go

Dockerfile

@@ -1,46 +1,43 @@
 # Use the official Python 3.11 slim image
 FROM python:3.11-slim
 
-# Install curl for keep-alive script and clean up
+# Install system dependencies required for Math/Data libraries
+# libgomp1 is CRITICAL for scikit-learn & numpy parallelization to prevent segfaults
 RUN apt-get update && \
-    apt-get install -y curl && \
-    rm -rf /var/lib/apt/lists/*
+    apt-get install -y --no-install-recommends \
+    curl \
+    gcc \
+    g++ \
+    libgomp1 \
+    && rm -rf /var/lib/apt/lists/*
 
 # Set the working directory inside the container
 WORKDIR /app
 
-# Create all required directories with proper permissions upfront
-RUN mkdir -p /app/generated_outputs && \
-    mkdir -p /app/generated_charts && \
-    mkdir -p /app/cache && \
-    chmod -R 777 /app/generated_outputs && \
-    chmod -R 777 /app/generated_charts && \
-    chmod -R 777 /app/cache
+# Create required directories with permissions
+RUN mkdir -p /app/generated_outputs /app/generated_charts /app/cache && \
+    chmod -R 777 /app/generated_outputs /app/generated_charts /app/cache
 
-# Create log files with proper permissions
-RUN touch /app/pandasai.log && \
-    touch /app/api_key_rotation.log && \
-    chmod 666 /app/pandasai.log && \
-    chmod 666 /app/api_key_rotation.log
+# Create log files
+RUN touch /app/pandasai.log /app/api_key_rotation.log && \
+    chmod 666 /app/pandasai.log /app/api_key_rotation.log
 
 # Set environment variables
 ENV MPLCONFIGDIR=/app/cache
 
-# Copy the requirements file first
+# Copy requirements and install
 COPY requirements.txt .
-
-# Install dependencies
 RUN pip install --no-cache-dir -r requirements.txt
 
-# Copy the rest of the application code
+# Copy application code
 COPY . .
 
-# Ensure the user has write permissions to all directories
-RUN chown -R 1000:1000 /app && \
-    chmod -R 777 /app
+# Ensure permissions
+RUN chown -R 1000:1000 /app && chmod -R 777 /app
 
-# Expose port 7860 (required by Hugging Face Spaces)
+# Expose port
 EXPOSE 7860
 
-# Keep-alive command (pings every 5 minutes) + start Uvicorn
-CMD bash -c "while true; do curl -s https://code-api-executor.hf.space/ping >/dev/null && sleep 300; done & uvicorn controller:app --host 0.0.0.0 --port 7860"
+# CMD: Run the python script directly.
+# This ensures the 'if __name__ == "__main__":' block in your code actually runs.
+CMD bash -c "while true; do curl -s https://code-api-executor.hf.space/ping >/dev/null && sleep 300; done & python controller.py"

controller.py

@@ -25,6 +25,7 @@ from bson import ObjectId
 import mysql.connector
 import psycopg2
 from psycopg2.extras import RealDictCursor
+import uvicorn
 
 # --- Existing Services ---
 from csv_analysis_service import execute_analysis_logic
@@ -467,19 +468,22 @@ async def batch_execute_mongo(payload: BatchRequest[ExecutorPayload], token: str
 # ==============================================================================
 
 if __name__ == "__main__":
-    import uvicorn
     host = os.getenv("HOST", "0.0.0.0")
     port = int(os.getenv("PORT", 7860))
     
-    # Scale processes: 16 Cores - 2 = 14 Workers
+    # Calculate workers (save 2 cores for system overhead)
+    # Ensure at least 1 worker exists
     num_workers = max(1, multiprocessing.cpu_count() - 2)
     
     print(f"Starting production server on {host}:{port} with {num_workers} workers...")
+    print("Using 'asyncio' loop to prevent Pandas/Numpy segfaults.")
 
     uvicorn.run(
         "controller:app",  
         host=host, 
         port=port, 
         workers=num_workers,
-        loop="auto",
+        loop="asyncio",
+        log_level="info",
+        access_log=False
     )