apiKey);
+		else return this.buildAbstractedClient(fields);
+	}
+	buildParameters() {
+		const ret = { outputDimensionality: this.dimensions };
+		let key;
+		for (key in ret) if (ret[key] === void 0) delete ret[key];
+		return ret;
+	}
+	vertexResponseToValues(response) {
+		const predictions = response?.data?.predictions ?? [];
+		return predictions.map((prediction) => prediction.embeddings.values);
+	}
+	aiStudioResponseToValues(response) {
+		const value = response?.data?.embedding?.values ?? [];
+		return [value];
+	}
+	responseToValues(response) {
+		switch (this.connection.platform) {
+			case "gcp": return this.vertexResponseToValues(response);
+			case "gai": return this.aiStudioResponseToValues(response);
+			default: throw new Error(`Unknown response platform: ${this.connection.platform}`);
+		}
+	}
+	/**
+	* Takes an array of documents as input and returns a promise that
+	* resolves to a 2D array of embeddings for each document. It splits the
+	* documents into chunks and makes requests to the Google Vertex AI API to
+	* generate embeddings.
+	* @param documents An array of documents to be embedded.
+	* @returns A promise that resolves to a 2D array of embeddings for each document.
+	*/
+	async embedDocuments(documents) {
+		const chunkSize = 1;
+		const instanceChunks = chunkArray(documents.map((document) => ({ content: document })), chunkSize);
+		const parameters = this.buildParameters();
+		const options = {};
+		const responses = await Promise.all(instanceChunks.map((instances) => this.connection.request(instances, parameters, options)));
+		const result = responses?.map((response) => this.responseToValues(response)).flat() ?? [];
+		return result;
+	}
+	/**
+	* Takes a document as input and returns a promise that resolves to an
+	* embedding for the document. It calls the embedDocuments method with the
+	* document as the input.
+	* @param document A document to be embedded.
+	* @returns A promise that resolves to an embedding for the document.
+	*/
+	async embedQuery(document) {
+		const data = await this.embedDocuments([document]);
+		return data[0];
+	}
+};
+
+//#endregion
+export { BaseGoogleEmbeddings };
+//# sourceMappingURL=embeddings.js.map

package/dist/embeddings.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"embeddings.js","names":["fields: BaseGoogleEmbeddingsParams<AuthOptions> | undefined","caller: AsyncCaller","client: GoogleAbstractedClient","streaming: boolean","input: VertexEmbeddingsInstance[]","parameters: VertexEmbeddingsParameters","parts: GeminiPartText[]","instance: VertexEmbeddingsInstance","ret: AIStudioEmbeddingsRequest","key: keyof AIStudioEmbeddingsRequest","fields: BaseGoogleEmbeddingsParams<AuthOptions>","apiKey: string","fields?: GoogleConnectionParams<AuthOptions>","ret: VertexEmbeddingsParameters","key: keyof VertexEmbeddingsParameters","response: VertexEmbeddingsResponse","predictions: VertexEmbeddingsResponsePrediction[]","prediction: VertexEmbeddingsResponsePrediction","response: AIStudioEmbeddingsResponse","value: number[]","response: GoogleEmbeddingsResponse","documents: string[]","instanceChunks: VertexEmbeddingsInstance[][]","result: number[][]","document: string"],"sources":["../src/embeddings.ts"],"sourcesContent":["import { Embeddings } from \"@langchain/core/embeddings\";\nimport {\n  AsyncCaller,\n  AsyncCallerCallOptions,\n} from \"@langchain/core/utils/async_caller\";\nimport { chunkArray } from \"@langchain/core/utils/chunk_array\";\nimport { getEnvironmentVariable } from \"@langchain/core/utils/env\";\n\nimport { GoogleAIConnection } from \"./connection.js\";\nimport { ApiKeyGoogleAuth, GoogleAbstractedClient } from \"./auth.js\";\nimport {\n  BaseGoogleEmbeddingsOptions,\n  BaseGoogleEmbeddingsParams,\n  GoogleConnectionParams,\n  VertexEmbeddingsInstance,\n  GoogleEmbeddingsResponse,\n  VertexEmbeddingsParameters,\n  GoogleEmbeddingsRequest,\n  VertexEmbeddingsResponse,\n  AIStudioEmbeddingsResponse,\n  VertexEmbeddingsResponsePrediction,\n  AIStudioEmbeddingsRequest,\n  GeminiPartText,\n  VertexEmbeddingsRequest,\n} from \"./types.js\";\n\nclass EmbeddingsConnection<\n  CallOptions extends AsyncCallerCallOptions,\n  AuthOptions\n> extends GoogleAIConnection<\n  CallOptions,\n  VertexEmbeddingsInstance[],\n  AuthOptions,\n  GoogleEmbeddingsResponse\n> {\n  convertSystemMessageToHumanContent: boolean | undefined;\n\n  constructor(\n    fields: BaseGoogleEmbeddingsParams<AuthOptions> | undefined,\n    caller: AsyncCaller,\n    client: GoogleAbstractedClient,\n    streaming: boolean\n  ) {\n    super(fields, caller, client, streaming);\n  }\n\n  buildUrlMethodAiStudio(): string {\n    return \"embedContent\";\n  }\n\n  buildUrlMethodVertex(): string {\n    return \"predict\";\n  }\n\n  async buildUrlMethod(): Promise<string> {\n    switch (this.platform) {\n      case \"gcp\":\n        return this.buildUrlMethodVertex();\n      case \"gai\":\n        return this.buildUrlMethodAiStudio();\n      default:\n        throw new Error(\n          `Unknown platform when building method: ${this.platform}`\n        );\n    }\n  }\n\n  get modelPublisher(): string {\n    // All the embedding models are currently published by \"google\"\n    return \"google\";\n  }\n\n  formatDataAiStudio(\n    input: VertexEmbeddingsInstance[],\n    parameters: VertexEmbeddingsParameters\n  ): AIStudioEmbeddingsRequest {\n    const parts: GeminiPartText[] = input.map(\n      (instance: VertexEmbeddingsInstance) => ({\n        text: instance.content,\n      })\n    );\n    const content = {\n      parts,\n    };\n    const outputDimensionality = parameters?.outputDimensionality;\n\n    const ret: AIStudioEmbeddingsRequest = {\n      content,\n      outputDimensionality,\n    };\n\n    // Remove undefined attributes\n    let key: keyof AIStudioEmbeddingsRequest;\n    for (key in ret) {\n      if (ret[key] === undefined) {\n        delete ret[key];\n      }\n    }\n\n    return ret;\n  }\n\n  formatDataVertex(\n    input: VertexEmbeddingsInstance[],\n    parameters: VertexEmbeddingsParameters\n  ): VertexEmbeddingsRequest {\n    return {\n      instances: input,\n      parameters,\n    };\n  }\n\n  async formatData(\n    input: VertexEmbeddingsInstance[],\n    parameters: VertexEmbeddingsParameters\n  ): Promise<GoogleEmbeddingsRequest> {\n    switch (this.platform) {\n      case \"gcp\":\n        return this.formatDataVertex(input, parameters);\n      case \"gai\":\n        return this.formatDataAiStudio(input, parameters);\n      default:\n        throw new Error(\n          `Unknown platform to format embeddings ${this.platform}`\n        );\n    }\n  }\n}\n\n/**\n * Enables calls to Google APIs for generating\n * text embeddings.