@5ss/ai-tools 2.11.1

package/dist/modules/document-extract/index.d.ts

@@ -0,0 +1,438 @@
+import { a as ModuleDefinition, o as ToolContext, s as ToolDefinition } from "../../types-Ccb0DXY9.js";
+import "../../index-CZX5OeOF.js";
+import { C as TextractTextChunk, S as TextractStatusInput, _ as TextractExtractResult, b as TextractExtractTextInput, d as MAX_BATCH_EXTRACT, p as MAX_INLINE_EXTRACT_CHARS, v as TextractExtractTextBatchInput, x as TextractOutputMode, y as TextractExtractTextBatchOutput } from "../../index-CdFmEfZO.js";
+import { z } from "zod";
+//#region src/modules/document-extract/contracts.d.ts
+declare const documentExtractAuthSchema: z.ZodDiscriminatedUnion<[z.ZodObject<{
+  access_key_id: z.ZodString;
+  secret_access_key: z.ZodString;
+  region: z.ZodString;
+  bucket: z.ZodString;
+  session_token: z.ZodOptional<z.ZodString>;
+  key_prefix: z.ZodOptional<z.ZodString>;
+  poll_timeout_ms: z.ZodOptional<z.ZodInt>;
+  poll_interval_ms: z.ZodOptional<z.ZodInt>;
+  provider: z.ZodLiteral<"textract">;
+}, z.core.$strip>], "provider">;
+type DocumentExtractAuth = z.infer<typeof documentExtractAuthSchema>;
+/** Capability I/O — re-export Textract shapes under seam names (vendor owns presentation). */
+declare const extractOutputModeSchema: z.ZodEnum<{
+  artifact: "artifact";
+  chunks: "chunks";
+  inline: "inline";
+}>;
+type ExtractOutputMode = TextractOutputMode;
+type ExtractTextChunk = TextractTextChunk;
+declare const extractResultSchema: z.ZodObject<{
+  status: z.ZodEnum<{
+    failed: "failed";
+    pending: "pending";
+    succeeded: "succeeded";
+  }>;
+  job_id: z.ZodOptional<z.ZodString>;
+  text: z.ZodOptional<z.ZodString>;
+  page_count: z.ZodOptional<z.ZodInt>;
+  error: z.ZodOptional<z.ZodString>;
+  source: z.ZodOptional<z.ZodObject<{
+    store: z.ZodEnum<{
+      host: "host";
+      object: "object";
+    }>;
+    key: z.ZodString;
+    media_type: z.ZodOptional<z.ZodString>;
+    filename: z.ZodOptional<z.ZodString>;
+    byte_length: z.ZodOptional<z.ZodInt>;
+  }, z.core.$strip>>;
+  output: z.ZodOptional<z.ZodEnum<{
+    artifact: "artifact";
+    chunks: "chunks";
+    inline: "inline";
+  }>>;
+  artifact: z.ZodOptional<z.ZodObject<{
+    store: z.ZodEnum<{
+      host: "host";
+      object: "object";
+    }>;
+    key: z.ZodString;
+    media_type: z.ZodOptional<z.ZodString>;
+    filename: z.ZodOptional<z.ZodString>;
+    byte_length: z.ZodOptional<z.ZodInt>;
+  }, z.core.$strip>>;
+  chunks: z.ZodOptional<z.ZodArray<z.ZodObject<{
+    id: z.ZodString;
+    index: z.ZodNumber;
+    text: z.ZodString;
+  }, z.core.$strip>>>;
+}, z.core.$strip>;
+declare const extractTextInputSchema: z.ZodObject<{
+  output: z.ZodOptional<z.ZodEnum<{
+    artifact: "artifact";
+    chunks: "chunks";
+    inline: "inline";
+  }>>;
+  destination_key: z.ZodOptional<z.ZodString>;
+  chunk: z.ZodOptional<z.ZodObject<{
+    max_chars: z.ZodOptional<z.ZodInt>;
+    overlap: z.ZodOptional<z.ZodInt>;
+  }, z.core.$strip>>;
+  source: z.ZodObject<{
+    store: z.ZodEnum<{
+      host: "host";
+      object: "object";
+    }>;
+    key: z.ZodString;
+    media_type: z.ZodOptional<z.ZodString>;
+    filename: z.ZodOptional<z.ZodString>;
+    byte_length: z.ZodOptional<z.ZodInt>;
+  }, z.core.$strip>;
+}, z.core.$strip>;
+declare const extractTextBatchInputSchema: z.ZodObject<{
+  output: z.ZodOptional<z.ZodEnum<{
+    artifact: "artifact";
+    chunks: "chunks";
+    inline: "inline";
+  }>>;
+  destination_key: z.ZodOptional<z.ZodString>;
+  chunk: z.ZodOptional<z.ZodObject<{
+    max_chars: z.ZodOptional<z.ZodInt>;
+    overlap: z.ZodOptional<z.ZodInt>;
+  }, z.core.$strip>>;
+  sources: z.ZodArray<z.ZodObject<{
+    store: z.ZodEnum<{
+      host: "host";
+      object: "object";
+    }>;
+    key: z.ZodString;
+    media_type: z.ZodOptional<z.ZodString>;
+    filename: z.ZodOptional<z.ZodString>;
+    byte_length: z.ZodOptional<z.ZodInt>;
+  }, z.core.$strip>>;
+  destination_key_prefix: z.ZodOptional<z.ZodString>;
+}, z.core.$strip>;
+declare const statusInputSchema: z.ZodObject<{
+  output: z.ZodOptional<z.ZodEnum<{
+    artifact: "artifact";
+    chunks: "chunks";
+    inline: "inline";
+  }>>;
+  destination_key: z.ZodOptional<z.ZodString>;
+  chunk: z.ZodOptional<z.ZodObject<{
+    max_chars: z.ZodOptional<z.ZodInt>;
+    overlap: z.ZodOptional<z.ZodInt>;
+  }, z.core.$strip>>;
+  job_id: z.ZodString;
+}, z.core.$strip>;
+declare const extractTextBatchOutputSchema: z.ZodObject<{
+  results: z.ZodArray<z.ZodObject<{
