Neo4j override on telegram bot messages

pxy_bots/views.py

@@ -1,50 +1,52 @@
 import json
-from telegram import Update, Bot
+import logging
 from django.http import JsonResponse
 from django.views.decorators.csrf import csrf_exempt
 from asgiref.sync import sync_to_async
-from .models import TelegramBot
+from telegram import Update, Bot
+from django.utils import timezone
+from django.conf import settings
+
+from .models import TelegramBot, TelegramConversation, TelegramMessage
 from pxy_langchain.services import LangchainAIService
 from .handlers import dream_city_command, start, help_command, handle_location
-import logging
-from .models import TelegramConversation, TelegramMessage, TelegramBot
-from django.utils import timezone
-
 
 logger = logging.getLogger(__name__)
 
 @csrf_exempt
 async def telegram_webhook(request, bot_name):
     """
-    Webhook view that handles Telegram updates asynchronously and logs messages.
+    Telegram webhook handler that logs inbound/outbound messages and
+    responds with AI or commands, resilient to Neo4j or send failures.
+    Always returns 200 OK to prevent Telegram retries.
     """
     try:
         logger.info(f"Webhook called for bot: {bot_name}")
 
-        # 1) Fetch the bot instance
+        # 1) Load bot configuration
         try:
             bot_instance = await sync_to_async(TelegramBot.objects.get)(
                 name=bot_name, is_active=True
             )
-            logger.info(f"Loaded bot configuration: {bot_instance}")
         except TelegramBot.DoesNotExist:
             logger.error(f"Bot '{bot_name}' not found or inactive.")
-            return JsonResponse({"error": f"Bot '{bot_name}' not found."}, status=400)
+            return JsonResponse({"status": "ok", "error": f"Bot '{bot_name}' not found."}, status=200)
 
+        # 2) Ensure POST
         if request.method != "POST":
-            logger.warning("Received non-POST request")
-            return JsonResponse({"error": "Invalid request method"}, status=405)
+            logger.warning("Received non-POST request to Telegram webhook.")
+            return JsonResponse({"status": "ok"}, status=200)
 
-        # 2) Parse the incoming update
+        # 3) Parse the update JSON
         try:
             payload = json.loads(request.body.decode("utf-8"))
             update = Update.de_json(payload, Bot(token=bot_instance.token))
-            logger.info(f"Update received: {update}")
+            logger.info(f"Parsed update: {update}")
         except Exception as e:
-            logger.error(f"Failed to parse update: {e}")
-            return JsonResponse({"error": "Invalid JSON payload"}, status=400)
+            logger.error(f"Failed to parse update JSON: {e}")
+            return JsonResponse({"status": "ok", "error": "Invalid JSON payload"}, status=200)
 
-        # 3) Log conversation & inbound message
+        # 4) Log inbound message
         user_id = str(update.effective_user.id)
         conv, _ = await sync_to_async(TelegramConversation.objects.get_or_create)(
             bot=bot_instance,
@@ -52,47 +54,62 @@ async def telegram_webhook(request, bot_name):
             defaults={'started_at': timezone.now()}
         )
         incoming_text = update.message.text or ""
-        await sync_to_async(TelegramMessage.objects.create)(
-            conversation=conv,
-            direction=TelegramMessage.IN,
-            content=incoming_text
-        )
+        try:
+            await sync_to_async(TelegramMessage.objects.create)(
+                conversation=conv,
+                direction=TelegramMessage.IN,
+                content=incoming_text
+            )
+        except Exception as log_in_err:
+            logger.error(f"Error logging inbound message: {log_in_err}")
 
