From cc1cb0d2a75785c94437f5c48718af58f15edc13 Mon Sep 17 00:00:00 2001
From: siribeesu <siribeesu07@gmail.com>
Date: Thu, 23 Jul 2026 17:23:34 +0530
Subject: [PATCH] Add extraction input size validation

---
 backend/utils/extractValidation.js | 38 ++++++++++++++++++++++++++++++
 server.js                          | 15 ++++++++----
 tests/extractValidation.test.js    | 23 ++++++++++++++++++
 3 files changed, 72 insertions(+), 4 deletions(-)
 create mode 100644 backend/utils/extractValidation.js
 create mode 100644 tests/extractValidation.test.js

diff --git a/backend/utils/extractValidation.js b/backend/utils/extractValidation.js
new file mode 100644
index 00000000..b357aeef
--- /dev/null
+++ b/backend/utils/extractValidation.js
@@ -0,0 +1,38 @@
+const MAX_EXTRACTION_TEXT_LENGTH = 12000;
+
+function validateExtractionText(text) {
+  if (typeof text !== 'string') {
+    return {
+      ok: false,
+      statusCode: 400,
+      error: 'Text must be provided as a string.',
+    };
+  }
+
+  const trimmedText = text.trim();
+  if (!trimmedText) {
+    return {
+      ok: false,
+      statusCode: 400,
+      error: 'Text is required',
+    };
+  }
+
+  if (trimmedText.length > MAX_EXTRACTION_TEXT_LENGTH) {
+    return {
+      ok: false,
+      statusCode: 413,
+      error: `Text exceeds the maximum allowed length of ${MAX_EXTRACTION_TEXT_LENGTH} characters. Please provide a smaller excerpt for AI extraction.`,
+    };
+  }
+
+  return {
+    ok: true,
+    text: trimmedText,
+  };
+}
+
+module.exports = {
+  MAX_EXTRACTION_TEXT_LENGTH,
+  validateExtractionText,
+};
diff --git a/server.js b/server.js
index b5267cb0..98d72f34 100644
--- a/server.js
+++ b/server.js
@@ -5,10 +5,11 @@ const { db, initDb } = require('./database');
 const { GoogleGenAI } = require('@google/genai');
 const path = require('path');
 const csvDownloadRouter = require('./backend/routers/csvDownload.router.js');
+const { validateExtractionText } = require('./backend/utils/extractValidation.js');
 
 const app = express();
 app.use(cors());
-app.use(express.json());
+app.use(express.json({ limit: '1mb' }));
 
 const page404Path = path.join(__dirname, '404.html');
 const page500Path = path.join(__dirname, 'error.html');
@@ -530,7 +531,13 @@ app.delete('/api/tasks/:id', (req, res) => {
 // ================= AI EXTRACTION =================
 app.post('/api/extract', async (req, res) => {
   const { text } = req.body;
-  if (!text) return res.status(400).json({ error: 'Text is required' });
+  const validation = validateExtractionText(text);
+
+  if (!validation.ok) {
+    return res.status(validation.statusCode).json({ error: validation.error });
+  }
+
+  const safeText = validation.text;
 
   if (ai) {
     try {
@@ -540,7 +547,7 @@ Return ONLY a raw JSON array (no markdown, no backticks, no explanation).
 Each object must have: title (string), subject_name (string), due_at (ISO 8601 datetime), notes (string), confidence_score (number 0-100), priority ("low"|"medium"|"high"), icon (emoji).
 IMPORTANT: Do not strip hashtags from the task description! If the original text contains hashtag labels (e.g. #urgent, #Group), you MUST include them at the end of the 'title' field (e.g. 'Read chapter 1 #urgent').
 
-Text: "${text}"
+Text: "${safeText}"
 `;
       const response = await ai.models.generateContent({
         model: 'gemini-2.5-flash',
@@ -559,7 +566,7 @@ Text: "${text}"
   }
 
   // NLP heuristic fallback (no API key, or Gemini failed)
-  const tasks = nlpExtractTasksFromText(text);
+  const tasks = nlpExtractTasksFromText(safeText);
   return res.json(tasks);
 });
 // ================= AUTH =================
diff --git a/tests/extractValidation.test.js b/tests/extractValidation.test.js
new file mode 100644
index 00000000..7cb4d3fd
--- /dev/null
+++ b/tests/extractValidation.test.js
@@ -0,0 +1,23 @@
+const test = require('node:test');
+const assert = require('node:assert/strict');
+
+const {
+  MAX_EXTRACTION_TEXT_LENGTH,
+  validateExtractionText,
+} = require('../backend/utils/extractValidation.js');
+
+test('validateExtractionText accepts a normal-sized prompt', () => {
+  const result = validateExtractionText('Submit math homework by tomorrow.');
+
+  assert.equal(result.ok, true);
+  assert.equal(result.text, 'Submit math homework by tomorrow.');
+});
+
+test('validateExtractionText rejects oversized input with a helpful error', () => {
+  const longText = 'a'.repeat(MAX_EXTRACTION_TEXT_LENGTH + 1);
+  const result = validateExtractionText(longText);
+
+  assert.equal(result.ok, false);
+  assert.equal(result.statusCode, 413);
+  assert.match(result.error, /maximum allowed length/i);
+});