+    index: z.ZodInt;
+    ok: z.ZodBoolean;
+    value: z.ZodOptional<z.ZodObject<{
+      status: z.ZodEnum<{
+        failed: "failed";
+        pending: "pending";
+        succeeded: "succeeded";
+      }>;
+      job_id: z.ZodOptional<z.ZodString>;
+      text: z.ZodOptional<z.ZodString>;
+      page_count: z.ZodOptional<z.ZodInt>;
+      error: z.ZodOptional<z.ZodString>;
+      source: z.ZodOptional<z.ZodObject<{
+        store: z.ZodEnum<{
+          host: "host";
+          object: "object";
+        }>;
+        key: z.ZodString;
+        media_type: z.ZodOptional<z.ZodString>;
+        filename: z.ZodOptional<z.ZodString>;
+        byte_length: z.ZodOptional<z.ZodInt>;
+      }, z.core.$strip>>;
+      output: z.ZodOptional<z.ZodEnum<{
+        artifact: "artifact";
+        chunks: "chunks";
+        inline: "inline";
+      }>>;
+      artifact: z.ZodOptional<z.ZodObject<{
+        store: z.ZodEnum<{
+          host: "host";
+          object: "object";
+        }>;
+        key: z.ZodString;
+        media_type: z.ZodOptional<z.ZodString>;
+        filename: z.ZodOptional<z.ZodString>;
+        byte_length: z.ZodOptional<z.ZodInt>;
+      }, z.core.$strip>>;
+      chunks: z.ZodOptional<z.ZodArray<z.ZodObject<{
+        id: z.ZodString;
+        index: z.ZodNumber;
+        text: z.ZodString;
+      }, z.core.$strip>>>;
+    }, z.core.$strip>>;
+    error: z.ZodOptional<z.ZodObject<{
+      code: z.ZodString;
+      message: z.ZodString;
+      retryable: z.ZodOptional<z.ZodBoolean>;
+    }, z.core.$strip>>;
+  }, z.core.$strip>>;
+  succeeded: z.ZodInt;
+  failed: z.ZodInt;
+}, z.core.$strip>;
+type ExtractResult = TextractExtractResult;
+type ExtractTextInput = TextractExtractTextInput;
+type ExtractTextBatchInput = TextractExtractTextBatchInput;
+type StatusInput = TextractStatusInput;
+type ExtractTextBatchOutput = TextractExtractTextBatchOutput;
+/** Provider + public client surface (same ops; vendor does the work). */
+type DocumentExtractOps = {
+  extractText: (input: ExtractTextInput) => Promise<ExtractResult>;
+  getStatus: (input: StatusInput) => Promise<ExtractResult>;
+  extractTextBatch: (input: ExtractTextBatchInput) => Promise<ExtractTextBatchOutput>;
+};
+//#endregion
+//#region src/modules/document-extract/client.d.ts
+declare class DocumentExtractClient implements DocumentExtractOps {
+  #private;
+  constructor(ops: DocumentExtractOps);
+  static fromContext(ctx: ToolContext): DocumentExtractClient;
+  static fromAuth(auth: DocumentExtractAuth, ctx?: ToolContext): DocumentExtractClient;
+  extractText(input: ExtractTextInput): Promise<{
+    status: "failed" | "pending" | "succeeded";
+    job_id?: string | undefined;
+    text?: string | undefined;
+    page_count?: number | undefined;
+    error?: string | undefined;
+    source?: {
+      store: "host" | "object";
+      key: string;
+      media_type?: string | undefined;
+      filename?: string | undefined;
+      byte_length?: number | undefined;
+    } | undefined;
+    output?: "artifact" | "chunks" | "inline" | undefined;
+    artifact?: {
+      store: "host" | "object";
+      key: string;
+      media_type?: string | undefined;
+      filename?: string | undefined;
+      byte_length?: number | undefined;
+    } | undefined;
+    chunks?: {
+      id: string;
+      index: number;
+      text: string;
+    }[] | undefined;
+  }>;
+  getStatus(input: StatusInput): Promise<{
+    status: "failed" | "pending" | "succeeded";
+    job_id?: string | undefined;
+    text?: string | undefined;
+    page_count?: number | undefined;
+    error?: string | undefined;
+    source?: {
+      store: "host" | "object";
+      key: string;
+      media_type?: string | undefined;
+      filename?: string | undefined;
+      byte_length?: number | undefined;
+    } | undefined;
+    output?: "artifact" | "chunks" | "inline" | undefined;
+    artifact?: {
+      store: "host" | "object";
+      key: string;
+      media_type?: string | undefined;
+      filename?: string | undefined;
+      byte_length?: number | undefined;
+    } | undefined;
+    chunks?: {
+      id: string;
+      index: number;
+      text: string;
+    }[] | undefined;
+  }>;
+  extractTextBatch(input: ExtractTextBatchInput): Promise<{
+    results: {
+      index: number;
+      ok: boolean;
+      value?: {
+        status: "failed" | "pending" | "succeeded";
+        job_id?: string | undefined;
+        text?: string | undefined;
+        page_count?: number | undefined;
+        error?: string | undefined;
