@credal/actions 0.2.120 → 0.2.122

package/dist/utils/pdf.d.ts

@@ -1 +1 @@
-export declare function extractTextFromPdf(buffer: ArrayBuffer): Promise<string>;
+export declare function extractTextFromPdf(input: ArrayBuffer | Uint8Array): Promise<string>;

package/dist/utils/pdf.js

@@ -7,34 +7,23 @@ var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, ge
         step((generator = generator.apply(thisArg, _arguments || [])).next());
     });
 };
-import PDFParser from "pdf2json";
-export function extractTextFromPdf(buffer) {
+// npm i pdfjs-dist
+import { getDocument } from "pdfjs-dist";
+export function extractTextFromPdf(input) {
     return __awaiter(this, void 0, void 0, function* () {
-        try {
-            const extractedText = yield new Promise((resolve, reject) => {
-                const pdfParser = new PDFParser();
-                pdfParser.on("pdfParser_dataError", (errData) => {
-                    reject(errData.parserError || new Error("PDF parsing failed"));
-                });
-                pdfParser.on("pdfParser_dataReady", (pdfData) => {
-                    try {
-                        const text = pdfData.Pages.map((page) => page.Texts.map((textItem) => {
-                            // Handle cases where R array might be empty or have multiple runs
-                            return textItem.R.map((run) => decodeURIComponent(run.T)).join("");
-                        }).join("")).join("\n");
-                        resolve(text);
-                    }
-                    catch (error) {
-                        reject(error);
-                    }
-                });
-                pdfParser.parseBuffer(Buffer.from(buffer));
-            });
-            return extractedText;
-        }
-        catch (error) {
-            console.error("Error extracting PDF text:", error);
-            throw error;
+        const data = input instanceof Uint8Array ? input : new Uint8Array(input);
+        // Load PDF
+        const loadingTask = getDocument({ data });
+        const pdf = yield loadingTask.promise;
+        const pages = [];
+        for (let i = 1; i <= pdf.numPages; i++) {
+            const page = yield pdf.getPage(i);
+            const content = yield page.getTextContent();
+            // content.items is typed as TextItem | TextMarkedContent
+            const strings = content.items.map(item => ("str" in item ? item.str : "")).join(" ");
+            pages.push(strings.trim());
         }
+        yield pdf.destroy();
+        return pages.join("\n\n");
     });
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@credal/actions",
-  "version": "0.2.120",
+  "version": "0.2.122",
   "type": "module",
   "description": "AI Actions by Credal AI",
   "sideEffects": false,
@@ -33,6 +33,7 @@
     "@types/jsonwebtoken": "^9.0.9",
     "@types/node": "^22.10.1",
     "@types/node-forge": "^1.3.11",
+    "@types/pdf-parse": "^1.1.5",
     "@typescript-eslint/eslint-plugin": "^8.18.0",
     "@typescript-eslint/parser": "^8.18.0",
     "eslint": "^9.16.0",
@@ -68,6 +69,7 @@
     "node-forge": "^1.3.1",
     "p-limit": "^7.1.1",
     "pdf2json": "^3.1.6",
+    "pdfjs-dist": "^5.4.149",
     "resend": "^4.7.0",
     "snowflake-sdk": "^2.0.2",
     "ts-node": "^10.9.2",