programmingNot4u/bangla-ocr-data-web

# ✍️ BanglaOCR: A Platform to Collect Bangla Handwriting for OCR Models **Bangla OCR Handwriting Data** is a full-stack web application designed to **crowdsource handwritten Bangla text** for building large-scale OCR datasets. It provides a clean submission portal for users and a robust backend for data moderation, tracking, and export. --- ## 🌐 Live Demo * bangla-ocr-data-web.vercel.… (frontend) * hossainimamrony71.pythonany… (API Endpoint) --- ## 🧠 Key Features * Weighted random Bangla prompt selection * Direct, secure client-side uploads to **Cloudinary** * Real-time status updates during submission * RESTful API using **Django + DRF** * Admin-only moderation dashboard (via API for now) * Verified submission export as `.zip` with `CSV` labels * Decoupled architecture: Vanilla JS frontend + DRF backend --- ## 🔧 Backend: Django REST Framework The backend powers secure APIs for serving prompts, logging submissions, moderating content, and exporting verified data. ### 📦 API Endpoints #### `GET /api/prompt/` * **Purpose:** Serve a Bangla text prompt using weighted random logic (prompts with fewer submissions appear more) * **Permissions:** Public * **Response:** ```json { "id": 1, "text": "বাংলা আমার মাতৃভাষা।" } ``` --- #### `GET /api/cloudinary-config/` * **Purpose:** Provide a secure signature for client-side Cloudinary uploads * **Permissions:** Public * **Response:** ```json { "api_key": "...", "cloud_name": "...", "timestamp": "...", "signature": "..." } ``` --- #### `POST /api/submissions/` * **Purpose:** Log metadata after successful Cloudinary upload * **Permissions:** Public * **Body:** ```json { "prompt": 1, "image_url": "res.cloudinary.com", "public_id": "bangla_ocr_handwriting_data/..." } ``` --- #### `GET /api/moderator/pending/` * **Purpose:** View unreviewed submissions for moderation * **Permissions:** Admin only * **Returns:** List of pending entries with image URLs and prom …