+        source?: {
+          store: "host" | "object";
+          key: string;
+          media_type?: string | undefined;
+          filename?: string | undefined;
+          byte_length?: number | undefined;
+        } | undefined;
+        output?: "artifact" | "chunks" | "inline" | undefined;
+        artifact?: {
+          store: "host" | "object";
+          key: string;
+          media_type?: string | undefined;
+          filename?: string | undefined;
+          byte_length?: number | undefined;
+        } | undefined;
+        chunks?: {
+          id: string;
+          index: number;
+          text: string;
+        }[] | undefined;
+      } | undefined;
+      error?: {
+        code: string;
+        message: string;
+        retryable?: boolean | undefined;
+      } | undefined;
+    }[];
+    succeeded: number;
+    failed: number;
+  }>;
+}
+//#endregion
+//#region src/modules/document-extract/module.d.ts
+declare const documentExtractTextTool: ToolDefinition<{
+  output?: "artifact" | "chunks" | "inline" | undefined;
+  destination_key?: string | undefined;
+  chunk?: {
+    max_chars?: number | undefined;
+    overlap?: number | undefined;
+  } | undefined;
+  source: {
+    store: "host" | "object";
+    key: string;
+    media_type?: string | undefined;
+    filename?: string | undefined;
+    byte_length?: number | undefined;
+  };
+}, {
+  status: "failed" | "pending" | "succeeded";
+  job_id?: string | undefined;
+  text?: string | undefined;
+  page_count?: number | undefined;
+  error?: string | undefined;
+  source?: {
+    store: "host" | "object";
+    key: string;
+    media_type?: string | undefined;
+    filename?: string | undefined;
+    byte_length?: number | undefined;
+  } | undefined;
+  output?: "artifact" | "chunks" | "inline" | undefined;
+  artifact?: {
+    store: "host" | "object";
+    key: string;
+    media_type?: string | undefined;
+    filename?: string | undefined;
+    byte_length?: number | undefined;
+  } | undefined;
+  chunks?: {
+    id: string;
+    index: number;
+    text: string;
+  }[] | undefined;
+}>;
+declare const documentExtractStatusTool: ToolDefinition<{
+  output?: "artifact" | "chunks" | "inline" | undefined;
+  destination_key?: string | undefined;
+  chunk?: {
+    max_chars?: number | undefined;
+    overlap?: number | undefined;
+  } | undefined;
+  job_id: string;
+}, {
+  status: "failed" | "pending" | "succeeded";
+  job_id?: string | undefined;
+  text?: string | undefined;
+  page_count?: number | undefined;
+  error?: string | undefined;
+  source?: {
+    store: "host" | "object";
+    key: string;
+    media_type?: string | undefined;
+    filename?: string | undefined;
+    byte_length?: number | undefined;
+  } | undefined;
+  output?: "artifact" | "chunks" | "inline" | undefined;
+  artifact?: {
+    store: "host" | "object";
+    key: string;
+    media_type?: string | undefined;
+    filename?: string | undefined;
+    byte_length?: number | undefined;
+  } | undefined;
+  chunks?: {
+    id: string;
+    index: number;
+    text: string;
+  }[] | undefined;
+}>;
+declare const documentExtractTextBatchTool: ToolDefinition<{
+  output?: "artifact" | "chunks" | "inline" | undefined;
+  destination_key?: string | undefined;
+  chunk?: {
+    max_chars?: number | undefined;
+    overlap?: number | undefined;
+  } | undefined;
+  sources: {
+    store: "host" | "object";
+    key: string;
+    media_type?: string | undefined;
+    filename?: string | undefined;
+    byte_length?: number | undefined;
+  }[];
+  destination_key_prefix?: string | undefined;
+}, {
+  results: {
+    index: number;
+    ok: boolean;
+    value?: {
+      status: "failed" | "pending" | "succeeded";
+      job_id?: string | undefined;
+      text?: string | undefined;
+      page_count?: number | undefined;
+      error?: string | undefined;
+      source?: {
+        store: "host" | "object";
+        key: string;
+        media_type?: string | undefined;
+        filename?: string | undefined;
+        byte_length?: number | undefined;
+      } | undefined;
+      output?: "artifact" | "chunks" | "inline" | undefined;
+      artifact?: {
+        store: "host" | "object";
+        key: string;
+        media_type?: string | undefined;
+        filename?: string | undefined;
+        byte_length?: number | undefined;
+      } | undefined;
+      chunks?: {
+        id: string;
+        index: number;
+        text: string;
+      }[] | undefined;
+    } | undefined;
