LLM refactor, NPU ollama based, publisher update json query to llm

app_urls/fetcher/views.py

@@ -14,16 +14,6 @@ import json
 ####################################################################################################
 
 def llm(request):
-
-    def stream_response(model, text):
-        msg_content = {
-            "role": "user", 
-            "content": text,
-        }
-        response = OllamaClient().client.chat(model=model, messages=[msg_content], stream=True)
-        for chunk in response:
-            yield chunk["message"]["content"]  # Stream each chunk of text
-
     if request.method == 'POST':
         try:
             body_data = json.loads(request.body)
@@ -33,7 +23,7 @@ def llm(request):
             if message is None:
                 return JsonResponse({'error': 'No message found in request'}, status=400)
             
-            return StreamingHttpResponse(stream_response(model, message), content_type="text/plain")
+            return StreamingHttpResponse(OllamaClient().generate_stream(model, message), content_type="text/plain")
         except json.JSONDecodeError:
             return JsonResponse({'error': 'Invalid JSON'}, status=400)
 
@@ -55,13 +45,15 @@ def url_detail_view(request, id):
         url_content = {}
     
     ollama = OllamaClient()
+    # prompt_content = "{}\n{}\n{}".format(url_content.title, url_content.description, url_content.content)
+    prompt_content = "{}".format(url_content.content)
 
     context = {
         'url_item': url_item,
         'sources': url_sources,
         'searches': url_searches,
         'models': ollama.get_models(),
-        'prompt': ollama.get_prompt(),
+        'prompt': ollama.get_prompt(prompt_content),
         'url_content': url_content,
         'url_canonical': url_canonical,
     }