\n */\nexport abstract class BaseGoogleEmbeddings<AuthOptions>\n  extends Embeddings\n  implements BaseGoogleEmbeddingsParams<AuthOptions>\n{\n  model: string;\n\n  dimensions?: number;\n\n  private connection: EmbeddingsConnection<\n    BaseGoogleEmbeddingsOptions,\n    AuthOptions\n  >;\n\n  constructor(fields: BaseGoogleEmbeddingsParams<AuthOptions>) {\n    super(fields);\n\n    this.model = fields.model;\n    this.dimensions = fields.dimensions ?? fields.outputDimensionality;\n\n    this.connection = new EmbeddingsConnection(\n      { ...fields, ...this },\n      this.caller,\n      this.buildClient(fields),\n      false\n    );\n  }\n\n  abstract buildAbstractedClient(\n    fields?: GoogleConnectionParams<AuthOptions>\n  ): GoogleAbstractedClient;\n\n  buildApiKeyClient(apiKey: string): GoogleAbstractedClient {\n    return new ApiKeyGoogleAuth(apiKey);\n  }\n\n  buildApiKey(\n    fields?: GoogleConnectionParams<AuthOptions>\n  ): string | undefined {\n    return fields?.apiKey ?? getEnvironmentVariable(\"GOOGLE_API_KEY\");\n  }\n\n  buildClient(\n    fields?: GoogleConnectionParams<AuthOptions>\n  ): GoogleAbstractedClient {\n    const apiKey = this.buildApiKey(fields);\n    if (apiKey) {\n      return this.buildApiKeyClient(apiKey);\n    } else {\n      return this.buildAbstractedClient(fields);\n    }\n  }\n\n  buildParameters(): VertexEmbeddingsParameters {\n    const ret: VertexEmbeddingsParameters = {\n      outputDimensionality: this.dimensions,\n    };\n\n    // Remove undefined attributes\n    let key: keyof VertexEmbeddingsParameters;\n    for (key in ret) {\n      if (ret[key] === undefined) {\n        delete ret[key];\n      }\n    }\n\n    return ret;\n  }\n\n  vertexResponseToValues(response: VertexEmbeddingsResponse): number[][] {\n    const predictions: VertexEmbeddingsResponsePrediction[] =\n      response?.data?.predictions ?? [];\n    return predictions.map(\n      (prediction: VertexEmbeddingsResponsePrediction): number[] =>\n        prediction.embeddings.values\n    );\n  }\n\n  aiStudioResponseToValues(response: AIStudioEmbeddingsResponse): number[][] {\n    const value: number[] = response?.data?.embedding?.values ?? [];\n    return [value];\n  }\n\n  responseToValues(response: GoogleEmbeddingsResponse): number[][] {\n    switch (this.connection.platform) {\n      case \"gcp\":\n        return this.vertexResponseToValues(\n          response as VertexEmbeddingsResponse\n        );\n      case \"gai\":\n        return this.aiStudioResponseToValues(\n          response as AIStudioEmbeddingsResponse\n        );\n      default:\n        throw new Error(\n          `Unknown response platform: ${this.connection.platform}`\n        );\n    }\n  }\n\n  /**\n   * Takes an array of documents as input and returns a promise that\n   * resolves to a 2D array of embeddings for each document. It splits the\n   * documents into chunks and makes requests to the Google Vertex AI API to\n   * generate embeddings.\n   * @param documents An array of documents to be embedded.\n   * @returns A promise that resolves to a 2D array of embeddings for each document.\n   */\n  async embedDocuments(documents: string[]): Promise<number[][]> {\n    // Vertex \"text-\" models could do up 5 documents at once,\n    // but the \"gemini-embedding-001\" can only do 1.\n    // AI Studio can only do a chunk size of 1.