+    error?: {
+      code: string;
+      message: string;
+      retryable?: boolean | undefined;
+    } | undefined;
+  }[];
+  succeeded: number;
+  failed: number;
+}>;
+declare const documentExtractModule: ModuleDefinition<{
+  access_key_id: string;
+  secret_access_key: string;
+  region: string;
+  bucket: string;
+  session_token?: string | undefined;
+  key_prefix?: string | undefined;
+  poll_timeout_ms?: number | undefined;
+  poll_interval_ms?: number | undefined;
+  provider: "textract";
+}>;
+//#endregion
+export { type DocumentExtractAuth, DocumentExtractClient, type ExtractOutputMode, type ExtractResult, type ExtractTextBatchInput, type ExtractTextBatchOutput, type ExtractTextChunk, type ExtractTextInput, MAX_BATCH_EXTRACT, MAX_INLINE_EXTRACT_CHARS, type StatusInput, documentExtractAuthSchema, documentExtractModule, documentExtractStatusTool, documentExtractTextBatchTool, documentExtractTextTool, extractOutputModeSchema, extractResultSchema, extractTextBatchInputSchema, extractTextBatchOutputSchema, extractTextInputSchema, statusInputSchema };
+//# sourceMappingURL=index.d.ts.map

package/dist/modules/document-extract/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","names":[],"sources":["../../../src/modules/document-extract/contracts.ts","../../../src/modules/document-extract/client.ts","../../../src/modules/document-extract/module.ts"],"mappings":";;;;;cAiCa,2BAAyB,EAAA,uBAAA,EAAA;;;;;;;;;;GAAwE,EAAA,KAAA;KAElG,sBAAsB,EAAE,aAAa;;cAGpC,yBAAuB,EAAA;;;;;KACxB,oBAAoB;KAGpB,mBAAmB;cAElB,qBAAmB,EAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAA8B,EAAA,KAAA;cACjD,wBAAsB,EAAA;;;;;;;;;;;;;;;;;;;;;GAAiC,EAAA,KAAA;cACvD,6BAA2B,EAAA;;;;;;;;;;;;;;;;;;;;;;GAAsC,EAAA,KAAA;cACjE,mBAAiB,EAAA;;;;;;;;;;;;GAA4B,EAAA,KAAA;cAC7C,8BAA4B,EAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAAuC,EAAA,KAAA;KAEpE,gBAAgB;KAChB,mBAAmB;KACnB,wBAAwB;KACxB,cAAc;KACd,yBAAyB;;KAGzB;EACX,cAAc,OAAO,qBAAqB,QAAQ;EAClD,YAAY,OAAO,gBAAgB,QAAQ;EAC3C,mBAAmB,OAAO,0BAA0B,QAAQ;;;;cC3BhD,iCAAiC;;EAG7C,YAAY,KAAK;SAIV,YAAY,KAAK,cAAc;SAK/B,SAAS,MAAM,qBAAqB,MAAK,cAAmB;EAInE,YAAY,OAAO,mBAAgB;;;;;;;;;;;;;;;;;;;;;;;;;;;EAInC,UAAU,OAAO,cAAW;;;;;;;;;;;;;;;;;;;;;;;;;;;EAI5B,iBAAiB,OAAO,wBAAqB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;cC3CjC,yBAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;cAYA,2BAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;cAYA,8BAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;cAYA,uBAAA"}

package/dist/modules/document-extract/index.js

@@ -0,0 +1,127 @@
+import { n as defineModule, r as defineTool } from "../../define-CvpJ_pNP.js";
+import { n as requireAuth } from "../../provider-CgDAlg6K.js";
+import { S as textractStatusInputSchema, _ as textractExtractResultSchema, a as TextractClient, b as textractExtractTextInputSchema, f as MAX_INLINE_EXTRACT_CHARS, h as textractAuthSchema, u as MAX_BATCH_EXTRACT, v as textractExtractTextBatchInputSchema, x as textractOutputModeSchema, y as textractExtractTextBatchOutputSchema } from "../../textract-CPLiB22H.js";
+import { z } from "zod";
+//#region src/modules/document-extract/contracts.ts
+/** Host auth: vendor credentials + provider discriminator. */
+const textractDocumentExtractAuthSchema = textractAuthSchema.extend({ provider: z.literal("textract") });
+const documentExtractAuthSchema = z.discriminatedUnion("provider", [textractDocumentExtractAuthSchema]);
+/** Capability I/O — re-export Textract shapes under seam names (vendor owns presentation). */
+const extractOutputModeSchema = textractOutputModeSchema;
+const extractResultSchema = textractExtractResultSchema;
+const extractTextInputSchema = textractExtractTextInputSchema;
+const extractTextBatchInputSchema = textractExtractTextBatchInputSchema;
+const statusInputSchema = textractStatusInputSchema;
+const extractTextBatchOutputSchema = textractExtractTextBatchOutputSchema;
+//#endregion
+//#region src/modules/document-extract/providers/textract.ts
+var TextractDocumentExtractProvider = class {
+	#client;
+	constructor(auth, options = {}) {
+		const { provider: _provider, ...vendorAuth } = auth;
+		this.#client = new TextractClient(vendorAuth, options);
+	}
+	extractText(input) {
+		return this.#client.extractText(input);
+	}
+	getStatus(input) {
+		return this.#client.getStatus(input);
+	}
+	extractTextBatch(input) {
+		return this.#client.extractTextBatch(input);
+	}
+};
+//#endregion
+//#region src/modules/document-extract/client.ts
+/**
+* Document-extract seam client — picks a provider class from host auth.
