Refactor Docker and entrypoint configurations; update NLTK data handling and logging paths

Dockerfile

@@ -103,6 +103,6 @@ COPY deployment/supervisord.conf /etc/supervisord.conf
 ENTRYPOINT ["/app/deployment/gradio_entrypoint.sh"]
 
 # Ports
-EXPOSE 80 7860
+EXPOSE 7860
 
 CMD ["supervisord", "-c", "/etc/supervisord.conf"]

deployment/gradio_entrypoint.sh

@@ -1,27 +1,51 @@
 #!/bin/bash
-set -e
+set -u
 
 echo "Starting AdvisorAI Data Pipeline with Gradio..."
 
-# Create necessary directories
-mkdir -p /data/logs /data/nltk_data
+# Determine writable data dir via existing Python config logic
+NLTK_DIR=$(python - <<'PY'
+import os
+try:
+    from src.config import DATA_DIR
+except Exception:
+    # fallback order
+    for p in ['/data', '/app/data', '/tmp']:
+        try:
+            os.makedirs(p, exist_ok=True)
+            test = os.path.join(p, '.wtest')
+            open(test,'w').close(); os.remove(test)
+            DATA_DIR = p
+            break
+        except Exception:
+            continue
+    else:
+        DATA_DIR = '/tmp'
+
+nl = os.path.join(DATA_DIR, 'nltk_data')
+os.makedirs(nl, exist_ok=True)
+print(nl)
+PY
+)
 
-# Set proper permissions
-chmod -R 777 /data
+export NLTK_DATA="$NLTK_DIR"
+echo "NLTK_DATA set to: $NLTK_DATA"
 
-# Download NLTK data if needed
-python -c "
-import nltk
+# Best-effort NLTK downloads (do not fail on errors)
+python - <<'PY'
 import os
-os.environ['NLTK_DATA'] = '/data/nltk_data'
+print('Preparing NLTK into', os.environ.get('NLTK_DATA'))
 try:
-    nltk.download('punkt', download_dir='/data/nltk_data', quiet=True)
-    nltk.download('stopwords', download_dir='/data/nltk_data', quiet=True)
-    nltk.download('vader_lexicon', download_dir='/data/nltk_data', quiet=True)
-    print('NLTK data downloaded successfully')
+    import nltk
+    for pkg in ['punkt', 'stopwords', 'vader_lexicon']:
+        try:
+            nltk.download(pkg, download_dir=os.environ.get('NLTK_DATA'), quiet=True)
+            print('Downloaded', pkg)
+        except Exception as e:
+            print('NLTK download failed for', pkg, e)
 except Exception as e:
-    print(f'NLTK download failed: {e}')
-"
+    print('NLTK import failed:', e)
+PY
 
 echo "Starting services..."
 exec "$@"

deployment/nginx.conf

@@ -1,5 +1,6 @@
 server {
-    listen 80;
+    # On Spaces Docker, the platform expects the app to listen on 7860
+    listen 7860;
     
     # Increase timeouts to handle long-running operations
     proxy_connect_timeout       60s;
@@ -18,20 +19,10 @@ server {
     client_body_timeout         60s;
     client_header_timeout       60s;
 
-    # -- health-check: proxy to gradio app --
+    # -- health-check: serve OK directly --
     location = /health {
-        proxy_pass       http://127.0.0.1:7860/;
-        proxy_set_header Host $host;
-        proxy_set_header X-Real-IP $remote_addr;
-        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
-        proxy_set_header X-Forwarded-Proto $scheme;
-        
-        # Shorter timeouts for health checks
-        proxy_connect_timeout 10s;
-        proxy_send_timeout 10s;
-        proxy_read_timeout 10s;
-        
-        # don't log upstream body
+        default_type text/plain;
+        return 200 'OK\n';
         access_log off;
     }
 

deployment/supervisord.conf

@@ -6,7 +6,7 @@ pidfile=/tmp/supervisord.pid
 loglevel=info
 
 [program:gradio]
-command=python /app/src/api/gradio_main.py
+command=/bin/sh -c 'PORT=${PORT:-7860} python /app/src/api/gradio_main.py'
 directory=/app
 autostart=true
 autorestart=true
@@ -14,24 +14,24 @@ stdout_logfile=/dev/stdout
 stderr_logfile=/dev/stderr
 stdout_logfile_maxbytes=0
 stderr_logfile_maxbytes=0
-startsecs=10
+startsecs=5
 startretries=3
-stopwaitsecs=30
+stopwaitsecs=20
 killasgroup=true
 stopasgroup=true
 environment=PYTHONPATH="/app:/app/src:/app/src/api:/app/src/data_cloud:/app/src/fetchers:/app/src/merge"
 