\n    // TODO: Make this configurable\n    const chunkSize = 1;\n    const instanceChunks: VertexEmbeddingsInstance[][] = chunkArray(\n      documents.map((document) => ({\n        content: document,\n      })),\n      chunkSize\n    );\n    const parameters: VertexEmbeddingsParameters = this.buildParameters();\n    const options = {};\n    const responses = await Promise.all(\n      instanceChunks.map((instances) =>\n        this.connection.request(instances, parameters, options)\n      )\n    );\n    const result: number[][] =\n      responses?.map((response) => this.responseToValues(response)).flat() ??\n      [];\n    return result;\n  }\n\n  /**\n   * Takes a document as input and returns a promise that resolves to an\n   * embedding for the document. It calls the embedDocuments method with the\n   * document as the input.\n   * @param document A document to be embedded.\n   * @returns A promise that resolves to an embedding for the document.\n   */\n  async embedQuery(document: string): Promise<number[]> {\n    const data = await this.embedDocuments([document]);\n    return data[0];\n  }\n}\n"],"mappings":";;;;;;;AA0BA,IAAM,uBAAN,cAGU,mBAKR;CACA;CAEA,YACEA,QACAC,QACAC,QACAC,WACA;EACA,MAAM,QAAQ,QAAQ,QAAQ,UAAU;CACzC;CAED,yBAAiC;AAC/B,SAAO;CACR;CAED,uBAA+B;AAC7B,SAAO;CACR;CAED,MAAM,iBAAkC;AACtC,UAAQ,KAAK,UAAb;GACE,KAAK,MACH,QAAO,KAAK,sBAAsB;GACpC,KAAK,MACH,QAAO,KAAK,wBAAwB;GACtC,QACE,OAAM,IAAI,MACR,CAAC,uCAAuC,EAAE,KAAK,UAAU;EAE9D;CACF;CAED,IAAI,iBAAyB;AAE3B,SAAO;CACR;CAED,mBACEC,OACAC,YAC2B;EAC3B,MAAMC,QAA0B,MAAM,IACpC,CAACC,cAAwC,EACvC,MAAM,SAAS,QAChB,GACF;EACD,MAAM,UAAU,EACd,MACD;EACD,MAAM,uBAAuB,YAAY;EAEzC,MAAMC,MAAiC;GACrC;GACA;EACD;EAGD,IAAIC;AACJ,OAAK,OAAO,IACV,KAAI,IAAI,SAAS,QACf,OAAO,IAAI;AAIf,SAAO;CACR;CAED,iBACEL,OACAC,YACyB;AACzB,SAAO;GACL,WAAW;GACX;EACD;CACF;CAED,MAAM,WACJD,OACAC,YACkC;AAClC,UAAQ,KAAK,UAAb;GACE,KAAK,MACH,QAAO,KAAK,iBAAiB,OAAO,WAAW;GACjD,KAAK,MACH,QAAO,KAAK,mBAAmB,OAAO,WAAW;GACnD,QACE,OAAM,IAAI,MACR,CAAC,sCAAsC,EAAE,KAAK,UAAU;EAE7D;CACF;AACF;;;;;AAMD,IAAsB,uBAAtB,cACU,WAEV;CACE;CAEA;CAEA,AAAQ;CAKR,YAAYK,QAAiD;EAC3D,MAAM,OAAO;EAEb,KAAK,QAAQ,OAAO;EACpB,KAAK,aAAa,OAAO,cAAc,OAAO;EAE9C,KAAK,aAAa,IAAI,qBACpB;GAAE,GAAG;GAAQ,GAAG;EAAM,GACtB,KAAK,QACL,KAAK,YAAY,OAAO,EACxB;CAEH;CAMD,kBAAkBC,QAAwC;AACxD,SAAO,IAAI,iBAAiB;CAC7B;CAED,YACEC,QACoB;AACpB,SAAO,QAAQ,UAAU,uBAAuB,iBAAiB;CAClE;CAED,YACEA,QACwB;EACxB,MAAM,SAAS,KAAK,YAAY,OAAO;AACvC,MAAI,OACF,QAAO,KAAK,kBAAkB,OAAO;MAErC,QAAO,KAAK,sBAAsB,OAAO;CAE5C;CAED,kBAA8C;EAC5C,MAAMC,MAAkC,EACtC,sBAAsB,KAAK,WAC5B;EAGD,IAAIC;AACJ,OAAK,OAAO,IACV,KAAI,IAAI,SAAS,QACf,OAAO,IAAI;AAIf,SAAO;CACR;CAED,uBAAuBC,UAAgD;EACrE,MAAMC,cACJ,UAAU,MAAM,eAAe,CAAE;AACnC,SAAO,YAAY,IACjB,CAACC,eACC,WAAW,WAAW,OACzB;CACF;CAED,yBAAyBC,UAAkD;EACzE,MAAMC,QAAkB,UAAU,MAAM,WAAW,UAAU,CAAE;AAC/D,SAAO,CAAC,KAAM;CACf;CAED,iBAAiBC,UAAgD;AAC/D,UAAQ,KAAK,WAAW,UAAxB;GACE,KAAK,MACH,QAAO,KAAK,uBACV,SACD;GACH,KAAK,MACH,QAAO,KAAK,yBACV,SACD;GACH,QACE,OAAM,IAAI,MACR,CAAC,2BAA2B,EAAE,KAAK,WAAW,UAAU;EAE7D;CACF;;;;;;;;;CAUD,MAAM,eAAeC,WAA0C;EAK7D,MAAM,YAAY;EAClB,MAAMC,iBAA+C,WACnD,UAAU,IAAI,CAAC,cAAc,EAC3B,SAAS,SACV,GAAE,EACH,UACD;EACD,MAAMjB,aAAyC,KAAK,iBAAiB;EACrE,MAAM,UAAU,CAAE;EAClB,MAAM,YAAY,MAAM,QAAQ,IAC9B,eAAe,IAAI,CAAC,cAClB,KAAK,WAAW,QAAQ,WAAW,YAAY,QAAQ,CACxD,CACF;EACD,MAAMkB,SACJ,WAAW,IAAI,CAAC,aAAa,KAAK,iBAAiB,SAAS,CAAC,CAAC,MAAM,IACpE,CAAE;AACJ,SAAO;CACR;;;;;;;;CASD,MAAM,WAAWC,UAAqC;EACpD,MAAM,OAAO,MAAM,KAAK,eAAe,CAAC,QAAS,EAAC;AAClD,SAAO,KAAK;CACb;AACF"}