+* Presentation (inline | artifact | chunks) lives in the vendor client.
+*/
+function transportOptions(ctx) {
+	return {
+		...ctx.fetch && { fetch: ctx.fetch },
+		...ctx.signal && { signal: ctx.signal }
+	};
+}
+function providerFor(auth, ctx) {
+	const options = transportOptions(ctx);
+	switch (auth.provider) {
+		case "textract": return new TextractDocumentExtractProvider(auth, options);
+	}
+}
+var DocumentExtractClient = class DocumentExtractClient {
+	#ops;
+	constructor(ops) {
+		this.#ops = ops;
+	}
+	static fromContext(ctx) {
+		const auth = requireAuth(ctx, documentExtractAuthSchema);
+		return new DocumentExtractClient(providerFor(auth, ctx));
+	}
+	static fromAuth(auth, ctx = {}) {
+		return new DocumentExtractClient(providerFor(auth, ctx));
+	}
+	extractText(input) {
+		return this.#ops.extractText(input);
+	}
+	getStatus(input) {
+		return this.#ops.getStatus(input);
+	}
+	extractTextBatch(input) {
+		return this.#ops.extractTextBatch(input);
+	}
+};
+//#endregion
+//#region src/modules/document-extract/module.ts
+const documentExtractTextTool = defineTool({
+	id: "document-extract-text",
+	name: "extractDocumentText",
+	description: "Extract text from a document ArtifactRef in object storage. Polls until done or timeout. Optional output: inline (default; fails if over size limit), artifact (write text to object storage), or chunks (split for RAG). If still running after the wait budget, returns pending and job_id for document-extract-status.",
+	inputSchema: extractTextInputSchema,
+	outputSchema: extractResultSchema,
+	sideEffect: "read",
+	runtime: "both",
+	execute: async (input, ctx) => DocumentExtractClient.fromContext(ctx).extractText(input)
+});
+const documentExtractStatusTool = defineTool({
+	id: "document-extract-status",
+	name: "getDocumentExtractStatus",
+	description: "Check a text-extraction job by job_id from document-extract-text. Same optional output modes as extract (inline, artifact, chunks). Does not start a new job.",
+	inputSchema: statusInputSchema,
+	outputSchema: extractResultSchema,
+	sideEffect: "read",
+	runtime: "both",
+	execute: async (input, ctx) => DocumentExtractClient.fromContext(ctx).getStatus(input)
+});
+const documentExtractTextBatchTool = defineTool({
+	id: "document-extract-text-batch",
+	name: "extractDocumentTextBatch",
+	description: "Extract text from up to 10 document ArtifactRefs with optional output modes (inline, artifact, chunks). Returns per-item status without aborting the whole batch.",
+	inputSchema: extractTextBatchInputSchema,
+	outputSchema: extractTextBatchOutputSchema,
+	sideEffect: "read",
+	runtime: "both",
+	execute: async (input, ctx) => DocumentExtractClient.fromContext(ctx).extractTextBatch(input)
+});
+const documentExtractModule = defineModule({
+	id: "document-extract",
+	title: "Document Extract",
+	description: "Extract text from scanned or extraction-dependent documents. Choose inline text, a text ArtifactRef, or retrieval chunks. Extraction polls within the initial call; use the status tool only when a job_id is returned as pending.",
+	runtime: "both",
+	auth: {
+		type: "custom",
+		schema: documentExtractAuthSchema
+	},
+	categories: ["documents", "ocr"],
+	classification: "pii",
+	tags: ["textract", "extract"],
+	tools: [
+		documentExtractTextTool,
+		documentExtractStatusTool,
+		documentExtractTextBatchTool
+	]
+});
+//#endregion
+export { DocumentExtractClient, MAX_BATCH_EXTRACT, MAX_INLINE_EXTRACT_CHARS, documentExtractAuthSchema, documentExtractModule, documentExtractStatusTool, documentExtractTextBatchTool, documentExtractTextTool, extractOutputModeSchema, extractResultSchema, extractTextBatchInputSchema, extractTextBatchOutputSchema, extractTextInputSchema, statusInputSchema };
