@vertesia/workflow 0.42.2

package/lib/cjs/activities/extractDocumentText.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"extractDocumentText.js","sourceRoot":"","sources":["../../../src/activities/extractDocumentText.ts"],"names":[],"mappings":";;AAwBA,kDA6HC;AApJD,mDAA2C;AAC3C,uDAA0D;AAC1D,uDAAyD;AACzD,iDAA6D;AAC7D,wEAAgE;AAChE,4CAA+C;AAC/C,wDAAgF;AAChF,gDAA2D;AAC3D,kDAAiD;AAEjD,YAAY;AACZ,MAAM,IAAI,GAAoB;IAC1B,IAAI,EAAE,qBAAqB;CAC9B,CAAA;AAUM,KAAK,UAAU,mBAAmB,CAAC,OAAoC;IAC1E,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,GAAG,MAAM,IAAA,kCAAa,EAAC,OAAO,CAAC,CAAC;IAE1D,MAAM,CAAC,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC;QAChC,KAAK,EAAE,EAAE,GAAG,EAAE,QAAQ,EAAE;QACxB,KAAK,EAAE,CAAC;QACR,MAAM,EAAE,OAAO;KAClB,CAAC,CAAA;IACF,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,CAAkB,CAAC;IAClC,IAAI,CAAC,GAAG,EAAE,CAAC;QACP,cAAG,CAAC,KAAK,CAAC,YAAY,QAAQ,YAAY,CAAC,CAAC;QAC5C,MAAM,IAAI,2BAAe,CAAC,YAAY,QAAQ,YAAY,EAAE,OAAO,CAAC,SAAS,CAAC,CAAC;IACnF,CAAC;IAED,cAAG,CAAC,IAAI,CAAC,8BAA8B,GAAG,CAAC,EAAE,EAAE,CAAC,CAAC;IAGjD,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,IAAI,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,MAAM,EAAE,CAAC;QAC7C,IAAI,GAAG,CAAC,IAAI,EAAE,CAAC;YACX,OAAO,cAAc,CAAC,GAAG,EAAE,GAAG,CAAC,IAAI,EAAE,sCAAoB,CAAC,OAAO,EAAE,oCAAoC,CAAC,CAAC;QAC7G,CAAC;aAAM,CAAC;YACJ,OAAO,cAAc,CAAC,GAAG,EAAE,EAAE,EAAE,sCAAoB,CAAC,KAAK,EAAE,yBAAyB,CAAC,CAAC;QAC1F,CAAC;IACL,CAAC;IAED,gDAAgD;IAChD,IAAI,GAAG,CAAC,IAAI,IAAI,GAAG,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,GAAG,CAAC,SAAS,KAAK,GAAG,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC;QACxE,OAAO,cAAc,CAAC,GAAG,EAAE,GAAG,CAAC,IAAI,EAAE,sCAAoB,CAAC,OAAO,EAAE,wBAAwB,CAAC,CAAC;IACjG,CAAC;IAED,IAAI,UAAkB,CAAC;IACvB,IAAI,CAAC;QACD,UAAU,GAAG,MAAM,IAAA,4BAAiB,EAAC,MAAM,EAAE,GAAG,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;IACrE,CAAC;IAAC,OAAO,CAAM,EAAE,CAAC;QACd,cAAG,CAAC,KAAK,CAAC,uBAAuB,CAAC,EAAE,CAAC,CAAC;QACtC,OAAO,cAAc,CAAC,GAAG,EAAE,EAAE,EAAE,sCAAoB,CAAC,KAAK,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC;IAC1E,CAAC;IAGD,IAAI,GAAW,CAAC;IAEhB,QAAQ,GAAG,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC;QAEvB,KAAK,iBAAiB;YAClB,qCAAqC;YACrC,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,GAAG,IAAI,GAAG,IAAI,EAAE,CAAC;gBACtC,GAAG,GAAG,MAAM,IAAA,2BAAe,EAAC,UAAU,CAAC,CAAC;YAC5C,CAAC;iBAAM,CAAC;gBACJ,GAAG,GAAG,MAAM,IAAA,8BAAqB,EAAC,UAAU,CAAC,CAAC;YAClD,CAAC;YACD,MAAM;QAEV,KAAK,YAAY;YACb,GAAG,GAAG,UAAU,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAA;YACjC,MAAM;QAEV,MAAM;QACN,KAAK,yEAAyE;YAC1E,GAAG,GAAG,MAAM,IAAA,0BAAc,EAAC,UAAU,EAAE,MAAM,CAAC,CAAC;YAC/C,MAAM;QAEV,MAAM;QACN,KAAK,WAAW;YACZ,GAAG,GAAG,MAAM,IAAA,0BAAc,EAAC,UAAU,EAAE,MAAM,CAAC,CAAC;YAC/C,MAAM;QAEV,cAAc;QACd,KAAK,yCAAyC;YAC1C,GAAG,GAAG,MAAM,IAAA,0BAAc,EAAC,UAAU,EAAE,KAAK,CAAC,CAAC;YAC9C,MAAM;QAEV,KAAK;QACL,KAAK,iBAAiB;YAClB,GAAG,GAAG,MAAM,IAAA,0BAAc,EAAC,UAAU,EAAE,KAAK,CAAC,CAAC;YAC9C,MAAM;QAEV,UAAU;QACV,KAAK,eAAe;YAChB,GAAG,GAAG,UAAU,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;YAClC,MAAM;QAEV,KAAK;QACL,KAAK,UAAU;YACX,GAAG,GAAG,UAAU,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;YAClC,MAAM;QAEV,YAAY;QACZ,KAAK,wBAAwB;YACzB,GAAG,GAAG,UAAU,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;YAClC,MAAM;QAEV,YAAY;QACZ,KAAK,wBAAwB;YACzB,GAAG,GAAG,UAAU,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;YAClC,MAAM;QAEV,MAAM;QACN,KAAK,kBAAkB;YACnB,GAAG,GAAG,UAAU,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;YAClC,MAAM;QAEV;YACI,IAAI,WAAW,CAAC,UAAU,CAAC,EAAE,CAAC;gBAC1B,GAAG,GAAG,UAAU,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,6BAA6B;gBAChE,MAAM;YACV,CAAC;YACD,OAAO,cAAc,CAAC,GAAG,EAAE,GAAG,CAAC,IAAI,IAAI,EAAE,EAAE,sCAAoB,CAAC,OAAO,EAAE,0BAA0B,GAAG,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,CAAC;IAC/H,CAAC;IAGD,MAAM,UAAU,GAAG,IAAA,uBAAW,EAAC,GAAG,CAAC,CAAC;IACpC,MAAM,IAAI,GAAG,GAAG,CAAC,OAAO,CAAC,IAAI,IAAI,IAAA,cAAG,EAAC,GAAG,CAAC,CAAC;IAE1C,MAAM,UAAU,GAA+B;QAC3C,IAAI,EAAE,GAAG;QACT,SAAS,EAAE,IAAI;QACf,MAAM,EAAE;YACJ,GAAG,UAAU;YACb,IAAI,EAAE,IAAI;SACb;KACJ,CAAA;IAED,MAAM,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,EAAE,UAAU,CAAC,CAAC;IAEhD,OAAO,cAAc,CAAC,GAAG,EAAE,GAAG,EAAE,sCAAoB,CAAC,OAAO,CAAC,CAAC;AAClE,CAAC;AAED,SAAS,cAAc,CAAC,GAAkB,EAAE,IAAY,EAAE,MAA4B,EAAE,OAAgB;IACpG,OAAO;QACH,MAAM;QACN,OAAO;QACP,MAAM,EAAE,GAAG,CAAC,MAAM;QAClB,GAAG,EAAE,IAAI,CAAC,MAAM;QAChB,QAAQ,EAAE,GAAG,CAAC,EAAE;QAChB,OAAO,EAAE,CAAC,CAAC,IAAI;KAClB,CAAA;AAEL,CAAC;AAGD,yDAAyD;AACzD,SAAS,WAAW,CAAC,GAAW;IAC5B,IAAI,GAAG,CAAC,MAAM,GAAG,GAAG,GAAG,IAAI,EAAE,CAAC;QAC1B,MAAM,CAAC,GAAG,GAAG,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;QAC/B,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACf,OAAO,IAAI,CAAC;QAChB,CAAC;IACL,CAAC;IACD,OAAO,KAAK,CAAC;AACjB,CAAC"}

package/lib/cjs/activities/generateDocumentProperties.js

@@ -0,0 +1,59 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.generateDocumentProperties = generateDocumentProperties;
+const activity_1 = require("@temporalio/activity");
+const ActivityContext_js_1 = require("../dsl/setup/ActivityContext.js");
+const executeInteraction_js_1 = require("./executeInteraction.js");
+const INT_EXTRACT_INFORMATION = "sys:ExtractInformation";
+async function generateDocumentProperties(payload) {
+    const context = await (0, ActivityContext_js_1.setupActivity)(payload);
+    const { params, client, objectId } = context;
+    const interactionName = params.interactionName ?? INT_EXTRACT_INFORMATION;
+    const project = await context.fetchProject();
+    const doc = await client.objects.retrieve(objectId, "+text");
+    const type = doc.type ? await client.types.retrieve(doc.type.id) : undefined;
+    if (!doc?.text && !params.use_vision && !doc?.content?.type?.startsWith("image/")) {
+        activity_1.log.warn(`Object ${objectId} not found or text is empty`);
+        return { status: "failed", error: "no-text" };
+    }
+    if (!type || !type.object_schema) {
+        activity_1.log.info(`Object ${objectId} has no schema`);
+        return { document: objectId, status: "skipped", message: "no schema defined on type" };
+    }
+    const getImageRef = () => {
+        if (doc.content?.type?.startsWith("image/")) {
+            return "store:" + doc.id;
+        }
+        if (params.use_vision && doc.content?.type?.startsWith("application/pdf")) {
+            return "store:" + doc.id;
+        }
+        activity_1.log.info(`Object ${objectId} is not an image or pdf`);
+        return undefined;
+    };
+    const promptData = {
+        content: doc.text ?? undefined,
+        image: getImageRef() ?? undefined,
+        human_context: project?.configuration?.human_context ?? undefined,
+    };
+    activity_1.log.info(` Extracting information from object ${objectId} with type ${type.name}`, payload.debug_mode ? { params, } : undefined);
+    const infoRes = await (0, executeInteraction_js_1.executeInteractionFromActivity)(client, interactionName, {
+        ...params,
+        include_previous_error: true,
+        result_schema: type.object_schema,
+    }, promptData, payload.debug_mode ?? false);
+    activity_1.log.info(`Extracted information from object ${objectId} with type ${type.name}`, { runId: infoRes.id });