-        # 4) Route commands or AI
+        # 5) Handle commands or AI response
         if update.message:
-            # built-in commands
-            if update.message.text == "/start":
-                await start(update)
-            elif update.message.text == "/help":
-                await help_command(update)
-            elif update.message.text == "/dream_city":
-                await dream_city_command(update)
-            elif update.message.location:
-                await handle_location(update)
-            else:
-                # AI fallback
-                assistant_instance = await sync_to_async(LangchainAIService)(
-                    bot_instance.assistant
-                )
-                start_time = timezone.now()
-                bot_response = await sync_to_async(
-                    assistant_instance.generate_response
-                )(update.message.text)
-                response_time = int((timezone.now() - start_time).total_seconds() * 1000)
+            try:
+                text = update.message.text or ""
+                if text == "/start":
+                    await start(update)
+                elif text == "/help":
+                    await help_command(update)
+                elif text == "/dream_city":
+                    await dream_city_command(update)
+                elif update.message.location:
+                    await handle_location(update)
+                else:
+                    # AI fallback with resilience
+                    try:
+                        assistant = await sync_to_async(LangchainAIService)(bot_instance.assistant)
+                        start_time = timezone.now()
+                        bot_response = await sync_to_async(assistant.generate_response)(text)
+                        response_time = int((timezone.now() - start_time).total_seconds() * 1000)
+                    except Exception as ai_err:
+                        logger.error(f"AI service error: {ai_err}")
+                        bot_response = "Lo siento, el servicio de IA no está disponible."
+                        response_time = 0
 
-                # Send reply
-                await update.message.reply_text(bot_response)
+                    # Send reply (skipped in DEBUG)
+                    if not settings.DEBUG:
+                        try:
+                            await update.message.reply_text(bot_response)
+                        except Exception as send_err:
+                            logger.error(f"Error sending message to Telegram: {send_err}")
 
-                # 5) Log outbound message
-                await sync_to_async(TelegramMessage.objects.create)(
-                    conversation=conv,
-                    direction=TelegramMessage.OUT,
-                    content=bot_response,
-                    response_time_ms=response_time
-                )
+                    # Log outbound message
+                    try:
+                        await sync_to_async(TelegramMessage.objects.create)(
+                            conversation=conv,
+                            direction=TelegramMessage.OUT,
+                            content=bot_response,
+                            response_time_ms=response_time
+                        )
+                    except Exception as log_out_err:
+                        logger.error(f"Error logging outbound message: {log_out_err}")
+            except Exception as cmd_err:
+                logger.error(f"Error processing Telegram commands: {cmd_err}")
 
-        return JsonResponse({"status": "ok"})
+        # 6) Always return 200 OK
+        return JsonResponse({"status": "ok"}, status=200)
 
     except Exception as e:
-        logger.error(f"Error in telegram_webhook: {e}")
-        return JsonResponse({"error": f"Unexpected error: {e}"}, status=500)
+        logger.error(f"Unexpected error in telegram_webhook: {e}")
+        return JsonResponse({"status": "ok", "error": str(e)}, status=200)

pxy_neo4j/neo4j_connector.py

@@ -9,10 +9,9 @@ from datetime import datetime
 from langchain.prompts import ChatPromptTemplate
 from langchain_community.chains.graph_qa.cypher import GraphCypherQAChain
 
-
 logger = logging.getLogger(__name__)
 
-DEFAULT_PROFILE_NAME = "DefaultNeo4jProfile"  # Define a default profile
+DEFAULT_PROFILE_NAME = "DefaultNeo4jProfile"
 
 class Neo4jDatabase:
     """
@@ -28,84 +27,99 @@ class Neo4jDatabase:
                 logger.warning("No profile specified. Using default Neo4j profile.")
                 self.profile = Neo4jProfile.objects.get(name=DEFAULT_PROFILE_NAME)
 
-            # Connect to the assigned Neo4j instance
-            self.graph = Neo4jGraph(
-                url=self.profile.uri,
-                username=self.profile.username,
-                password=self.profile.password
-            )
+            # Attempt to connect to the assigned Neo4j instance
+            try:
+                self.graph = Neo4jGraph(
+                    url=self.profile.uri,
+                    username=self.profile.username,
+                    password=self.profile.password
+                )
+            except Exception as e:
+                logger.error(f"Failed to initialize Neo4jGraph: {e}")
+                self.graph = None
 
