Upd user_id

helpers/namer.py

@@ -79,7 +79,9 @@ Return only the session name, nothing else."""
                 system_prompt=sys_prompt,
                 user_prompt=user_prompt,
                 gemini_rotator=None,
-                nvidia_rotator=nvidia_rotator
+                nvidia_rotator=nvidia_rotator,
+                user_id=user_id,
+                context="session_naming"
             )
             
             logger.info(f"[NAMER] Raw API response: {response}")

legacy.py

@@ -673,7 +673,7 @@ async def generate_report(
         
         try:
             selection_filter = {"provider": "gemini", "model": os.getenv("GEMINI_MED", "gemini-2.5-flash")}
-            filter_response = await generate_answer_with_model(selection_filter, filter_sys, filter_user, gemini_rotator, nvidia_rotator)
+            filter_response = await generate_answer_with_model(selection_filter, filter_sys, filter_user, gemini_rotator, nvidia_rotator, user_id="system", context="legacy_filter")
             logger.info(f"[REPORT] Raw filter response: {filter_response}")
             # Try to parse the filter response to get relevant chunks
             import json
@@ -709,7 +709,7 @@ async def generate_report(
     try:
         # Step 1: Outline with Flash/Med
         selection_outline = {"provider": "gemini", "model": os.getenv("GEMINI_MED", "gemini-2.5-flash")}
-        outline_md = await generate_answer_with_model(selection_outline, sys_outline, user_outline, gemini_rotator, nvidia_rotator)
+        outline_md = await generate_answer_with_model(selection_outline, sys_outline, user_outline, gemini_rotator, nvidia_rotator, user_id="system", context="legacy_outline")
     except Exception as e:
         logger.warning(f"Report outline failed: {e}")
         outline_md = "# Report Outline\n\n- Introduction\n- Key Topics\n- Conclusion"
@@ -731,7 +731,7 @@ async def generate_report(
 
     try:
         selection_report = {"provider": "gemini", "model": os.getenv("GEMINI_PRO", "gemini-2.5-pro")}
-        report_md = await generate_answer_with_model(selection_report, sys_report, user_report, gemini_rotator, nvidia_rotator)
+        report_md = await generate_answer_with_model(selection_report, sys_report, user_report, gemini_rotator, nvidia_rotator, user_id="system", context="legacy_report")
     except Exception as e:
         logger.error(f"Report generation failed: {e}")
         report_md = outline_md + "\n\n" + file_summary

memo/sessions.py

@@ -174,7 +174,9 @@ Is this a context switch?"""
                         system_prompt=sys_prompt,
                         user_prompt=user_prompt,
                         gemini_rotator=None,
-                        nvidia_rotator=nvidia_rotator
+                        nvidia_rotator=nvidia_rotator,
+                        user_id="system",
+                        context="context_switch_detection"
                     )
                     
                     # Parse JSON response

utils/api/router.py

@@ -135,11 +135,11 @@ async def generate_answer_with_model(selection: Dict[str, Any], system_prompt: s
             if model in [GEMINI_PRO, GEMINI_MED]:
                 logger.info(f"Falling back from {model} to NVIDIA_LARGE")
                 fallback_selection = {"provider": "nvidia_large", "model": NVIDIA_LARGE}
-                return await generate_answer_with_model(fallback_selection, system_prompt, user_prompt, gemini_rotator, nvidia_rotator)
+                return await generate_answer_with_model(fallback_selection, system_prompt, user_prompt, gemini_rotator, nvidia_rotator, user_id, context)
             elif model == GEMINI_SMALL:
                 logger.info(f"Falling back from {model} to NVIDIA_SMALL")
                 fallback_selection = {"provider": "nvidia", "model": NVIDIA_SMALL}
-                return await generate_answer_with_model(fallback_selection, system_prompt, user_prompt, gemini_rotator, nvidia_rotator)
+                return await generate_answer_with_model(fallback_selection, system_prompt, user_prompt, gemini_rotator, nvidia_rotator, user_id, context)
             else:
                 logger.error(f"No fallback defined for Gemini model: {model}")
                 return "I couldn't parse the model response."

utils/service/pdf.py

@@ -691,7 +691,7 @@ Return only the formatted references, one per line, numbered sequentially."""
         user_prompt = f"Format these sources in IEEE style:\n\n{source_data}"
         
         selection = {"provider": "nvidia", "model": "meta/llama-3.1-8b-instruct"}
-        response = await generate_answer_with_model(selection, sys_prompt, user_prompt, None, nvidia_rotator)
+        response = await generate_answer_with_model(selection, sys_prompt, user_prompt, None, nvidia_rotator, user_id="system", context="pdf_citation")
         
         # Parse the response into individual references
         references = [line.strip() for line in response.split('\n') if line.strip() and line.strip().startswith('[')]