+    await client.objects.update(doc.id, {
+        properties: {
+            ...infoRes.result,
+            etag: doc.text_etag
+        },
+        text: infoRes.result.description ?? undefined,
+        generation_run_info: {
+            id: infoRes.id,
+            date: new Date().toISOString(),
+            model: infoRes.modelId,
+        }
+    });
+    return { status: "completed" };
+}
+//# sourceMappingURL=generateDocumentProperties.js.map

package/lib/cjs/activities/generateDocumentProperties.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"generateDocumentProperties.js","sourceRoot":"","sources":["../../../src/activities/generateDocumentProperties.ts"],"names":[],"mappings":";;AAsBA,gEAsEC;AA3FD,mDAA2C;AAC3C,wEAAgE;AAEhE,mEAAqG;AAErG,MAAM,uBAAuB,GAAG,wBAAwB,CAAA;AAgBjD,KAAK,UAAU,0BAA0B,CAAC,OAAoC;IACjF,MAAM,OAAO,GAAG,MAAM,IAAA,kCAAa,EAAmC,OAAO,CAAC,CAAC;IAC/E,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,GAAG,OAAO,CAAC;IAC7C,MAAM,eAAe,GAAG,MAAM,CAAC,eAAe,IAAI,uBAAuB,CAAC;IAE1E,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,YAAY,EAAE,CAAC;IAE7C,MAAM,GAAG,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;IAC7D,MAAM,IAAI,GAAG,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;IAE7E,IAAI,CAAC,GAAG,EAAE,IAAI,IAAI,CAAC,MAAM,CAAC,UAAU,IAAI,CAAC,GAAG,EAAE,OAAO,EAAE,IAAI,EAAE,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;QAChF,cAAG,CAAC,IAAI,CAAC,UAAU,QAAQ,6BAA6B,CAAC,CAAC;QAC1D,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,KAAK,EAAE,SAAS,EAAE,CAAA;IACjD,CAAC;IAED,IAAI,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,aAAa,EAAE,CAAC;QAC/B,cAAG,CAAC,IAAI,CAAC,UAAU,QAAQ,gBAAgB,CAAC,CAAC;QAC7C,OAAO,EAAE,QAAQ,EAAE,QAAQ,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,2BAA2B,EAAE,CAAC;IAC3F,CAAC;IAED,MAAM,WAAW,GAAG,GAAG,EAAE;QACrB,IAAI,GAAG,CAAC,OAAO,EAAE,IAAI,EAAE,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;YAC1C,OAAO,QAAQ,GAAG,GAAG,CAAC,EAAE,CAAC;QAC7B,CAAC;QAED,IAAI,MAAM,CAAC,UAAU,IAAI,GAAG,CAAC,OAAO,EAAE,IAAI,EAAE,UAAU,CAAC,iBAAiB,CAAC,EAAE,CAAC;YACxE,OAAO,QAAQ,GAAG,GAAG,CAAC,EAAE,CAAC;QAC7B,CAAC;QAED,cAAG,CAAC,IAAI,CAAC,UAAU,QAAQ,yBAAyB,CAAC,CAAC;QACtD,OAAO,SAAS,CAAA;IACpB,CAAC,CAAA;IAED,MAAM,UAAU,GAAG;QACf,OAAO,EAAE,GAAG,CAAC,IAAI,IAAI,SAAS;QAC9B,KAAK,EAAE,WAAW,EAAE,IAAI,SAAS;QACjC,aAAa,EAAE,OAAO,EAAE,aAAa,EAAE,aAAa,IAAI,SAAS;KACpE,CAAA;IAED,cAAG,CAAC,IAAI,CAAC,uCAAuC,QAAQ,cAAc,IAAI,CAAC,IAAI,EAAE,EAAE,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,EAAE,MAAM,GAAG,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC;IAEjI,MAAM,OAAO,GAAG,MAAM,IAAA,sDAA8B,EAChD,MAAM,EACN,eAAe,EACf;QACI,GAAG,MAAM;QACT,sBAAsB,EAAE,IAAI;QAC5B,aAAa,EAAE,IAAI,CAAC,aAAa;KACpC,EACD,UAAU,EACV,OAAO,CAAC,UAAU,IAAI,KAAK,CAC9B,CAAC;IAEF,cAAG,CAAC,IAAI,CAAC,qCAAqC,QAAQ,cAAc,IAAI,CAAC,IAAI,EAAE,EAAE,EAAE,KAAK,EAAE,OAAO,CAAC,EAAE,EAAE,CAAC,CAAC;IACxG,MAAM,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,EAAE;QAChC,UAAU,EAAE;YACR,GAAG,OAAO,CAAC,MAAM;YACjB,IAAI,EAAE,GAAG,CAAC,SAAS;SACtB;QACD,IAAI,EAAE,OAAO,CAAC,MAAM,CAAC,WAAW,IAAI,SAAS;QAC7C,mBAAmB,EAAE;YACjB,EAAE,EAAE,OAAO,CAAC,EAAE;YACd,IAAI,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;YAC9B,KAAK,EAAE,OAAO,CAAC,OAAO;SACzB;KACJ,CAAC,CAAC;IAGH,OAAO,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;AAEnC,CAAC"}

package/lib/cjs/activities/generateEmbeddings.js

@@ -0,0 +1,292 @@
+"use strict";
+var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    var desc = Object.getOwnPropertyDescriptor(m, k);
+    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
+      desc = { enumerable: true, get: function() { return m[k]; } };
+    }
+    Object.defineProperty(o, k2, desc);
+}) : (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    o[k2] = m[k];
+}));
+var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
+    Object.defineProperty(o, "default", { enumerable: true, value: v });
+}) : function(o, v) {
+    o["default"] = v;
+});
+var __importStar = (this && this.__importStar) || (function () {
+    var ownKeys = function(o) {
+        ownKeys = Object.getOwnPropertyNames || function (o) {
+            var ar = [];
+            for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
+            return ar;
+        };
+        return ownKeys(o);
+    };
+    return function (mod) {
+        if (mod && mod.__esModule) return mod;
+        var result = {};
+        if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
+        __setModuleDefault(result, mod);
+        return result;
+    };
+})();
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.generateEmbeddings = generateEmbeddings;
+const common_1 = require("@vertesia/common");
+const activity_1 = require("@temporalio/activity");
+const tf = __importStar(require("@tensorflow/tfjs-node"));
+const ActivityContext_js_1 = require("../dsl/setup/ActivityContext.js");
+const errors_js_1 = require("../errors.js");
+const blobs_js_1 = require("../utils/blobs.js");
+const tokens_js_1 = require("../utils/tokens.js");
+async function generateEmbeddings(payload) {
+    const { params, client, objectId, fetchProject } = await (0, ActivityContext_js_1.setupActivity)(payload);
+    const { force, type } = params;
+    const projectData = await fetchProject();
+    const config = projectData?.configuration.embeddings[type];
+    if (!projectData) {
+        throw new errors_js_1.NoDocumentFound('Project not found', [payload.project_id]);
+    }
+    if (!config) {
+        throw new errors_js_1.NoDocumentFound('Embeddings configuration not found', [objectId]);
+    }
+    if (!projectData) {
+        throw new errors_js_1.NoDocumentFound('Project not found', [payload.project_id]);
+    }
+    if (!projectData?.configuration.embeddings[type]?.enabled) {
+        activity_1.log.info(`Embeddings generation disabled for type ${type} on project: ${projectData.name} (${projectData.namespace})`, { config });
+        return { id: objectId, status: "skipped", message: `Embeddings generation disabled for type ${type}` };
+    }
+    activity_1.log.info(`${type} embedding generation starting for object ${objectId}`, { force, config });
+    if (!config.environment) {
+        throw new Error('No environment found in project configuration. Set environment in project configuration to generate embeddings.');
+    }
+    const document = await client.objects.retrieve(objectId, "+text +parts +embeddings +tokens +properties");
+    if (!document) {
+        throw new errors_js_1.NoDocumentFound('Document not found', [objectId]);
+    }
+    if (!document.content) {
+        throw new errors_js_1.NoDocumentFound('Document content not found', [objectId]);
+    }
+    let res;
+    switch (type) {
+        case common_1.SupportedEmbeddingTypes.text:
+            res = await generateTextEmbeddings({
+                client,
+                config,
+                document,
+                type
+            });
+            break;
+        case common_1.SupportedEmbeddingTypes.properties:
+            res = await generateTextEmbeddings({
+                client,
+                config,
+                document,
+                type,
+            });
+            break;
+        case common_1.SupportedEmbeddingTypes.image:
+            res = await generateImageEmbeddings({
+                client,
+                config,
+                document,
+                type
+            });
+            break;
+        default:
+            res = { id: objectId, status: "failed", message: `unsupported embedding type: ${type}` };
+    }
+    return res;
+}
+async function generateTextEmbeddings({ document, client, type, config }) {
+    // if (!force && document.embeddings[type]?.etag === (document.text_etag ?? md5(document.text))) {
+    //     return { id: objectId, status: "skipped", message: "embeddings already generated" }
+    // }
+    if (!document) {
+        return { status: "error", message: "document is null or undefined" };
+    }
+    if (type !== common_1.SupportedEmbeddingTypes.text && type !== common_1.SupportedEmbeddingTypes.properties) {
+        return { id: document.id, status: "failed", message: `unsupported embedding type: ${type}` };
+    }
+    if (type === common_1.SupportedEmbeddingTypes.text && !document.text) {
+        return { id: document.id, status: "failed", message: "no text found" };
+    }
+    if (type === common_1.SupportedEmbeddingTypes.properties && !document?.properties) {
+        return { id: document.id, status: "failed", message: "no properties found" };
+    }
+    const { environment, model } = config;
+    // Count tokens if not already done
+    if (!document.tokens?.count && type === common_1.SupportedEmbeddingTypes.text) {
+        activity_1.log.debug('Updating token count for document: ' + document.id);
+        const tokensData = (0, tokens_js_1.countTokens)(document.text);
+        await client.objects.update(document.id, {
+            tokens: {
+                ...tokensData,
+                etag: document.text_etag ?? (0, blobs_js_1.md5)(document.text)
+            }
+        });
+        document.tokens = {
+            ...tokensData,
+            etag: document.text_etag ?? (0, blobs_js_1.md5)(document.text)
+        };
+    }
+    const maxTokens = config.max_tokens ?? 8000;
+    //generate embeddings for the main doc if document isn't too large
+    //if too large, we'll just generate embeddings for the parts
+    //then we can generate embeddings for the main document by averaging the tensors
+    activity_1.log.info(`Generating ${type} embeddings for document ${document.id}`);
+    if (type === common_1.SupportedEmbeddingTypes.text && document.tokens?.count && document.tokens?.count > maxTokens) {
+        activity_1.log.info('Document too large, generating embeddings for parts');
+        if (!document.parts || document.parts.length === 0) {
+            return { id: document.id, status: "skipped", message: "no parts found" };
+        }
+        const docParts = await Promise.all(document.parts?.map(async (partId) => client.objects.retrieve(partId, "+text +embeddings +properties +tokens")));
+        activity_1.log.info(`Retrieved ${docParts.length} parts`);
+        const generatePartEmbeddings = async (part, i) => {
+            try {
+                activity_1.log.info(`Generating embeddings for part ${part.id}`, { text_len: part.text?.length });
+                if (!part.text) {
+                    return { id: part.id, number: i, result: null, status: "skipped", message: "no text found" };
+                }
+                if (part.tokens?.count && part.tokens.count > maxTokens) {
+                    activity_1.log.info('Part too large, skipping embeddings generation for part', { part: part.id, tokens: part.tokens.count });
+                    return { id: part.id, number: i, result: null, message: "part too large" };
+                }
+                const e = await generateEmbeddingsFromStudio(part.text, environment, client, model).catch(e => {
+                    activity_1.log.error('Error generating embeddings for part', { part: part.id, tokens: part.tokens, text_length: part.text?.length, error: e });
+                    return null;
+                });
+                if (!e || !e.values) {
+                    return { id: part.id, number: i, result: null, message: "no embeddings generated" };
+                }
+                activity_1.log.info(`Embeddings generated for part ${part.id}, updating object in the store.`);
+                await client.objects.setEmbedding(part.id, common_1.SupportedEmbeddingTypes.text, {
+                    values: e.values,
+                    model: e.model,
+                    etag: part.text_etag
+                }).catch(err => {
+                    activity_1.log.info(`Error updating embeddings on part ${part.id}`);
+                    return { id: part.id, number: i, result: null, message: "error setting embeddings on part", error: err.message };
+                });
+                activity_1.log.info('Generated embeddings for part: ' + part.id);
+                return { id: part.id, number: i, result: e };
+            }
+            catch (err) {
+                activity_1.log.info(`Error generating ${type} embeddings for part ${part.id} of ${document.id}`, { error: err });
+                return { id: part.id, number: i, result: null, message: "error generating embeddings", error: err.message };
+            }
+        };
+        const promises = docParts.map((p, i) => generatePartEmbeddings(p, i));
+        const res = await Promise.all(promises);
+        // let i = 0;
+        // for (const p of docParts) {
+        //     log.info(`Processing part ${p.id}`)
+        //     const r = await generatePartEmbeddings(p, i++);
+        //     res.push(r)
+        // }
+        // Filter out parts without embeddings
+        const validEmbeddings = res.filter(item => item.result !== null);
+        // Compute the document-level embedding using TensorFlow for attention mechanism
+        activity_1.log.info('Computing document-level embedding using TF');
+        const documentEmbedding = computeAttentionEmbedding(validEmbeddings.map(item => item.result.values));
+        // Save the document-level embedding
+        await client.objects.setEmbedding(document.id, type, {
+            values: documentEmbedding,
+            model: "attention",
+            etag: document.text_etag
+        });
+        return { id: document.id, status: "completed", parts: docParts.map(i => i.id), len: documentEmbedding.length, part_embeddings: res.map(r => { return { id: r.id, status: r.status, error: r.error, message: r.message }; }) };
+    }
+    else {
+        activity_1.log.info(`Generating ${type} embeddings for document`);
+        const res = await generateEmbeddingsFromStudio(JSON.stringify(document[type]), environment, client);
+        if (!res || !res.values) {
+            return { id: document.id, status: "failed", message: "no embeddings generated" };
+        }
+        activity_1.log.info(`${type} embeddings generated for document ${document.id}`, { len: res.values.length });
+        await client.objects.setEmbedding(document.id, type, {
+            values: res.values,
+            model: res.model,
+            etag: document.text_etag
+        });
+        return { id: document.id, type, status: "completed", len: res.values.length };
