@editneo/pdf 0.1.0

package/dist/worker.d.ts

@@ -0,0 +1,3 @@
+import { NeoBlock } from '@editneo/core';
+export declare function extractBlocksFromPdf(data: ArrayBuffer): Promise<NeoBlock[]>;
+//# sourceMappingURL=worker.d.ts.map

package/dist/worker.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"worker.d.ts","sourceRoot":"","sources":["../src/worker.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAa,MAAM,eAAe,CAAC;AAOpD,wBAAsB,oBAAoB,CAAC,IAAI,EAAE,WAAW,GAAG,OAAO,CAAC,QAAQ,EAAE,CAAC,CAkEjF"}

package/dist/worker.js

@@ -0,0 +1,79 @@
+import * as pdfjs from 'pdfjs-dist';
+import { v4 as uuid } from 'uuid';
+// Configure worker - in a real app, this needs careful path handling or a bundler plugin
+// For now, we assume the worker is loaded by the main thread or handled by the build system
+// pdfjs.GlobalWorkerOptions.workerSrc = ...; 
+export async function extractBlocksFromPdf(data) {
+    const loadingTask = pdfjs.getDocument(data);
+    const pdf = await loadingTask.promise;
+    const blocks = [];
+    for (let i = 1; i <= pdf.numPages; i++) {
+        const page = await pdf.getPage(i);
+        const textContent = await page.getTextContent();
+        const opList = await page.getOperatorList();
+        // Simplistic mode font size calculation for the page
+        const fontSizes = {};
+        for (const item of textContent.items) {
+            if ('height' in item) {
+                const height = Math.floor(item.height);
+                fontSizes[height] = (fontSizes[height] || 0) + 1;
+            }
+        }
+        let modeFontSize = 12;
+        let maxCount = 0;
+        for (const size in fontSizes) {
+            if (fontSizes[size] > maxCount) {
+                maxCount = fontSizes[size];
+                modeFontSize = Number(size);
+            }
+        }
+        // Heuristic 1: Text extraction
+        // This is a simplified extraction that doesn't handle layout perfectly but follows the plan
+        let currentBlockText = '';
+        let currentBlockType = 'paragraph';
+        let lastY = -1;
+        for (const item of textContent.items) {
+            if (!('str' in item))
+                continue;
+            // Simple line grouping test
+            if (lastY !== -1 && Math.abs(item.transform[5] - lastY) > 5) {
+                if (currentBlockText.trim()) {
+                    blocks.push(createBlock(currentBlockType, currentBlockText));
+                }
+                currentBlockText = '';
+                // Reset type based on new line height
+                if (item.height > modeFontSize * 2)
+                    currentBlockType = 'heading-1';
+                else if (item.height > modeFontSize * 1.5)
+                    currentBlockType = 'heading-2';
+                else
+                    currentBlockType = 'paragraph';
+            }
+            currentBlockText += item.str + ' ';
+            lastY = item.transform[5];
+        }
+        if (currentBlockText.trim()) {
+            blocks.push(createBlock(currentBlockType, currentBlockText));
+        }
+        // Heuristic 2: Images (Simplistic mock logic as real extraction form opList is complex)
+        // Real implementation requires processing opList.fnArray and argsArray for PaintImageXObject
+        // For this task, we will just add a placeholder if we detect image ops
+        if (opList.fnArray.includes(pdfjs.OPS.paintImageXObject)) {
+            blocks.push(createBlock('image', '', { src: 'placeholder-image-url' }));
+        }
+    }
+    return blocks;
+}
+function createBlock(type, text, props = {}) {
+    return {
+        id: uuid(),
+        type,
+        content: [{ text: text.trim() }],
+        props,
+        children: [],
+        parentId: null,
+        createdAt: Date.now(),
+        updatedAt: Date.now(),
+    };
+}
+//# sourceMappingURL=worker.js.map