+
+//# sourceMappingURL=index.js.map

package/dist/modules/document-extract/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","names":["#client","#ops"],"sources":["../../../src/modules/document-extract/contracts.ts","../../../src/modules/document-extract/providers/textract.ts","../../../src/modules/document-extract/client.ts","../../../src/modules/document-extract/module.ts"],"sourcesContent":["import { z } from 'zod'\n\nimport {\n\tMAX_BATCH_EXTRACT,\n\tMAX_INLINE_EXTRACT_CHARS,\n\ttextractAuthSchema,\n\ttextractExtractResultSchema,\n\ttextractExtractTextBatchInputSchema,\n\ttextractExtractTextBatchOutputSchema,\n\ttextractExtractTextInputSchema,\n\ttextractOutputModeSchema,\n\ttextractStatusInputSchema,\n\ttextractTextChunkSchema\n} from '../../vendors/textract'\nimport type {\n\tTextractExtractResult,\n\tTextractExtractTextBatchInput,\n\tTextractExtractTextBatchOutput,\n\tTextractExtractTextInput,\n\tTextractOutputMode,\n\tTextractStatusInput,\n\tTextractTextChunk\n} from '../../vendors/textract'\n\nexport { MAX_BATCH_EXTRACT, MAX_INLINE_EXTRACT_CHARS }\n\n/** Host auth: vendor credentials + provider discriminator. */\nexport const textractDocumentExtractAuthSchema = textractAuthSchema.extend({\n\tprovider: z.literal('textract')\n})\n\nexport type TextractDocumentExtractAuth = z.infer<typeof textractDocumentExtractAuthSchema>\n\nexport const documentExtractAuthSchema = z.discriminatedUnion('provider', [textractDocumentExtractAuthSchema])\n\nexport type DocumentExtractAuth = z.infer<typeof documentExtractAuthSchema>\n\n/** Capability I/O — re-export Textract shapes under seam names (vendor owns presentation). */\nexport const extractOutputModeSchema = textractOutputModeSchema\nexport type ExtractOutputMode = TextractOutputMode\n\nexport const extractTextChunkSchema = textractTextChunkSchema\nexport type ExtractTextChunk = TextractTextChunk\n\nexport const extractResultSchema = textractExtractResultSchema\nexport const extractTextInputSchema = textractExtractTextInputSchema\nexport const extractTextBatchInputSchema = textractExtractTextBatchInputSchema\nexport const statusInputSchema = textractStatusInputSchema\nexport const extractTextBatchOutputSchema = textractExtractTextBatchOutputSchema\n\nexport type ExtractResult = TextractExtractResult\nexport type ExtractTextInput = TextractExtractTextInput\nexport type ExtractTextBatchInput = TextractExtractTextBatchInput\nexport type StatusInput = TextractStatusInput\nexport type ExtractTextBatchOutput = TextractExtractTextBatchOutput\n\n/** Provider + public client surface (same ops; vendor does the work). */\nexport type DocumentExtractOps = {\n\textractText: (input: ExtractTextInput) => Promise<ExtractResult>\n\tgetStatus: (input: StatusInput) => Promise<ExtractResult>\n\textractTextBatch: (input: ExtractTextBatchInput) => Promise<ExtractTextBatchOutput>\n}\n","/**\n * Textract provider for the document-extract seam.\n * Thin wrap of `TextractClient` — presentation is owned by the vendor.\n */\n\nimport type { HttpServiceOptions } from '../../../transport/http-service'\nimport { TextractClient } from '../../../vendors/textract'\nimport type {\n\tDocumentExtractOps,\n\tExtractResult,\n\tExtractTextBatchInput,\n\tExtractTextBatchOutput,\n\tExtractTextInput,\n\tStatusInput,\n\tTextractDocumentExtractAuth\n} from '../contracts'\n\nexport type TextractDocumentExtractProviderOptions = Pick<HttpServiceOptions, 'fetch' | 'signal'>\n\nexport class TextractDocumentExtractProvider implements DocumentExtractOps {\n\treadonly #client: TextractClient\n\n\tconstructor(auth: TextractDocumentExtractAuth, options: TextractDocumentExtractProviderOptions = {}) {\n\t\tconst { provider: _provider, ...vendorAuth } = auth\n\t\tthis.