-[program:nginx]
-command=/usr/sbin/nginx -g 'daemon off;'
-autostart=true
-autorestart=true
-stdout_logfile=/dev/stdout
-stderr_logfile=/dev/stderr
-stdout_logfile_maxbytes=0
-stderr_logfile_maxbytes=0
-startsecs=5
-startretries=3
-stopwaitsecs=10
+; [program:nginx]
+; command=/usr/sbin/nginx -g 'daemon off;'
+; autostart=true
+; autorestart=true
+; stdout_logfile=/dev/stdout
+; stderr_logfile=/dev/stderr
+; stdout_logfile_maxbytes=0
+; stderr_logfile_maxbytes=0
+; startsecs=5
+; startretries=3
+; stopwaitsecs=10
 
 [program:scheduler]
 ; wait 180 s before first run, then your scheduler.py handles its own 30 min sleeps

src/api/gradio_main.py

@@ -9,6 +9,18 @@ from datetime import datetime, timedelta
 import psutil
 from pathlib import Path
 
+# Internal config for paths and markers
+try:
+    from src.config import DATA_DIR, LOG_DIR, LAST_RUN_PATH
+except Exception:
+    # Fallbacks if import path differs in Spaces
+    try:
+        from config import DATA_DIR, LOG_DIR, LAST_RUN_PATH  # type: ignore
+    except Exception:
+        DATA_DIR = os.environ.get('DATA_DIR', '/data')
+        LOG_DIR = os.environ.get('LOG_DIR', os.path.join(DATA_DIR, 'logs'))
+        LAST_RUN_PATH = os.environ.get('LAST_RUN_PATH', '/tmp/last_run.txt')
+
 # Add src to Python path for imports
 sys.path.insert(0, '/app/src')
 sys.path.insert(0, '/app')
@@ -37,7 +49,7 @@ def get_health_status():
         scheduler_running = False
         last_run_time = "Unknown"
         try:
-            last_run_file = "/app/deployment/last_run.txt"
+            last_run_file = LAST_RUN_PATH
             if os.path.exists(last_run_file):
                 with open(last_run_file, 'r') as f:
                     last_run_str = f.read().strip()
@@ -71,10 +83,10 @@ def get_pipeline_status():
     """Get data pipeline status"""
     try:
         data_dirs = [
-            "/data/merged/features",
-            "/data/merged/train",
-            "/data/alpaca",
-            "/data/advisorai-data"
+            os.path.join(DATA_DIR, 'merged', 'features'),
+            os.path.join(DATA_DIR, 'merged', 'train'),
+            os.path.join(DATA_DIR, 'alpaca'),
+            os.path.join(DATA_DIR, 'advisorai-data'),
         ]
         
         recent_files = 0
@@ -113,10 +125,10 @@ def get_recent_files():
     """Get list of recent files in the data directories"""
     try:
         base_paths = [
-            "/data/merged/features",
-            "/data/merged/train", 
-            "/data/alpaca",
-            "/data/advisorai-data/features"
+            os.path.join(DATA_DIR, 'merged', 'features'),
+            os.path.join(DATA_DIR, 'merged', 'train'),
+            os.path.join(DATA_DIR, 'alpaca'),
+            os.path.join(DATA_DIR, 'advisorai-data', 'features'),
         ]
         
         recent_files = []
@@ -129,7 +141,7 @@ def get_recent_files():
                             stat = os.stat(file_path)
                             recent_files.append({
                                 "File": file,
-                                "Path": file_path.replace("/data/", ""),
+                                "Path": file_path.replace(DATA_DIR.rstrip('/') + '/', ""),
                                 "Size": f"{stat.st_size / (1024**2):.2f} MB",
                                 "Modified": datetime.fromtimestamp(stat.st_mtime).strftime("%Y-%m-%d %H:%M")
                             })
@@ -215,14 +227,19 @@ with gr.Blocks(title="AdvisorAI Data Pipeline Monitor", theme=gr.themes.Soft())
     def refresh_dashboard():
         health = get_health_status()
         pipeline = get_pipeline_status()
-        return json.dumps(health, indent=2), json.dumps(pipeline, indent=2)
+        # JSON components accept dicts directly
+        return health, pipeline
     
     def refresh_files():
         files = get_recent_files()
-        if files and isinstance(files[0], dict) and "Error" not in files[0]:
-            return [[f["File"], f["Path"], f["Size"], f["Modified"]] for f in files]
-        else:
-            return [["Error", str(files), "", ""]]
+        if not files:
+            return []
+        if isinstance(files, list) and isinstance(files[0], dict) and "Error" not in files[0]:
+            rows = []
+            for f in files:
+                rows.append([f.get("File",""), f.get("Path",""), f.get("Size",""), f.get("Modified","")])
+            return rows
+        return [["Error", str(files), "", ""]]
     
     def refresh_logs():
         return get_logs()
@@ -256,9 +273,10 @@ with gr.Blocks(title="AdvisorAI Data Pipeline Monitor", theme=gr.themes.Soft())
 
 if __name__ == "__main__":
     logger.info("Starting Gradio app...")
+    port = int(os.environ.get("PORT", "7860"))
     app.launch(
         server_name="0.0.0.0",
-        server_port=7860,
+        server_port=port,
         share=False,
         show_error=True,
         quiet=False