reverted s3 code

main.py

@@ -140,11 +140,8 @@ async def process_analysis_request(request: Request):
         **Strategy 2: Generate a Single Python Script.**
         If the user's request is complex, requires multiple steps, data manipulation, or cannot be handled by a single specialized tool, you MUST generate a plan containing a SINGLE step that uses the `python_interpreter`. This single step must contain a complete, self-contained Python script that performs all the necessary actions and prints the final JSON output.
 
-        **Expert Advice for Large Datasets (S3 / Parquet):**
-        - **If the user's request involves querying large Parquet files, especially from an S3 path, the most performant and reliable method is to use the `duckdb` library.**
-        - **To avoid permission errors in a secure environment, your script MUST start by setting environment variables to point configuration directories to the writable `/tmp` folder.** 
         CRITICAL RULES:
-        1.**REMOTE DATA SOURCES:** If the user's request describes a data source using a URL or an S3 path (e.g., `s3://...`), your script MUST attempt to access that data directly. 
+        1.**REMOTE DATA SOURCES:** If the user's request describes a data source using a URL or an S3 path (e.g., `s3://...`), your script MUST attempt to access that data directly.
         2.  **CHOOSE A STRATEGY:** First, analyze the request. Is it a simple task for a specialized tool, or a complex one requiring a full script?
         3.  **TOOL NAMES:** You MUST use the exact tool names from the provided list. 
         4. **SPECIAL TOOL USAGE -