package/dist/worker.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"worker.js","sourceRoot":"","sources":["../src/worker.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,KAAK,MAAM,YAAY,CAAC;AAEpC,OAAO,EAAE,EAAE,IAAI,IAAI,EAAE,MAAM,MAAM,CAAC;AAElC,yFAAyF;AACzF,4FAA4F;AAC5F,8CAA8C;AAE9C,MAAM,CAAC,KAAK,UAAU,oBAAoB,CAAC,IAAiB;IAC1D,MAAM,WAAW,GAAG,KAAK,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC;IAC5C,MAAM,GAAG,GAAG,MAAM,WAAW,CAAC,OAAO,CAAC;IACtC,MAAM,MAAM,GAAe,EAAE,CAAC;IAE9B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,GAAG,CAAC,QAAQ,EAAE,CAAC,EAAE,EAAE,CAAC;QACvC,MAAM,IAAI,GAAG,MAAM,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;QAClC,MAAM,WAAW,GAAG,MAAM,IAAI,CAAC,cAAc,EAAE,CAAC;QAChD,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,eAAe,EAAE,CAAC;QAE5C,qDAAqD;QACrD,MAAM,SAAS,GAA2B,EAAE,CAAC;QAC7C,KAAK,MAAM,IAAI,IAAI,WAAW,CAAC,KAAK,EAAE,CAAC;YACrC,IAAI,QAAQ,IAAI,IAAI,EAAE,CAAC;gBACrB,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;gBACvC,SAAS,CAAC,MAAM,CAAC,GAAG,CAAC,SAAS,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;YACnD,CAAC;QACH,CAAC;QAED,IAAI,YAAY,GAAG,EAAE,CAAC;QACtB,IAAI,QAAQ,GAAG,CAAC,CAAC;QACjB,KAAK,MAAM,IAAI,IAAI,SAAS,EAAE,CAAC;YAC7B,IAAI,SAAS,CAAC,IAAI,CAAC,GAAG,QAAQ,EAAE,CAAC;gBAC/B,QAAQ,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC;gBAC3B,YAAY,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC;YAC9B,CAAC;QACH,CAAC;QAED,+BAA+B;QAC/B,4FAA4F;QAC5F,IAAI,gBAAgB,GAAG,EAAE,CAAC;QAC1B,IAAI,gBAAgB,GAAc,WAAW,CAAC;QAC9C,IAAI,KAAK,GAAG,CAAC,CAAC,CAAC;QAEf,KAAK,MAAM,IAAI,IAAI,WAAW,CAAC,KAAK,EAAE,CAAC;YACnC,IAAI,CAAC,CAAC,KAAK,IAAI,IAAI,CAAC;gBAAE,SAAS;YAE/B,4BAA4B;YAC5B,IAAI,KAAK,KAAK,CAAC,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;gBACzD,IAAI,gBAAgB,CAAC,IAAI,EAAE,EAAE,CAAC;oBAC1B,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,gBAAgB,EAAE,gBAAgB,CAAC,CAAC,CAAC;gBACjE,CAAC;gBACD,gBAAgB,GAAG,EAAE,CAAC;gBACtB,sCAAsC;gBACtC,IAAI,IAAI,CAAC,MAAM,GAAG,YAAY,GAAG,CAAC;oBAAE,gBAAgB,GAAG,WAAW,CAAC;qBAC9D,IAAI,IAAI,CAAC,MAAM,GAAG,YAAY,GAAG,GAAG;oBAAE,gBAAgB,GAAG,WAAW,CAAC;;oBACrE,gBAAgB,GAAG,WAAW,CAAC;YACzC,CAAC;YAED,gBAAgB,IAAI,IAAI,CAAC,GAAG,GAAG,GAAG,CAAC;YACnC,KAAK,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC;QAC9B,CAAC;QAED,IAAI,gBAAgB,CAAC,IAAI,EAAE,EAAE,CAAC;YAC1B,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,gBAAgB,EAAE,gBAAgB,CAAC,CAAC,CAAC;QACjE,CAAC;QAED,wFAAwF;QACxF,6FAA6F;QAC7F,uEAAuE;QACvE,IAAI,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,iBAAiB,CAAC,EAAE,CAAC;YACtD,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,OAAO,EAAE,EAAE,EAAE,EAAE,GAAG,EAAE,uBAAuB,EAAE,CAAC,CAAC,CAAC;QAC7E,CAAC;IACH,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,WAAW,CAAC,IAAe,EAAE,IAAY,EAAE,QAA6B,EAAE;IAC/E,OAAO;QACH,EAAE,EAAE,IAAI,EAAE;QACV,IAAI;QACJ,OAAO,EAAE,CAAC,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE,EAAE,CAAC;QAChC,KAAK;QACL,QAAQ,EAAE,EAAE;QACZ,QAAQ,EAAE,IAAI;QACd,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE;QACrB,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE;KACxB,CAAC;AACN,CAAC"}

package/package.json

@@ -0,0 +1,24 @@
+{
+  "name": "@editneo/pdf",
+  "version": "0.1.0",
+  "description": "PDF transmutation engine for EditNeo — converts PDFs into editable blocks",
+  "main": "./dist/worker.js",
+  "types": "./dist/worker.d.ts",
+  "files": ["dist"],
+  "scripts": {
+    "build": "tsc",
+    "check-types": "tsc --noEmit"
+  },
+  "publishConfig": {
+    "access": "public"
+  },
+  "dependencies": {
+    "pdfjs-dist": "3.11.174",
+    "@editneo/core": "*"
+  },
+  "devDependencies": {
+    "typescript": "^5.3.3",
+    "@types/pdfjs-dist": "^2.10.378"
+  },
+  "license": "MIT"
+}