#client = new TextractClient(vendorAuth, options)\n\t}\n\n\textractText(input: ExtractTextInput): Promise<ExtractResult> {\n\t\treturn this.#client.extractText(input)\n\t}\n\n\tgetStatus(input: StatusInput): Promise<ExtractResult> {\n\t\treturn this.#client.getStatus(input)\n\t}\n\n\textractTextBatch(input: ExtractTextBatchInput): Promise<ExtractTextBatchOutput> {\n\t\treturn this.#client.extractTextBatch(input)\n\t}\n}\n","/**\n * Document-extract seam client — picks a provider class from host auth.\n * Presentation (inline | artifact | chunks) lives in the vendor client.\n */\n\nimport { requireAuth } from '../../core/provider'\nimport type { ToolContext } from '../../core/types'\nimport type { HttpServiceOptions } from '../../transport/http-service'\nimport { documentExtractAuthSchema } from './contracts'\nimport type {\n\tDocumentExtractAuth,\n\tDocumentExtractOps,\n\tExtractTextBatchInput,\n\tExtractTextInput,\n\tStatusInput\n} from './contracts'\nimport { TextractDocumentExtractProvider } from './providers/textract'\n\nfunction transportOptions(ctx: ToolContext): Pick<HttpServiceOptions, 'fetch' | 'signal'> {\n\treturn {\n\t\t...(ctx.fetch && { fetch: ctx.fetch }),\n\t\t...(ctx.signal && { signal: ctx.signal })\n\t}\n}\n\nfunction providerFor(auth: DocumentExtractAuth, ctx: ToolContext): DocumentExtractOps {\n\tconst options = transportOptions(ctx)\n\tswitch (auth.provider) {\n\t\tcase 'textract':\n\t\t\treturn new TextractDocumentExtractProvider(auth, options)\n\t}\n}\n\nexport class DocumentExtractClient implements DocumentExtractOps {\n\treadonly #ops: DocumentExtractOps\n\n\tconstructor(ops: DocumentExtractOps) {\n\t\tthis.#ops = ops\n\t}\n\n\tstatic fromContext(ctx: ToolContext): DocumentExtractClient {\n\t\tconst auth = requireAuth(ctx, documentExtractAuthSchema)\n\t\treturn new DocumentExtractClient(providerFor(auth, ctx))\n\t}\n\n\tstatic fromAuth(auth: DocumentExtractAuth, ctx: ToolContext = {}): DocumentExtractClient {\n\t\treturn new DocumentExtractClient(providerFor(auth, ctx))\n\t}\n\n\textractText(input: ExtractTextInput) {\n\t\treturn this.#ops.extractText(input)\n\t}\n\n\tgetStatus(input: StatusInput) {\n\t\treturn this.#ops.getStatus(input)\n\t}\n\n\textractTextBatch(input: ExtractTextBatchInput) {\n\t\treturn this.#ops.extractTextBatch(input)\n\t}\n}\n","import { defineModule, defineTool } from '../../core/define'\nimport { DocumentExtractClient } from './client'\nimport {\n\tdocumentExtractAuthSchema,\n\textractResultSchema,\n\textractTextBatchInputSchema,\n\textractTextBatchOutputSchema,\n\textractTextInputSchema,\n\tstatusInputSchema\n} from './contracts'\n\nexport type { DocumentExtractAuth } from './contracts'\nexport { documentExtractAuthSchema }\n\nexport const documentExtractTextTool = defineTool({\n\tid: 'document-extract-text',\n\tname: 'extractDocumentText',\n\tdescription:\n\t\t'Extract text from a document ArtifactRef in object storage. Polls until done or timeout. Optional output: inline (default; fails if over size limit), artifact (write text to object storage), or chunks (split for RAG). If still running after the wait budget, returns pending and job_id for document-extract-status.',\n\tinputSchema: extractTextInputSchema,\n\toutputSchema: extractResultSchema,\n\tsideEffect: 'read',\n\truntime: 'both',\n\texecute: async (input, ctx) => DocumentExtractClient.fromContext(ctx).extractText(input)\n})\n\nexport const documentExtractStatusTool = defineTool({\n\tid: 'document-extract-status',\n\tname: 'getDocumentExtractStatus',\n\tdescription:\n\t\t'Check a text-extraction job by job_id from document-extract-text. Same optional output modes as extract (inline, artifact, chunks). Does not start a new job.',\n\tinputSchema: statusInputSchema,\n\toutputSchema: extractResultSchema,\n\tsideEffect: 'read',\n\truntime: 'both',\n\texecute: async (input, ctx) => DocumentExtractClient.fromContext(ctx).getStatus(input)\n})\n\nexport const documentExtractTextBatchTool = defineTool({\n\tid: 'document-extract-text-batch',\n\tname: 'extractDocumentTextBatch',\n\tdescription:\n\t\t'Extract text from up to 10 document ArtifactRefs with optional output modes (inline, artifact, chunks). Returns per-item status without aborting the whole batch.',\n\tinputSchema: extractTextBatchInputSchema,\n\toutputSchema: extractTextBatchOutputSchema,\n\tsideEffect: 'read',\n\truntime: 'both',\n\texecute: async (input, ctx) => DocumentExtractClient.fromContext(ctx).extractTextBatch(input)\n})\n\nexport const documentExtractModule = defineModule({\n\tid: 'document-extract',\n\ttitle: 'Document Extract',\n\tdescription:\n\t\t'Extract text from scanned or extraction-dependent documents. Choose inline text, a text ArtifactRef, or retrieval chunks. Extraction polls within the initial call; use the status tool only when a job_id is returned as pending.',\n\truntime: 'both',\n\tauth: { type: 'custom', schema: documentExtractAuthSchema },\n\tcategories: ['documents', 'ocr'],\n\tclassification: 'pii',\n\ttags: ['textract', 'extract'],\n\ttools: [documentExtractTextTool, documentExtractStatusTool, documentExtractTextBatchTool]\n})\n"],"mappings":";;;;;;AA2BA,MAAa,oCAAoC,mBAAmB,OAAO,EAC1E,UAAU,EAAE,QAAQ,UAAU,EAC/B,CAAC;AAID,MAAa,4BAA4B,EAAE,mBAAmB,YAAY,CAAC,iCAAiC,CAAC;;AAK7G,MAAa,0BAA0B;AAMvC,MAAa,sBAAsB;AACnC,MAAa,yBAAyB;AACtC,MAAa,8BAA8B;AAC3C,MAAa,oBAAoB;AACjC,MAAa,+BAA+B;;;AC7B5C,IAAa,kCAAb,MAA2E;CAC1E;CAEA,YAAY,MAAmC,UAAkD,CAAC,GAAG;EACpG,MAAM,EAAE,UAAU,WAAW,GAAG,eAAe;EAC/C,KAAKA,UAAU,IAAI,eAAe,YAAY,OAAO;CACtD;CAEA,YAAY,OAAiD;EAC5D,OAAO,KAAKA,QAAQ,YAAY,KAAK;CACtC;CAEA,UAAU,OAA4C;EACrD,OAAO,KAAKA,QAAQ,UAAU,KAAK;CACpC;CAEA,iBAAiB,OAA+D;EAC/E,OAAO,KAAKA,QAAQ,iBAAiB,KAAK;CAC3C;AACD;;;;;;;ACpBA,SAAS,iBAAiB,KAAgE;CACzF,OAAO;EACN,GAAI,IAAI,SAAS,EAAE,OAAO,IAAI,MAAM;EACpC,GAAI,IAAI,UAAU,EAAE,QAAQ,IAAI,OAAO;CACxC;AACD;AAEA,SAAS,YAAY,MAA2B,KAAsC;CACrF,MAAM,UAAU,iBAAiB,GAAG;CACpC,QAAQ,KAAK,UAAb;EACC,KAAK,YACJ,OAAO,IAAI,gCAAgC,MAAM,OAAO;CAC1D;AACD;AAEA,IAAa,wBAAb,MAAa,sBAAoD;CAChE;CAEA,YAAY,KAAyB;EACpC,KAAKC,OAAO;CACb;CAEA,OAAO,YAAY,KAAyC;EAC3D,MAAM,OAAO,YAAY,KAAK,yBAAyB;EACvD,OAAO,IAAI,sBAAsB,YAAY,MAAM,GAAG,CAAC;CACxD;CAEA,OAAO,SAAS,MAA2B,MAAmB,CAAC,GAA0B;EACxF,OAAO,IAAI,sBAAsB,YAAY,MAAM,GAAG,CAAC;CACxD;CAEA,YAAY,OAAyB;EACpC,OAAO,KAAKA,KAAK,YAAY,KAAK;CACnC;CAEA,UAAU,OAAoB;EAC7B,OAAO,KAAKA,KAAK,UAAU,KAAK;CACjC;CAEA,iBAAiB,OAA8B;EAC9C,OAAO,KAAKA,KAAK,iBAAiB,KAAK;CACxC;AACD;;;AC9CA,MAAa,0BAA0B,WAAW;CACjD,IAAI;CACJ,MAAM;CACN,aACC;CACD,aAAa;CACb,cAAc;CACd,YAAY;CACZ,SAAS;CACT,SAAS,OAAO,OAAO,QAAQ,sBAAsB,YAAY,GAAG,CAAC,CAAC,YAAY,KAAK;AACxF,CAAC;AAED,MAAa,4BAA4B,WAAW;CACnD,IAAI;CACJ,MAAM;CACN,aACC;CACD,aAAa;CACb,cAAc;CACd,YAAY;CACZ,SAAS;CACT,SAAS,OAAO,OAAO,QAAQ,sBAAsB,YAAY,GAAG,CAAC,CAAC,UAAU,KAAK;AACtF,CAAC;AAED,MAAa,+BAA+B,WAAW;CACtD,IAAI;CACJ,MAAM;CACN,aACC;CACD,aAAa;CACb,cAAc;CACd,YAAY;CACZ,SAAS;CACT,SAAS,OAAO,OAAO,QAAQ,sBAAsB,YAAY,GAAG,CAAC,CAAC,iBAAiB,KAAK;AAC7F,CAAC;AAED,MAAa,wBAAwB,aAAa;CACjD,IAAI;CACJ,OAAO;CACP,aACC;CACD,SAAS;CACT,MAAM;EAAE,MAAM;EAAU,QAAQ;CAA0B;CAC1D,YAAY,CAAC,aAAa,KAAK;CAC/B,gBAAgB;CAChB,MAAM,CAAC,YAAY,SAAS;CAC5B,OAAO;EAAC;EAAyB;EAA2B;CAA4B;AACzF,CAAC"}