+    }
+}
+async function generateImageEmbeddings({ document, client, type, config }) {
+    activity_1.log.info('Generating image embeddings for document ' + document.id, { content: document.content });
+    if (!document.content?.type?.startsWith('image/') && !document.content?.type?.includes('pdf')) {
+        return { id: document.id, type, status: "failed", message: "content is not an image" };
+    }
+    const { environment, model } = config;
+    const resRnd = await client.store.objects.getRendition(document.id, {
+        format: "image/png",
+        max_hw: 1024,
+        generate_if_missing: true
+    });
+    if (resRnd.status === 'generating') {
+        throw new Error("Rendition is generating, will retry later");
+    }
+    else if (resRnd.status === "failed" || !resRnd.rendition) {
+        throw new errors_js_1.NoDocumentFound("Rendition retrieval failed", [document.id]);
+    }
+    if (!resRnd.rendition.content.source) {
+        throw new errors_js_1.NoDocumentFound("No source found in rendition", [document.id]);
+    }
+    const image = await (0, blobs_js_1.fetchBlobAsBase64)(client, resRnd.rendition.content.source);
+    const res = await client.environments.embeddings(environment, {
+        image,
+        model
+    }).then(res => res).catch(e => {
+        activity_1.log.error('Error generating embeddings for image', { error: e });
+        throw e;
+    });
+    if (!res || !res.values) {
+        return { id: document.id, status: "failed", message: "no embeddings generated" };
+    }
+    await client.objects.setEmbedding(document.id, common_1.SupportedEmbeddingTypes.image, {
+        values: res.values,
+        model: res.model,
+        etag: document.text_etag
+    });
+    return { id: document.id, type, status: "completed", len: res.values.length };
+}
+async function generateEmbeddingsFromStudio(text, env, client, model) {
+    activity_1.log.info(`Generating embeddings for text of ${text.length} chars with environment ${env}`);
+    return client.environments.embeddings(env, {
+        text,
+        model
+    }).then(res => res).catch(e => {
+        activity_1.log.error('Error generating embeddings for text', { error: e });
+        throw e;
+    });
+}
+function computeAttentionEmbedding(embeddingsArray, axis = 0) {
+    if (embeddingsArray.length === 0)
+        return [];
+    activity_1.log.info('Computing attention embedding for', { embeddingsArrays: embeddingsArray.map(a => a.length) });
+    const start = new Date().getTime();
+    // Convert embeddings array to TensorFlow tensor
+    const embeddingsTensor = tf.tensor(embeddingsArray);
+    // Initialize trainable attention weights
+    const attentionWeights = tf.variable(tf.randomNormal([embeddingsArray.length]), true);
+    // Compute attention scoresje sui
+    const attentionScores = tf.softmax(attentionWeights);
+    // Compute weighted sum of embeddings
+    const weightedEmbeddings = tf.mul(embeddingsTensor.transpose(), attentionScores).transpose();
+    const documentEmbeddingTensor = tf.sum(weightedEmbeddings, axis);
+    // Convert the result back to a JavaScript array
+    const documentEmbedding = documentEmbeddingTensor.arraySync();
+    const duration = (new Date().getTime() - start);
+    activity_1.log.info(`Computed attention embeddings in ${duration}ms - array size: ${documentEmbedding.length}`, { length: documentEmbedding.length });
+    // Clean up tensors
+    embeddingsTensor.dispose();
+    attentionWeights.dispose();
+    attentionScores.dispose();
+    weightedEmbeddings.dispose();
+    documentEmbeddingTensor.dispose();
+    return documentEmbedding;
+}
+//# sourceMappingURL=generateEmbeddings.js.map

package/lib/cjs/activities/generateEmbeddings.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"generateEmbeddings.js","sourceRoot":"","sources":["../../../src/activities/generateEmbeddings.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAsBA,gDAuEC;AA5FD,6CAAwJ;AAExJ,mDAA2C;AAC3C,0DAA4C;AAC5C,wEAAgE;AAChE,4CAA+C;AAC/C,gDAA2D;AAC3D,kDAAiD;AAc1C,KAAK,UAAU,kBAAkB,CAAC,OAAoC;IACzE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,YAAY,EAAE,GAAG,MAAM,IAAA,kCAAa,EAA2B,OAAO,CAAC,CAAC;IAC1G,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,GAAG,MAAM,CAAC;IAE/B,MAAM,WAAW,GAAG,MAAM,YAAY,EAAE,CAAC;IACzC,MAAM,MAAM,GAAG,WAAW,EAAE,aAAa,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC;IAC3D,IAAI,CAAC,WAAW,EAAE,CAAC;QACf,MAAM,IAAI,2BAAe,CAAC,mBAAmB,EAAE,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC;IACzE,CAAC;IACD,IAAI,CAAC,MAAM,EAAE,CAAC;QACV,MAAM,IAAI,2BAAe,CAAC,oCAAoC,EAAE,CAAC,QAAQ,CAAC,CAAC,CAAA;IAC/E,CAAC;IAED,IAAI,CAAC,WAAW,EAAE,CAAC;QACf,MAAM,IAAI,2BAAe,CAAC,mBAAmB,EAAE,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC;IACzE,CAAC;IAED,IAAI,CAAC,WAAW,EAAE,aAAa,CAAC,UAAU,CAAC,IAAI,CAAC,EAAE,OAAO,EAAE,CAAC;QACxD,cAAG,CAAC,IAAI,CAAC,2CAA2C,IAAI,gBAAgB,WAAW,CAAC,IAAI,KAAK,WAAW,CAAC,SAAS,GAAG,EAAE,EAAE,MAAM,EAAE,CAAC,CAAC;QACnI,OAAO,EAAE,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,2CAA2C,IAAI,EAAE,EAAE,CAAA;IAC1G,CAAC;IAED,cAAG,CAAC,IAAI,CAAC,GAAG,IAAI,6CAA6C,QAAQ,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC,CAAC;IAE5F,IAAI,CAAC,MAAM,CAAC,WAAW,EAAE,CAAC;QACtB,MAAM,IAAI,KAAK,CAAC,iHAAiH,CAAC,CAAC;IACvI,CAAC;IAED,MAAM,QAAQ,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC,QAAQ,EAAE,8CAA8C,CAAC,CAAC;IAEzG,IAAI,CAAC,QAAQ,EAAE,CAAC;QACZ,MAAM,IAAI,2BAAe,CAAC,oBAAoB,EAAE,CAAC,QAAQ,CAAC,CAAC,CAAC;IAChE,CAAC;IAED,IAAI,CAAC,QAAQ,CAAC,OAAO,EAAE,CAAC;QACpB,MAAM,IAAI,2BAAe,CAAC,4BAA4B,EAAE,CAAC,QAAQ,CAAC,CAAC,CAAC;IACxE,CAAC;IAED,IAAI,GAAG,CAAC;IAER,QAAQ,IAAI,EAAE,CAAC;QACX,KAAK,gCAAuB,CAAC,IAAI;YAC7B,GAAG,GAAG,MAAM,sBAAsB,CAAC;gBAC/B,MAAM;gBACN,MAAM;gBACN,QAAQ;gBACR,IAAI;aACP,CAAC,CAAA;YACF,MAAM;QACV,KAAK,gCAAuB,CAAC,UAAU;YACnC,GAAG,GAAG,MAAM,sBAAsB,CAAC;gBAC/B,MAAM;gBACN,MAAM;gBACN,QAAQ;gBACR,IAAI;aACP,CAAC,CAAC;YACH,MAAM;QACV,KAAK,gCAAuB,CAAC,KAAK;YAC9B,GAAG,GAAG,MAAM,uBAAuB,CAAC;gBAChC,MAAM;gBACN,MAAM;gBACN,QAAQ;gBACR,IAAI;aACP,CAAC,CAAC;YACH,MAAM;QACV;YACI,GAAG,GAAG,EAAE,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,QAAQ,EAAE,OAAO,EAAE,+BAA+B,IAAI,EAAE,EAAE,CAAA;IAChG,CAAC;IAED,OAAO,GAAG,CAAC;AAEf,CAAC;AAYD,KAAK,UAAU,sBAAsB,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAmC;IACrG,kGAAkG;IAClG,0FAA0F;IAC1F,IAAI;IAEJ,IAAI,CAAC,QAAQ,EAAE,CAAC;QACZ,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO,EAAE,+BAA+B,EAAE,CAAA;IACxE,CAAC;IAED,IAAI,IAAI,KAAK,gCAAuB,CAAC,IAAI,IAAI,IAAI,KAAK,gCAAuB,CAAC,UAAU,EAAE,CAAC;QACvF,OAAO,EAAE,EAAE,EAAE,QAAQ,CAAC,EAAE,EAAE,MAAM,EAAE,QAAQ,EAAE,OAAO,EAAE,+BAA+B,IAAI,EAAE,EAAE,CAAA;IAChG,CAAC;IAED,IAAI,IAAI,KAAK,gCAAuB,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC;QAC1D,OAAO,EAAE,EAAE,EAAE,QAAQ,CAAC,EAAE,EAAE,MAAM,EAAE,QAAQ,EAAE,OAAO,EAAE,eAAe,EAAE,CAAA;IAC1E,CAAC;IACD,IAAI,IAAI,KAAK,gCAAuB,CAAC,UAAU,IAAI,CAAC,QAAQ,EAAE,UAAU,EAAE,CAAC;QACvE,OAAO,EAAE,EAAE,EAAE,QAAQ,CAAC,EAAE,EAAE,MAAM,EAAE,QAAQ,EAAE,OAAO,EAAE,qBAAqB,EAAE,CAAA;IAChF,CAAC;IAED,MAAM,EAAE,WAAW,EAAE,KAAK,EAAE,GAAG,MAAM,CAAC;IAEtC,mCAAmC;IACnC,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,KAAK,IAAI,IAAI,KAAK,gCAAuB,CAAC,IAAI,EAAE,CAAC;QACnE,cAAG,CAAC,KAAK,CAAC,qCAAqC,GAAG,QAAQ,CAAC,EAAE,