Update custom domain to urdu-sentiment.hmuhammadusman.com

README.md

@@ -3,7 +3,7 @@
 Welcome to the Urdu Sentiment and Emotion Analysis Engine project! This repository contains the code for a multilingual NLP system that classifies sentiment (Positive, Negative, Neutral) and emotion (Joy, Anger, Fear, Sadness) from Urdu, Roman Urdu, and mixed-language text using a fine-tuned XLM-RoBERTa transformer.
 
 ## Current Progress: Phase 9 (Modal.com Deployment — In Progress)
-The project has successfully completed Phases 1 through 8. The AI models are fully trained, uploaded to Hugging Face Hub (`usman-ai-dev/urdu-sentiment-xlmr` & `usman-ai-dev/urdu-emotion-xlmr`), and integrated into a production-ready **FastAPI** web server with Uvicorn. The frontend features a dark-mode Glassmorphism dashboard with an interactive 3D WebGL Three.js particle wave background, floating ambient glowing orbs, real-time cursor spotlight, Chart.js analytics, and automated live tweet feed streaming. Phase 9 deploys the full stack to **Modal.com** with a custom domain (`urdu-ai.hmuhammadusman.com`).
+The project has successfully completed Phases 1 through 8. The AI models are fully trained, uploaded to Hugging Face Hub (`usman-ai-dev/urdu-sentiment-xlmr` & `usman-ai-dev/urdu-emotion-xlmr`), and integrated into a production-ready **FastAPI** web server with Uvicorn. The frontend features a dark-mode Glassmorphism dashboard with an interactive 3D WebGL Three.js particle wave background, floating ambient glowing orbs, real-time cursor spotlight, Chart.js analytics, and automated live tweet feed streaming. Phase 9 deploys the full stack to **Modal.com** with a custom domain (`urdu-sentiment.hmuhammadusman.com`).
 
 ### Repository Structure
 - `app.py`: FastAPI Web Server exposing all REST API routes (`/analyze`, `/analytics`, `/detect-language`, `/live-feed`, `/health`).
@@ -67,5 +67,5 @@ The server will boot up and listen on `http://127.0.0.1:5000`.
 - **Phase 8** ✅: Models pushed to Hugging Face Hub. `predictor.py` updated to load from Hub.
 - **Phase 9**: Deploy full FastAPI stack to **Modal.com** (free $30/month credit tier).
   - Run: `modal deploy modal_app.py`
-  - Custom domain: `urdu-ai.hmuhammadusman.com`
+  - Custom domain: `urdu-sentiment.hmuhammadusman.com`
   - GitHub Actions auto-deploys on every push to `main`.

modal_app.py

@@ -1,19 +1,20 @@
 """
-modal_app.py — Phase 9: Modal.com Deployment Entrypoint
----------------------------------------------------------
-Wraps the existing FastAPI app (app.py) for serverless deployment
-on Modal.com with custom domain support.
-
-Deploy command:
-    modal deploy modal_app.py
-
-Local test command:
-    modal serve modal_app.py
+modal_app.py — Phase 9: Modal.com Deployment (Modal v1.5+)
+------------------------------------------------------------
+Deploy:  python -m modal deploy modal_app.py
+Test:    python -m modal serve modal_app.py
 """
 
 import modal
 
-# ── Docker image with all dependencies ────────────────────────────────────────
+# ── Source files to include (explicit list — no large dirs) ───────────────────
+SOURCE_FILES = [
+    "app.py",
+    "predictor.py",
+    "lang_detector.py",
+]
+
+# ── Build container image ─────────────────────────────────────────────────────
 image = (
     modal.Image.debian_slim(python_version="3.10")
     .pip_install(
@@ -28,38 +29,28 @@ image = (
         "scikit-learn==1.4.0",
         "accelerate==0.29.0",
     )
+    # Copy only the essential Python source files (not models/, urdu_env/, etc.)
+    .add_local_file("app.py",          remote_path="/app/app.py")
+    .add_local_file("predictor.py",    remote_path="/app/predictor.py")
+    .add_local_file("lang_detector.py",remote_path="/app/lang_detector.py")
+    # Copy frontend assets
+    .add_local_dir("templates",        remote_path="/app/templates")
+    .add_local_dir("static",           remote_path="/app/static")
 )
 
-# ── Modal app definition ───────────────────────────────────────────────────────
+# ── Modal app ─────────────────────────────────────────────────────────────────
 app = modal.App("urdu-sentiment-engine", image=image)
 
-# ── Mount local project files into the container ──────────────────────────────
-project_mount = modal.Mount.from_local_dir(
-    ".",
-    remote_path="/app",
-    # Exclude large/unnecessary directories
-    condition=lambda path: not any(
-        part in path for part in [
-            "models", "urdu_env", "__pycache__", ".git",
-            "data", "logs", "results", "training", "evaluation",
-            ".github", "kaggle_upload.zip",
-        ]
-    ),
-)
 
 @app.function(
-    mounts=[project_mount],
-    # Keeps one container warm to avoid cold starts on the first request
-    min_containers=1,
-    # Give enough CPU/memory for the two XLM-RoBERTa models (~4GB RAM)
-    memory=4096,
+    min_containers=1,   # Keep one warm to avoid cold starts
+    memory=4096,        # 4GB RAM for two XLM-RoBERTa models (~2.2GB)
     cpu=2.0,
-    # Models download from HF Hub on first cold-start then are cached
-    timeout=300,
+    timeout=300,        # Allow 5min on cold start for model download from HF Hub
 )
 @modal.asgi_app()
 def fastapi_app():
     import sys
     sys.path.insert(0, "/app")
-    from app import app as _app          # Import the existing FastAPI app
+    from app import app as _app
     return _app