-            # Load the correct AI model
-            self.llm = ChatOpenAI(
-                api_key=self.profile.openai_api_key,
-                model_name=self.profile.model_name,
-                temperature=0
-            )
+            # Initialize LLM if graph is available
+            try:
+                self.llm = ChatOpenAI(
+                    api_key=self.profile.openai_api_key,
+                    model_name=self.profile.model_name,
+                    temperature=0
+                )
+            except Exception as e:
+                logger.error(f"Failed to initialize ChatOpenAI: {e}")
+                self.llm = None
 
-            # Add the custom prompt to enforce English
-            english_prompt = ChatPromptTemplate.from_template(
-                "Transform the following text into a graph structure. All nodes, relationships, and properties should be in English, regardless of the original language."
-            )
-
-            # Apply the custom English prompt here
-            self.graph_transformer = LLMGraphTransformer(
-                llm=self.llm,
-                prompt=english_prompt
-            )
+            # Prepare graph transformer only if llm available
+            if self.llm:
+                english_prompt = ChatPromptTemplate.from_template(
+                    "Transform the following text into a graph structure. All nodes, relationships, and properties should be in English, regardless of the original language."
+                )
+                try:
+                    self.graph_transformer = LLMGraphTransformer(
+                        llm=self.llm,
+                        prompt=english_prompt
+                    )
+                except Exception as e:
+                    logger.error(f"Failed to initialize LLMGraphTransformer: {e}")
+                    self.graph_transformer = None
+            else:
+                self.graph_transformer = None
 
             logger.info(f"Neo4jDatabase initialized with profile: {self.profile.name}")
 
         except Neo4jProfile.DoesNotExist:
             logger.error(f"Neo4j profile '{profile_name}' not found.")
-            raise
+            self.graph = None
+            self.llm = None
+            self.graph_transformer = None
         except Exception as e:
-            logger.error(f"Failed to initialize Neo4jDatabase: {str(e)}")
-            raise
+            logger.error(f"Unexpected error initializing Neo4jDatabase: {e}")
+            self.graph = None
+            self.llm = None
+            self.graph_transformer = None
 
     def store_interaction(self, user_id, bot_id, user_message, bot_response, platform):
         """
         Stores a chatbot interaction as a structured graph in Neo4j.
         Converts messages into `Document` objects to work with `LLMGraphTransformer`.
         """
+        if not self.graph_transformer:
+            logger.warning("Graph transformer not available, skipping store_interaction.")
+            return
+
         try:
-
             timestamp = datetime.utcnow().isoformat()
-
-            # Convert messages into `Document` objects
             documents = [
                 Document(page_content=user_message, metadata={"role": "user", "user_id": user_id, "platform": platform, "created_at": timestamp}),
                 Document(page_content=bot_response, metadata={"role": "bot", "bot_id": bot_id, "platform": platform, "created_at": timestamp}),
             ]
-
-            # Convert text into structured graph documents
             graph_docs = self.graph_transformer.convert_to_graph_documents(documents)
-
-            # Store the structured graph data into Neo4j
-            self.graph.add_graph_documents(graph_docs, include_source=True)
-
-            logger.info(f"Stored interaction in Neo4j (Profile: {self.profile.name})")
+            if self.graph:
+                self.graph.add_graph_documents(graph_docs, include_source=True)
+                logger.info(f"Stored interaction in Neo4j (Profile: {self.profile.name})")
+            else:
+                logger.warning("No Neo4j graph available, skipping add_graph_documents.")
         except Exception as e:
-            logger.error(f"Failed to store interaction in Neo4j: {str(e)}")
-            raise
+            logger.error(f"Failed to store interaction in Neo4j: {e}")
 