CAAC,CAAC;QAC/D,MAAM,UAAU,GAAG,IAAA,uBAAW,EAAC,QAAQ,CAAC,IAAK,CAAC,CAAC;QAC/C,MAAM,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,EAAE;YACrC,MAAM,EAAE;gBACJ,GAAG,UAAU;gBACb,IAAI,EAAE,QAAQ,CAAC,SAAS,IAAI,IAAA,cAAG,EAAC,QAAQ,CAAC,IAAK,CAAC;aAClD;SACJ,CAAC,CAAC;QACH,QAAQ,CAAC,MAAM,GAAG;YACd,GAAG,UAAU;YACb,IAAI,EAAE,QAAQ,CAAC,SAAS,IAAI,IAAA,cAAG,EAAC,QAAQ,CAAC,IAAK,CAAC;SAClD,CAAC;IACN,CAAC;IAED,MAAM,SAAS,GAAG,MAAM,CAAC,UAAU,IAAI,IAAI,CAAC;IAE5C,kEAAkE;IAClE,4DAA4D;IAC5D,gFAAgF;IAChF,cAAG,CAAC,IAAI,CAAC,cAAc,IAAI,4BAA4B,QAAQ,CAAC,EAAE,EAAE,CAAC,CAAC;IACtE,IAAI,IAAI,KAAK,gCAAuB,CAAC,IAAI,IAAI,QAAQ,CAAC,MAAM,EAAE,KAAK,IAAI,QAAQ,CAAC,MAAM,EAAE,KAAK,GAAG,SAAS,EAAE,CAAC;QACxG,cAAG,CAAC,IAAI,CAAC,qDAAqD,CAAC,CAAC;QAEhE,IAAI,CAAC,QAAQ,CAAC,KAAK,IAAI,QAAQ,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACjD,OAAO,EAAE,EAAE,EAAE,QAAQ,CAAC,EAAE,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,gBAAgB,EAAE,CAAA;QAC5E,CAAC;QAED,MAAM,QAAQ,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,EAAE,GAAG,CAAC,KAAK,EAAE,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC,MAAM,EAAE,uCAAuC,CAAC,CAAC,CAAC,CAAC;QACpJ,cAAG,CAAC,IAAI,CAAC,aAAa,QAAQ,CAAC,MAAM,QAAQ,CAAC,CAAA;QAE9C,MAAM,sBAAsB,GAAG,KAAK,EAAE,IAAwB,EAAE,CAAS,EAAE,EAAE;YACzE,IAAI,CAAC;gBACD,cAAG,CAAC,IAAI,CAAC,kCAAkC,IAAI,CAAC,EAAE,EAAE,EAAE,EAAE,QAAQ,EAAE,IAAI,CAAC,IAAI,EAAE,MAAM,EAAE,CAAC,CAAA;gBACtF,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;oBACb,OAAO,EAAE,EAAE,EAAE,IAAI,CAAC,EAAE,EAAE,MAAM,EAAE,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,eAAe,EAAE,CAAA;gBAChG,CAAC;gBAED,IAAI,IAAI,CAAC,MAAM,EAAE,KAAK,IAAI,IAAI,CAAC,MAAM,CAAC,KAAK,GAAG,SAAS,EAAE,CAAC;oBACtD,cAAG,CAAC,IAAI,CAAC,yDAAyD,EAAE,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,EAAE,MAAM,EAAE,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC,CAAC;oBAClH,OAAO,EAAE,EAAE,EAAE,IAAI,CAAC,EAAE,EAAE,MAAM,EAAE,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,gBAAgB,EAAE,CAAA;gBAC9E,CAAC;gBAED,MAAM,CAAC,GAAG,MAAM,4BAA4B,CAAC,IAAI,CAAC,IAAI,EAAE,WAAW,EAAE,MAAM,EAAE,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE;oBAC1F,cAAG,CAAC,KAAK,CAAC,sCAAsC,EAAE,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,EAAE,MAAM,EAAE,IAAI,CAAC,MAAM,EAAE,WAAW,EAAE,IAAI,CAAC,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC,CAAC;oBACpI,OAAO,IAAI,CAAC;gBAChB,CAAC,CAAC,CAAC;gBAEH,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC;oBAClB,OAAO,EAAE,EAAE,EAAE,IAAI,CAAC,EAAE,EAAE,MAAM,EAAE,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,yBAAyB,EAAE,CAAA;gBACvF,CAAC;gBAED,cAAG,CAAC,IAAI,CAAC,iCAAiC,IAAI,CAAC,EAAE,iCAAiC,CAAC,CAAA;gBACnF,MAAM,MAAM,CAAC,OAAO,CAAC,YAAY,CAAC,IAAI,CAAC,EAAE,EAAE,gCAAuB,CAAC,IAAI,EACnE;oBACI,MAAM,EAAE,CAAC,CAAC,MAAM;oBAChB,KAAK,EAAE,CAAC,CAAC,KAAK;oBACd,IAAI,EAAE,IAAI,CAAC,SAAS;iBACvB,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE;oBACX,cAAG,CAAC,IAAI,CAAC,qCAAqC,IAAI,CAAC,EAAE,EAAE,CAAC,CAAC;oBACzD,OAAO,EAAE,EAAE,EAAE,IAAI,CAAC,EAAE,EAAE,MAAM,EAAE,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,kCAAkC,EAAE,KAAK,EAAE,GAAG,CAAC,OAAO,EAAE,CAAA;gBACpH,CAAC,CAAC,CAAA;gBAEN,cAAG,CAAC,IAAI,CAAC,iCAAiC,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC;gBACtD,OAAO,EAAE,EAAE,EAAE,IAAI,CAAC,EAAE,EAAE,MAAM,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,CAAA;YAChD,CAAC;YAAC,OAAO,GAAQ,EAAE,CAAC;gBAChB,cAAG,CAAC,IAAI,CAAC,oBAAoB,IAAI,wBAAwB,IAAI,CAAC,EAAE,OAAO,QAAQ,CAAC,EAAE,EAAE,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC;gBACtG,OAAO,EAAE,EAAE,EAAE,IAAI,CAAC,EAAE,EAAE,MAAM,EAAE,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,6BAA6B,EAAE,KAAK,EAAE,GAAG,CAAC,OAAO,EAAE,CAAA;YAC/G,CAAC;QACL,CAAC,CAAA;QAED,MAAM,QAAQ,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,sBAAsB,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAA;QACrE,MAAM,GAAG,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACxC,aAAa;QACb,8BAA8B;QAC9B,0CAA0C;QAC1C,sDAAsD;QACtD,kBAAkB;QAClB,IAAI;QAGJ,sCAAsC;QACtC,MAAM,eAAe,GAAG,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,KAAK,IAAI,CAA+D,CAAC;QAE/H,gFAAgF;QAChF,cAAG,CAAC,IAAI,CAAC,6CAA6C,CAAC,CAAC;QACxD,MAAM,iBAAiB,GAAG,yBAAyB,CAAC,eAAe,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC;QAErG,oCAAoC;QACpC,MAAM,MAAM,CAAC,OAAO,CAAC,YAAY,CAAC,QAAQ,CAAC,EAAE,EAAE,IAAI,EAC/C;YACI,MAAM,EAAE,iBAAiB;YACzB,KAAK,EAAE,WAAW;YAClB,IAAI,EAAE,QAAQ,CAAC,SAAS;SAC3B,CACJ,CAAC;QACF,OAAO,EAAE,EAAE,EAAE,QAAQ,CAAC,EAAE,EAAE,MAAM,EAAE,WAAW,EAAE,KAAK,EAAE,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,iBAAiB,CAAC,MAAM,EAAE,eAAe,EAAE,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,GAAG,OAAO,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC,KAAK,EAAE,OAAO,EAAE,CAAC,CAAC,OAAO,EAAE,CAAA,CAAC,CAAC,CAAC,EAAE,CAAA;IAEhO,CAAC;SAAM,CAAC;QACJ,cAAG,CAAC,IAAI,CAAC,cAAc,IAAI,0BAA0B,CAAC,CAAC;QAEvD,MAAM,GAAG,GAAG,MAAM,4BAA4B,CAAC,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,EAAE,WAAW,EAAE,MAAM,CAAC,CAAC;QACpG,IAAI,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC;YACtB,OAAO,EAAE,EAAE,EAAE,QAAQ,CAAC,EAAE,EAAE,MAAM,EAAE,QAAQ,EAAE,OAAO,EAAE,yBAAyB,EAAE,CAAA;QACpF,CAAC;QAED,cAAG,CAAC,IAAI,CAAC,GAAG,IAAI,sCAAsC,QAAQ,CAAC,EAAE,EAAE,EAAE,EAAE,GAAG,EAAE,GAAG,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,CAAC;QACjG,MAAM,MAAM,CAAC,OAAO,CAAC,YAAY,CAAC,QAAQ,CAAC,EAAE,EAAE,IAAI,EAC/C;YACI,MAAM,EAAE,GAAG,CAAC,MAAM;YAClB,KAAK,EAAE,GAAG,CAAC,KAAK;YAChB,IAAI,EAAE,QAAQ,CAAC,SAAS;SAC3B,CACJ,CAAC;QAEF,OAAO,EAAE,EAAE,EAAE,QAAQ,CAAC,EAAE,EAAE,IAAI,EAAE,MAAM,EAAE,WAAW,EAAE,GAAG,EAAE,GAAG,CAAC,MAAM,CAAC,MAAM,EAAE,CAAA;IAEjF,CAAC;AAEL,CAAC;AAED,KAAK,UAAU,uBAAuB,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAmC;IAEtG,cAAG,CAAC,IAAI,CAAC,2CAA2C,GAAG,QAAQ,CAAC,EAAE,EAAE,EAAE,OAAO,EAAE,QAAQ,CAAC,OAAO,EAAE,CAAC,CAAC;IACnG,IAAI,CAAC,QAAQ,CAAC,OAAO,EAAE,IAAI,EAAE,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,EAAE,IAAI,EAAE,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;QAC5F,OAAO,EAAE,EAAE,EAAE,QAAQ,CAAC,EAAE,EAAE,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,OAAO,EAAE,yBAAyB,EAAE,CAAA;IAC1F,CAAC;IACD,MAAM,EAAE,WAAW,EAAE,KAAK,EAAE,GAAG,MAAM,CAAA;IAErC,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,KAAK,CAAC,OAAO,CAAC,YAAY,CAAC,QAAQ,CAAC,EAAE,EAAE;QAChE,MAAM,EAAE,WAAW;QACnB,MAAM,EAAE,IAAI;QACZ,mBAAmB,EAAE,IAAI;KAC5B,CAAC,CAAC;IAEH,IAAI,MAAM,CAAC,MAAM,KAAK,YAAY,EAAE,CAAC;QACjC,MAAM,IAAI,KAAK,CAAC,2CAA2C,CAAC,CAAA;IAChE,CAAC;SAAM,IAAI,MAAM,CAAC,MAAM,KAAK,QAAQ,IAAI,CAAC,MAAM,CAAC,SAAS,EAAE,CAAC;QACzD,MAAM,IAAI,2BAAe,CAAC,4BAA4B,EAAE,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAA;IAC1E,CAAC;IAED,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC;QACnC,MAAM,IAAI,2BAAe,CAAC,8BAA8B,EAAE,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAA;IAC5E,CAAC;IAED,MAAM,KAAK,GAAG,MAAM,IAAA,4BAAiB,EAAC,MAAM,EAAE,MAAM,CAAC,SAAS,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;IAE/E,MAAM,GAAG,GAAG,MAAM,MAAM,CAAC,YAAY,CAAC,UAAU,CAAC,WAAW,EAAE;QAC1D,KAAK;QACL,KAAK;KACR,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE;QAC1B,cAAG,CAAC,KAAK,CAAC,uCAAuC,EAAE,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC,CAAA;QAChE,MAAM,CAAC,CAAC;IACZ,CAAC,CAAC,CAAC;IAEH,IAAI,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC;QACtB,OAAO,EAAE,EAAE,EAAE,QAAQ,CAAC,EAAE,EAAE,MAAM,EAAE,QAAQ,EAAE,OAAO,EAAE,yBAAyB,EAAE,CAAA;IACpF,CAAC;IAED,MAAM,MAAM,CAAC,OAAO,CAAC,YAAY,CAAC,QAAQ,CAAC,EAAE,EAAE,gCAAuB,CAAC,KAAK,EACxE;QACI,MAAM,EAAE,GAAG,CAAC,MAAM;QAClB,KAAK,EAAE,GAAG,CAAC,KAAK;QAChB,IAAI,EAAE,QAAQ,CAAC,SAAS;KAC3B,CACJ,CAAC;IAEF,OAAO,EAAE,EAAE,EAAE,QAAQ,CAAC,EAAE,EAAE,IAAI,EAAE,MAAM,EAAE,WAAW,EAAE,GAAG,EAAE,GAAG,CAAC,MAAM,CAAC,MAAM,EAAE,CAAA;AAEjF,CAAC;AAED,KAAK,UAAU,4BAA4B,CAAC,IAAY,EAAE,GAAW,EAAE,MAAwB,EAAE,KAAc;IAE3G,cAAG,CAAC,IAAI,CAAC,qCAAqC,IAAI,CAAC,MAAM,2BAA2B,GAAG,EAAE,CAAC,CAAC;IAE3F,OAAO,MAAM,CAAC,YAAY,CAAC,UAAU,CAAC,GAAG,EAAE;QACvC,IAAI;QACJ,KAAK;KACR,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE;QAC1B,cAAG,CAAC,KAAK,CAAC,sCAAsC,EAAE,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC,CAAA;QAC/D,MAAM,CAAC,CAAC;IACZ,CAAC,CAAC,CAAC;AAEP,CAAC;AAED,SAAS,yBAAyB,CAAC,eAA2B,EAAE,OAAe,CAAC;IAC5E,IAAI,eAAe,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAC5C,cAAG,CAAC,IAAI,CAAC,mCAAmC,EAAE,EAAE,gBAAgB,EAAE,eAAe,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;IACxG,MAAM,KAAK,GAAG,IAAI,IAAI,EAAE,CAAC,OAAO,EAAE,CAAC;IAEnC,gDAAgD;IAChD,MAAM,gBAAgB,GAAG,EAAE,CAAC,MAAM,CAAC,eAAe,CAAC,CAAC;IAEpD,yCAAyC;IACzC,MAAM,gBAAgB,GAAG,EAAE,CAAC,QAAQ,CAAC,EAAE,CAAC,YAAY,CAAC,CAAC,eAAe,CAAC,MAAM,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;IAEtF,iCAAiC;IACjC,MAAM,eAAe,GAAG,EAAE,CAAC,OAAO,CAAC,gBAAgB,CAAC,CAAC;IAErD,qCAAqC;IACrC,MAAM,kBAAkB,GAAG,EAAE,CAAC,GAAG,CAAC,gBAAgB,CAAC,SAAS,EAAE,EAAE,eAAe,CAAC,CAAC,SAAS,EAAE,CAAC;IAC7F,MAAM,uBAAuB,GAAG,EAAE,CAAC,GAAG,CAAC,kBAAkB,EAAE,IAAI,CAAC,CAAC;IAEjE,gDAAgD;IAChD,MAAM,iBAAiB,GAAG,uBAAuB,CAAC,SAAS,EAAc,CAAC;IAC1E,MAAM,QAAQ,GAAG,CAAC,IAAI,IAAI,EAAE,CAAC,OAAO,EAAE,GAAG,KAAK,CAAC,CAAC;IAChD,cAAG,CAAC,IAAI,CAAC,oCAAoC,QAAQ,oBAAoB,iBAAiB,CAAC,MAAM,EAAE,EAAE,EAAE,MAAM,EAAE,iBAAiB,CAAC,MAAM,EAAE,CAAC,CAAC;IAE3I,mBAAmB;IACnB,gBAAgB,CAAC,OAAO,EAAE,CAAC;IAC3B,gBAAgB,CAAC,OAAO,EAAE,CAAC;IAC3B,eAAe,CAAC,OAAO,EAAE,CAAC;IAC1B,kBAAkB,CAAC,OAAO,EAAE,CAAC;IAC7B,uBAAuB,CAAC,OAAO,EAAE,CAAC;IAElC,OAAO,iBAAiB,CAAC;AAC7B,CAAC"}

package/lib/cjs/activities/generateImageRendition.js

@@ -0,0 +1,104 @@
+"use strict";
+var __importDefault = (this && this.__importDefault) || function (mod) {
+    return (mod && mod.__esModule) ? mod : { "default": mod };
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.generateImageRendition = generateImageRendition;
+const activity_1 = require("@temporalio/activity");
+const fs_1 = __importDefault(require("fs"));
+const sharp_1 = __importDefault(require("sharp"));
+const image_js_1 = require("../conversion/image.js");
+const mutool_js_1 = require("../conversion/mutool.js");
+const ActivityContext_js_1 = require("../dsl/setup/ActivityContext.js");
+const errors_js_1 = require("../errors.js");
+const blobs_js_1 = require("../utils/blobs.js");
+const memory_js_1 = require("../utils/memory.js");
+async function generateImageRendition(payload) {
+    const { client, objectId, params } = await (0, ActivityContext_js_1.setupActivity)(payload);
+    const supportedNonImageInputTypes = ['application/pdf'];
+    const inputObject = await client.objects.retrieve(objectId).catch((err) => {
+        activity_1.log.error(`Failed to retrieve document ${objectId}`, err);
+        if (err.response?.status === 404) {
+            throw new errors_js_1.NoDocumentFound(`Document ${objectId} not found`, [objectId]);
+        }
+        throw err;
+    });
+    const renditionType = await client.types.getTypeByName('Rendition');
+    if (!params.format) {
+        activity_1.log.error(`Format not found`);
+        throw new errors_js_1.WorkflowParamNotFound(`format`);
+    }
+    if (!renditionType) {
+        activity_1.log.error(`Rendition type not found`);
+        throw new errors_js_1.NoDocumentFound(`Rendition type not found`, [objectId]);
+    }
+    if (!inputObject.content?.source) {
+        activity_1.log.error(`Document ${objectId} has no source`);
+        throw new errors_js_1.NoDocumentFound(`Document ${objectId} has no source`, [objectId]);
+    }
+    if (!inputObject.content.type || (!inputObject.content.type?.startsWith('image/') && !supportedNonImageInputTypes.includes(inputObject.content.type))) {
+        activity_1.log.error(`Document ${objectId} is not an image`);
+        throw new errors_js_1.NoDocumentFound(`Document ${objectId} is not an image or pdf: ${inputObject.content.type}`, [objectId]);
+    }
+    //array of rendition files to upload
+    let renditionPages = [];
+    //if PDF, convert to pages
+    if (inputObject.content.type === 'application/pdf') {
+        const pdfBuffer = await (0, blobs_js_1.fetchBlobAsBuffer)(client, inputObject.content.source);
+        const pages = await (0, mutool_js_1.pdfToImages)(pdfBuffer);
+        if (!pages.length) {
+            activity_1.log.error(`Failed to convert pdf to image`);
+            throw new Error(`Failed to convert pdf to image`);
+        }
+        renditionPages = [...pages];
+    }
+    else if (inputObject.content.type.startsWith('image/')) {
+        const tmpFile = await (0, blobs_js_1.saveBlobToTempFile)(client, inputObject.content.source);
+        const filestats = fs_1.default.statSync(tmpFile);
+        activity_1.log.info(`Image ${objectId} copied to ${tmpFile}`, { filestats });
+        renditionPages.push(tmpFile);
+    }
+    //generate rendition name, pass an index for multi parts
+    const getRenditionName = (index = 0) => {
+        const name = `renditions/${objectId}/${params.max_hw}/${index}.${params.format}`;
+        return name;
+    };
+    if (!renditionPages || !renditionPages.length) {
+        activity_1.log.error(`Failed to generate rendition for ${objectId}`);
+        throw new Error(`Failed to generate rendition for ${objectId}`);
+    }
+    activity_1.log.info(`Uploading rendition for ${objectId} with ${renditionPages.length} pages (max_hw: ${params.max_hw}, format: ${params.format})`, { renditionPages });
+    const uploads = renditionPages.map(async (page, i) => {
+        const pageId = getRenditionName(i);
+        const resized = (0, sharp_1.default)(page).pipe((0, image_js_1.imageResizer)(params.max_hw, params.format));
+        const source = new memory_js_1.NodeStreamSource(resized, pageId.replace('renditions/', '').replace('/', '_'), 'image/' + params.format, pageId);
+        activity_1.log.info(`Uploading rendition for ${objectId} page ${i} with max_hw: ${params.max_hw} and format: ${params.format}`);
+        return client.objects.upload(source).catch((err) => {
+            activity_1.log.error(`Failed to upload rendition for ${objectId} page ${i}`, err);
+            return Promise.resolve(null);
+        });
+    });
+    const uploaded = await Promise.all(uploads);
+    if (!uploaded || !uploaded.length || !uploaded[0]) {
+        activity_1.log.error(`Failed to upload rendition for ${objectId}`);
+        throw new Error(`Failed to upload rendition for ${objectId}`);
+    }
+    activity_1.log.info(`Creating rendition for ${objectId} with max_hw: ${params.max_hw} and format: ${params.format}`, { uploaded });
+    const rendition = await client.objects.create({
+        name: inputObject.name + ` [Rendition ${params.max_hw}]`,
+        type: renditionType.id,
+        parent: inputObject.id,
+        content: uploaded[0],
+        properties: {
+            mime_type: 'image/' + params.format,
+            source_etag: inputObject.content.source,
+            height: params.max_hw,
+            width: params.max_hw,
+            multipart: uploaded.length > 1,
+            total_parts: uploaded.length
+        }
+    });
+    activity_1.log.info(`Rendition ${rendition.id} created for ${objectId}`, { rendition });
+    return { id: rendition.id, format: params.format, status: "success" };
+}
+//# sourceMappingURL=generateImageRendition.js.map

package/lib/cjs/activities/generateImageRendition.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"generateImageRendition.js","sourceRoot":"","sources":["../../../src/activities/generateImageRendition.ts"],"names":[],"mappings":";;;;;AAwBA,wDA6GC;AApID,mDAA2C;AAC3C,4CAAoB;AACpB,kDAA0C;AAC1C,qDAAsD;AACtD,uDAAsD;AACtD,wEAAgE;AAChE,4CAAsE;AACtE,gDAA0E;AAC1E,kDAAsD;AAe/C,KAAK,UAAU,sBAAsB,CAAC,OAAoC;IAC7E,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,EAAE,GAAG,MAAM,IAAA,kCAAa,EAA+B,OAAO,CAAC,CAAC;IAEhG,MAAM,2BAA2B,GAAG,CAAC,iBAAiB,CAAC,CAAA;IACvD,MAAM,WAAW,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,KAAK,CAAC,CAAC,GAAG,EAAE,EAAE;QACtE,cAAG,CAAC,KAAK,CAAC,+BAA+B,QAAQ,EAAE,EAAE,GAAG,CAAC,CAAC;QAC1D,IAAI,GAAG,CAAC,QAAQ,EAAE,MAAM,KAAK,GAAG,EAAE,CAAC;YAC/B,MAAM,IAAI,2BAAe,CAAC,YAAY,QAAQ,YAAY,EAAE,CAAC,QAAQ,CAAC,CAAC,CAAC;QAC5E,CAAC;QACD,MAAM,GAAG,CAAC;IACd,CAAC,CAAC,CAAC;IACH,MAAM,aAAa,GAAG,MAAM,MAAM,CAAC,KAAK,CAAC,aAAa,CAAC,WAAW,CAAC,CAAC;IAEpE,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC;QACjB,cAAG,CAAC,KAAK,CAAC,kBAAkB,CAAC,CAAC;QAC9B,MAAM,IAAI,iCAAqB,CAAC,QAAQ,CAAC,CAAC;IAC9C,CAAC;IAED,IAAI,CAAC,aAAa,EAAE,CAAC;QACjB,cAAG,CAAC,KAAK,CAAC,0BAA0B,CAAC,CAAC;QACtC,MAAM,IAAI,2BAAe,CAAC,0BAA0B,EAAE,CAAC,QAAQ,CAAC,CAAC,CAAC;IACtE,CAAC;IAED,IAAI,CAAC,WAAW,CAAC,OAAO,EAAE,MAAM,EAAE,CAAC;QAC/B,cAAG,CAAC,KAAK,CAAC,YAAY,QAAQ,gBAAgB,CAAC,CAAC;QAChD,MAAM,IAAI,2BAAe,CAAC,YAAY,QAAQ,gBAAgB,EAAE,CAAC,QAAQ,CAAC,CAAC,CAAC;IAChF,CAAC;IAED,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,IAAI,IAAI,CAAC,CAAC,WAAW,CAAC,OAAO,CAAC,IAAI,EAAE,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,2BAA2B,CAAC,QAAQ,CAAC,WAAW,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;QACpJ,cAAG,CAAC,KAAK,CAAC,YAAY,QAAQ,kBAAkB,CAAC,CAAC;QAClD,MAAM,IAAI,2BAAe,CAAC,YAAY,QAAQ,4BAA4B,WAAW,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,QAAQ,CAAC,CAAC,CAAC;IACtH,CAAC;IAED,oCAAoC;IACpC,IAAI,cAAc,GAAa,EAAE,CAAC;IAElC,0BAA0B;IAC1B,IAAI,WAAW,CAAC,OAAO,CAAC,IAAI,KAAK,iBAAiB,EAAE,CAAC;QACjD,MAAM,SAAS,GAAG,MAAM,IAAA,4BAAiB,EAAC,MAAM,EAAE,WAAW,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;QAC9E,MAAM,KAAK,GAAG,MAAM,IAAA,uBAAW,EAAC,SAAS,CAAC,CAAC;QAC3C,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC;YAChB,cAAG,CAAC,KAAK,CAAC,gCAAgC,CAAC,CAAC;YAC5C,MAAM,IAAI,KAAK,CAAC,gCAAgC,CAAC,CAAC;QACtD,CAAC;QACD,cAAc,GAAG,CAAC,GAAG,KAAK,CAAC,CAAC;IAChC,CAAC;SAAM,IAAI,WAAW,CAAC,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;QACvD,MAAM,OAAO,GAAG,MAAM,IAAA,6BAAkB,EAAC,MAAM,EAAE,WAAW,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;QAC7E,MAAM,SAAS,GAAG,YAAE,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;QACvC,cAAG,CAAC,IAAI,CAAC,SAAS,QAAQ,cAAc,OAAO,EAAE,EAAE,EAAE,SAAS,EAAE,CAAC,CAAC;QAClE,cAAc,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACjC,CAAC;IAED,wDAAwD;IACxD,MAAM,gBAAgB,GAAG,CAAC,QAAgB,CAAC,EAAE,EAAE;QAC3C,MAAM,IAAI,GAAG,cAAc,QAAQ,IAAI,MAAM,CAAC,MAAM,IAAI,KAAK,IAAI,MAAM,CAAC,MAAM,EAAE,CAAC;QACjF,OAAO,IAAI,CAAC;IAChB,CAAC,CAAA;IAED,IAAI,CAAC,cAAc,IAAI,CAAC,cAAc,CAAC,MAAM,EAAE,CAAC;QAC5C,cAAG,CAAC,KAAK,CAAC,oCAAoC,QAAQ,EAAE,CAAC,CAAC;QAC1D,MAAM,IAAI,KAAK,CAAC,oCAAoC,QAAQ,EAAE,CAAC,CAAC;IACpE,CAAC;IAED,cAAG,CAAC,IAAI,CAAC,2BAA2B,QAAQ,SAAS,cAAc,CAAC,MAAM,mBAAmB,MAAM,CAAC,MAAM,aAAa,MAAM,CAAC,MAAM,GAAG,EAAE,EAAE,cAAc,EAAE,CAAC,CAAC;IAC7J,MAAM,OAAO,GAAG,cAAc,CAAC,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,CAAC,EAAE,EAAE;QACjD,MAAM,MAAM,GAAG,gBAAgB,CAAC,CAAC,CAAC,CAAC;QACnC,MAAM,OAAO,GAAG,IAAA,eAAK,EAAC,IAAI,CAAC,CAAC,IAAI,CAAC,IAAA,uBAAY,EAAC,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC;QAE7E,MAAM,MAAM,GAAG,IAAI,4BAAgB,CAC/B,OAAO,EACP,MAAM,CAAC,OAAO,CAAC,aAAa,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,GAAG,EAAE,GAAG,CAAC,EACnD,QAAQ,GAAG,MAAM,CAAC,MAAM,EACxB,MAAM,CACT,CAAA;QAED,cAAG,CAAC,IAAI,CAAC,2BAA2B,QAAQ,SAAS,CAAC,iBAAiB,MAAM,CAAC,MAAM,gBAAgB,MAAM,CAAC,MAAM,EAAE,CAAC,CAAC;QACrH,OAAO,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,CAAC,CAAC,GAAG,EAAE,EAAE;YAC/C,cAAG,CAAC,KAAK,CAAC,kCAAkC,QAAQ,SAAS,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC;YACvE,OAAO,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;QACjC,CAAC,CAAC,CAAC;IACP,CAAC,CAAC,CAAC;IAEH,MAAM,QAAQ,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;IAC5C,IAAI,CAAC,QAAQ,IAAI,CAAC,QAAQ,CAAC,MAAM,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC;QAChD,cAAG,CAAC,KAAK,CAAC,kCAAkC,QAAQ,EAAE,CAAC,CAAC;QACxD,MAAM,IAAI,KAAK,CAAC,kCAAkC,QAAQ,EAAE,CAAC,CAAC;IAClE,CAAC;IAGD,cAAG,CAAC,IAAI,CAAC,0BAA0B,QAAQ,iBAAiB,MAAM,CAAC,MAAM,gBAAgB,MAAM,CAAC,MAAM,EAAE,EAAE,EAAE,QAAQ,EAAE,CAAC,CAAC;IACxH,MAAM,SAAS,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC;QAC1C,IAAI,EAAE,WAAW,CAAC,IAAI,GAAG,eAAe,MAAM,CAAC,MAAM,GAAG;QACxD,IAAI,EAAE,aAAa,CAAC,EAAE;QACtB,MAAM,EAAE,WAAW,CAAC,EAAE;QACtB,OAAO,EAAE,QAAQ,CAAC,CAAC,CAAC;QACpB,UAAU,EAAE;YACR,SAAS,EAAE,QAAQ,GAAG,MAAM,CAAC,MAAM;YACnC,WAAW,EAAE,WAAW,CAAC,OAAO,CAAC,MAAM;YACvC,MAAM,EAAE,MAAM,CAAC,MAAM;YACrB,KAAK,EAAE,MAAM,CAAC,MAAM;YACpB,SAAS,EAAE,QAAQ,CAAC,MAAM,GAAG,CAAC;YAC9B,WAAW,EAAE,QAAQ,CAAC,MAAM;SACD;KAClC,CAAC,CAAC;IAEH,cAAG,CAAC,IAAI,CAAC,aAAa,SAAS,CAAC,EAAE,gBAAgB,QAAQ,EAAE,EAAE,EAAE,SAAS,EAAE,CAAC,CAAC;IAE7E,OAAO,EAAE,EAAE,EAAE,SAAS,CAAC,EAAE,EAAE,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,EAAE,SAAS,EAAE,CAAC;AAE1E,CAAC"}

package/lib/cjs/activities/generateOrAssignContentType.js

@@ -0,0 +1,103 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.generateOrAssignContentType = generateOrAssignContentType;
+const activity_1 = require("@temporalio/activity");
+const ActivityContext_js_1 = require("../dsl/setup/ActivityContext.js");
+const tokens_js_1 = require("../utils/tokens.js");
+const executeInteraction_js_1 = require("./executeInteraction.js");
+const INT_SELECT_DOCUMENT_TYPE = "sys:SelectDocumentType";
+const INT_GENERATE_METADATA_MODEL = "sys:GenerateMetadataModel";
+async function generateOrAssignContentType(payload) {
+    const context = await (0, ActivityContext_js_1.setupActivity)(payload);
+    const { params, client, objectId } = context;
+    const interactionName = params.interactionNames?.selectDocumentType ?? INT_SELECT_DOCUMENT_TYPE;
+    activity_1.log.info("SelectDocumentType for object: " + objectId, { payload });
+    const object = await client.objects.retrieve(objectId, "+text");
+    //Expects object.type to be null on first ingestion of content
+    //User initiated Content Type change via the Composable UI,
+    //sets object.type to null when they let Composable choose for them.
+    //sets object.type to chosen type (thus non-null) when user picks a type.