     def query_graph(self, user_query):
         """
         Queries the graph using GraphCypherQAChain and returns a structured response.
         """
-        logger.info(f"Calling Neo4j {user_query}")
+        if not self.graph or not self.llm:
+            logger.warning("Graph or LLM not available, cannot query.")
+            return None
+
         try:
-            # Use AI model to generate Cypher query
             qa_chain = GraphCypherQAChain.from_llm(
                 llm=self.llm,
                 graph=self.graph,
                 verbose=True,
                 allow_dangerous_requests=True
             )
-
             result = qa_chain.invoke({"query": user_query})
-            logger.info(f"Resulting Neo4j {result}")
-            return result['result']
-
+            logger.info(f"Resulting Neo4j query result: {result}")
+            return result.get('result')
         except Exception as e:
             logger.error(f"Graph query failed: {e}")
             return None

tests/fake_telegram_update.json

@@ -0,0 +1,21 @@
+{
+  "update_id": 100000000,
+  "message": {
+    "message_id": 1,
+    "from": {
+      "id": 123456789,
+      "is_bot": false,
+      "first_name": "Tester",
+      "username": "tester_user"
+    },
+    "chat": {
+      "id": 123456789,
+      "first_name": "Tester",
+      "username": "tester_user",
+      "type": "private"
+    },
+    "date": 1700000000,
+    "text": "Hola desde local!"
+  }
+}
+

tests/test_telegram_webhook.sh

@@ -0,0 +1,59 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# === Configuración ===
+BOT_NAME="PolisplexityBot"
+HOST="localhost"
+PORT="8011"    # Puerto donde corre tu servidor Django (devserver o Docker mapped)
+WEBHOOK_URL="http://${HOST}:${PORT}/bots/webhook/${BOT_NAME}/"
+PAYLOAD_FILE="$(dirname "$0")/fake_telegram_update.json"
+
+# === Ejecución ===
+echo "=== Test Telegram Webhook ==="
+echo "Bot:      $BOT_NAME"
+echo "URL:      $WEBHOOK_URL"
+echo "Payload:  $PAYLOAD_FILE"
+echo
+
+echo "1) Mostrando contenido del payload:"
+cat "$PAYLOAD_FILE"
+echo
+
+echo "2) Enviando payload al webhook (verbose):"
+# Archivo temporal para capturar body
+BODY_FILE=$(mktemp)
+# Enviamos con curl -v para depuración, guardamos respuesta en BODY_FILE y código en STATUS_CODE
+STATUS_CODE=$(curl -v \
+  -X POST \
+  -H "Content-Type: application/json" \
+  --data @"$PAYLOAD_FILE" \
+  -o "$BODY_FILE" \
+  -w "%{http_code}" \
+  "$WEBHOOK_URL" || true)
+
+echo
+ echo "→ HTTP Status: $STATUS_CODE"
+ echo "→ Response Body:"
+ cat "$BODY_FILE"
+ echo
+
+if [[ "$STATUS_CODE" != "200" ]]; then
+  echo "❌ ERROR: el webhook devolvió código $STATUS_CODE"
+  exit 1
+fi
+
+echo "3) Verificando en base de datos dentro del contenedor Docker..."
+# Usamos docker exec para acceder al entorno con dependencias instaladas
+docker exec -i polisplexity_django python manage.py shell << 'EOF'
+from pxy_bots.models import TelegramConversation, TelegramMessage
+
+print("Conversations:", TelegramConversation.objects.count())
+print("Messages:    ", TelegramMessage.objects.count())
+for m in TelegramMessage.objects.all():
+    print(f"  • [{m.direction:3}] (conv={m.conversation_id}) rt={m.response_time_ms or 0} ms ⇒ {m.content[:50]!r}")
+EOF
+
+# Limpieza temporal
+rm -f "$BODY_FILE"
+
+echo "✅ Test completado"