+    if (object.type) {
+        activity_1.log.warn(`Object ${objectId} has already a type. Skipping type creation.`);
+        return { status: "skipped", message: "Object already has a type: " + object.type.name };
+    }
+    if (!object || (!object.text && !object.content?.type?.startsWith("image/") && !object.content?.type?.startsWith("application/pdf"))) {
+        activity_1.log.info(`Object ${objectId} not found or text is empty and not an image`, { object });
+        return { status: "failed", error: "no-text" };
+    }
+    const types = await client.types.list();
+    //make a list of all existing types, and add hints if any
+    const existing_types = types.map(t => t.name).filter(n => !["DocumentPart", "Rendition"].includes(n));
+    if (params.typesHint) {
+        const newHints = params.typesHint.filter((t) => !existing_types.includes(t));
+        existing_types.push(...newHints);
+    }
+    const content = object.text ? (0, tokens_js_1.truncByMaxTokens)(object.text, params.truncate || 4000) : undefined;
+    const getImage = async () => {
+        if (object.content?.type?.includes("pdf") && object.text?.length && object.text?.length < 100) {
+            return "store:" + objectId;
+        }
+        if (!object.content?.type?.startsWith("image/")) {
+            return undefined;
+        }
+        const res = await client.objects.getRendition(objectId, { max_hw: 1024, format: "image/png", generate_if_missing: true });
+        if (!res.rendition && res.status === "generating") {
+            //throw to try again
+            throw new Error(`Rendition for object ${objectId} is in progress`);
+        }
+        else if (res.rendition) {
+            return "store:" + objectId;
+        }
+    };
+    const fileRef = await getImage();
+    activity_1.log.info("Execute SelectDocumentType interaction on content with \nexisting types: " + existing_types.join(","));
+    const res = await (0, executeInteraction_js_1.executeInteractionFromActivity)(client, interactionName, params, {
+        existing_types, content, image: fileRef
+    });
+    activity_1.log.info("Selected Content Type Result: " + JSON.stringify(res.result));
+    //if type is not identified or not present in the database, generate a new type
+    let selectedType = undefined;
+    selectedType = types.find(t => t.name === res.result.document_type);
+    if (!selectedType) {
+        activity_1.log.warn("Document type not idenfified: starting type generation");
+        const newType = await generateNewType(context, existing_types, content, fileRef);
+        selectedType = { id: newType.id, name: newType.name };
+    }
+    if (!selectedType) {
+        activity_1.log.error("Type not found: ", res.result);
+        throw new Error("Type not found: " + res.result.document_type);
+    }
+    //update object with selected type
+    await client.objects.update(objectId, {
+        type: selectedType.id,
+    });
+    return {
+        id: selectedType.id,
+        name: selectedType.name,
+        isNew: !types.find(t => t.name === selectedType.name)
+    };
+}
+async function generateNewType(context, existing_types, content, fileRef) {
+    const { client, params } = context;
+    const project = await context.fetchProject();
+    const interactionName = params.interactionNames?.generateMetadataModel ?? INT_GENERATE_METADATA_MODEL;
+    const genTypeRes = await (0, executeInteraction_js_1.executeInteractionFromActivity)(client, interactionName, params, {
+        existing_types: existing_types,
+        content: content,
+        human_context: project?.configuration?.human_context ?? undefined,
+        image: fileRef ? fileRef : undefined
+    });
+    if (!genTypeRes.result.document_type) {
+        activity_1.log.error("No name generated for type", genTypeRes);
+        throw new Error("No name generated for type");
+    }
+    activity_1.log.info("Generated schema for type", genTypeRes.result.metadata_schema);
+    const typeData = {
+        name: genTypeRes.result.document_type,
+        object_schema: genTypeRes.result.metadata_schema,
+        is_chunkable: genTypeRes.result.is_chunkable,
+    };
+    const type = await client.types.create(typeData);
+    return type;
+}
+//# sourceMappingURL=generateOrAssignContentType.js.map

package/lib/cjs/activities/generateOrAssignContentType.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"generateOrAssignContentType.js","sourceRoot":"","sources":["../../../src/activities/generateOrAssignContentType.ts"],"names":[],"mappings":";;AA8BA,kEAyFC;AAtHD,mDAA2C;AAC3C,wEAAiF;AACjF,kDAAoE;AACpE,mEAAqG;AAErG,MAAM,wBAAwB,GAAG,wBAAwB,CAAA;AACzD,MAAM,2BAA2B,GAAG,2BAA2B,CAAA;AAuBxD,KAAK,UAAU,2BAA2B,CAAC,OAAoC;IAClF,MAAM,OAAO,GAAG,MAAM,IAAA,kCAAa,EAAoC,OAAO,CAAC,CAAC;IAChF,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,GAAG,OAAO,CAAC;IAE7C,MAAM,eAAe,GAAG,MAAM,CAAC,gBAAgB,EAAE,kBAAkB,IAAI,wBAAwB,CAAC;IAGhG,cAAG,CAAC,IAAI,CAAC,iCAAiC,GAAG,QAAQ,EAAE,EAAE,OAAO,EAAE,CAAC,CAAC;IAEpE,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;IAEhE,8DAA8D;IAC9D,2DAA2D;IAC3D,oEAAoE;IACpE,yEAAyE;IACzE,IAAI,MAAM,CAAC,IAAI,EAAE,CAAC;QACd,cAAG,CAAC,IAAI,CAAC,UAAU,QAAQ,8CAA8C,CAAC,CAAC;QAC3E,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,6BAA6B,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;IAC5F,CAAC;IAED,IAAI,CAAC,MAAM,IAAI,CAAC,CAAC,MAAM,CAAC,IAAI,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,IAAI,EAAE,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,IAAI,EAAE,UAAU,CAAC,iBAAiB,CAAC,CAAC,EAAE,CAAC;QACnI,cAAG,CAAC,IAAI,CAAC,UAAU,QAAQ,8CAA8C,EAAE,EAAE,MAAM,EAAE,CAAC,CAAC;QACvF,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC;IAClD,CAAC;IAED,MAAM,KAAK,GAAG,MAAM,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;IAExC,yDAAyD;IACzD,MAAM,cAAc,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,cAAc,EAAE,WAAW,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC;IACtG,IAAI,MAAM,CAAC,SAAS,EAAE,CAAC;QACnB,MAAM,QAAQ,GAAG,MAAM,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC,CAAS,EAAE,EAAE,CAAC,CAAC,cAAc,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC;QACrF,cAAc,CAAC,IAAI,CAAC,GAAG,QAAQ,CAAC,CAAC;IACrC,CAAC;IAED,MAAM,OAAO,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,IAAA,4BAAgB,EAAC,MAAM,CAAC,IAAI,EAAE,MAAM,CAAC,QAAQ,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;IAEjG,MAAM,QAAQ,GAAG,KAAK,IAAI,EAAE;QACxB,IAAI,MAAM,CAAC,OAAO,EAAE,IAAI,EAAE,QAAQ,CAAC,KAAK,CAAC,IAAI,MAAM,CAAC,IAAI,EAAE,MAAM,IAAI,MAAM,CAAC,IAAI,EAAE,MAAM,GAAG,GAAG,EAAE,CAAC;YAC5F,OAAO,QAAQ,GAAG,QAAQ,CAAA;QAC9B,CAAC;QACD,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,IAAI,EAAE,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;YAC9C,OAAO,SAAS,CAAC;QACrB,CAAC;QACD,MAAM,GAAG,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC,YAAY,CAAC,QAAQ,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,WAAW,EAAE,mBAAmB,EAAE,IAAI,EAAE,CAAC,CAAC;QAC1H,IAAI,CAAC,GAAG,CAAC,SAAS,IAAI,GAAG,CAAC,MAAM,KAAK,YAAY,EAAE,CAAC;YAChD,oBAAoB;YACpB,MAAM,IAAI,KAAK,CAAC,wBAAwB,QAAQ,iBAAiB,CAAC,CAAC;QACvE,CAAC;aAAM,IAAI,GAAG,CAAC,SAAS,EAAE,CAAC;YACvB,OAAO,QAAQ,GAAG,QAAQ,CAAC;QAC/B,CAAC;IACL,CAAC,CAAA;IAED,MAAM,OAAO,GAAG,MAAM,QAAQ,EAAE,CAAC;IAEjC,cAAG,CAAC,IAAI,CAAC,2EAA2E,GAAG,cAAc,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;IAEjH,MAAM,GAAG,GAAG,MAAM,IAAA,sDAA8B,EAAC,MAAM,EAAE,eAAe,EAAE,MAAM,EAAE;QAC9E,cAAc,EAAE,OAAO,EAAE,KAAK,EAAE,OAAO;KAC1C,CAAC,CAAC;IAEH,cAAG,CAAC,IAAI,CAAC,gCAAgC,GAAG,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC;IAExE,+EAA+E;IAC/E,IAAI,YAAY,GAA6C,SAAS,CAAC;IAEvE,YAAY,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,GAAG,CAAC,MAAM,CAAC,aAAa,CAAC,CAAC;IAEpE,IAAI,CAAC,YAAY,EAAE,CAAC;QAChB,cAAG,CAAC,IAAI,CAAC,wDAAwD,CAAC,CAAC;QACnE,MAAM,OAAO,GAAG,MAAM,eAAe,CAAC,OAAO,EAAE,cAAc,EAAE,OAAO,EAAE,OAAO,CAAC,CAAC;QAEjF,YAAY,GAAG,EAAE,EAAE,EAAE,OAAO,CAAC,EAAE,EAAE,IAAI,EAAE,OAAO,CAAC,IAAI,EAAE,CAAC;IAC1D,CAAC;IAED,IAAI,CAAC,YAAY,EAAE,CAAC;QAChB,cAAG,CAAC,KAAK,CAAC,kBAAkB,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;QAC1C,MAAM,IAAI,KAAK,CAAC,kBAAkB,GAAG,GAAG,CAAC,MAAM,CAAC,aAAa,CAAC,CAAC;IACnE,CAAC;IAED,kCAAkC;IAClC,MAAM,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,QAAQ,EAAE;QAClC,IAAI,EAAE,YAAY,CAAC,EAAE;KACxB,CAAC,CAAC;IAEH,OAAO;QACH,EAAE,EAAE,YAAY,CAAC,EAAE;QACnB,IAAI,EAAE,YAAY,CAAC,IAAI;QACvB,KAAK,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,YAAY,CAAC,IAAI,CAAC;KACxD,CAAC;AACN,CAAC;AAED,KAAK,UAAU,eAAe,CAAC,OAAwB,EAAE,cAAwB,EAAE,OAAgB,EAAE,OAAgB;IACjH,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC;IAEnC,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,YAAY,EAAE,CAAC;IAC7C,MAAM,eAAe,GAAG,MAAM,CAAC,gBAAgB,EAAE,qBAAqB,IAAI,2BAA2B,CAAC;IAEtG,MAAM,UAAU,GAAG,MAAM,IAAA,sDAA8B,EAAC,MAAM,EAAE,eAAe,EAAE,MAAM,EAAE;QACrF,cAAc,EAAE,cAAc;QAC9B,OAAO,EAAE,OAAO;QAChB,aAAa,EAAE,OAAO,EAAE,aAAa,EAAE,aAAa,IAAI,SAAS;QACjE,KAAK,EAAE,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,SAAS;KACvC,CAAC,CAAC;IAGH,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,aAAa,EAAE,CAAC;QACnC,cAAG,CAAC,KAAK,CAAC,4BAA4B,EAAE,UAAU,CAAC,CAAC;QACpD,MAAM,IAAI,KAAK,CAAC,4BAA4B,CAAC,CAAC;IAClD,CAAC;IAED,cAAG,CAAC,IAAI,CAAC,2BAA2B,EAAE,UAAU,CAAC,MAAM,CAAC,eAAe,CAAC,CAAC;IACzE,MAAM,QAAQ,GAAmC;QAC7C,IAAI,EAAE,UAAU,CAAC,MAAM,CAAC,aAAa;QACrC,aAAa,EAAE,UAAU,CAAC,MAAM,CAAC,eAAe;QAChD,YAAY,EAAE,UAAU,CAAC,MAAM,CAAC,YAAY;KAC/C,CAAA;IAED,MAAM,IAAI,GAAG,MAAM,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;IAEjD,OAAO,IAAI,CAAC;AAEhB,CAAC"}