Upload folder using huggingface_hub

pyproject.toml

@@ -19,6 +19,7 @@ dependencies = [
     "langchain_google_genai>=3.0.3",
     "langchain_groq>=1.0.1",
 
+
     # =======================
     # LANGCHAIN / LANGGRAPH
     # =======================
@@ -30,6 +31,7 @@ dependencies = [
     "langchain-text-splitters>=1.0.0",
     "langchain-chroma>=1.0.0",
     "html2text>=2025.4.15",
+    "langfuse>=2.0.0",
 
     # =======================
     # VECTOR DB / INDEXING
@@ -119,10 +121,7 @@ dependencies = [
     # =======================
     # OBSERVABILITY
     # =======================
-    "opentelemetry-api",
-    "opentelemetry-sdk",
-    "opentelemetry-exporter-otlp",
-    "opentelemetry-instrumentation-openai",
+
 ]
 
 [dependency-groups]

src/chatbot/aagents/orchestrator_agent.py

@@ -8,6 +8,8 @@ from aagents.input_validation_agent import input_validation_guardrail
 from agents import Agent, OpenAIChatCompletionsModel, Runner, function_tool
 from openai import AsyncOpenAI
 from dotenv import load_dotenv
+from langsmith import wrappers
+
 
 load_dotenv()
 
@@ -15,6 +17,7 @@ load_dotenv()
 GEMINI_BASE_URL = "https://generativelanguage.googleapis.com/v1beta/openai/"
 google_api_key = os.getenv("GOOGLE_API_KEY")
 gemini_client = AsyncOpenAI(base_url=GEMINI_BASE_URL, api_key=google_api_key)
+gemini_client = wrappers.wrap_openai(gemini_client)
 gemini_model = OpenAIChatCompletionsModel(
     model="gemini-2.0-flash",
     openai_client=gemini_client

src/chatbot/app.py

@@ -2,10 +2,24 @@ import os
 import glob
 import uuid
 import asyncio
+import langfuse
+import logging
 import streamlit as st
 from aagents.orchestrator_agent import orchestrator_agent
 from agents import Runner, trace, SQLiteSession
 from agents.exceptions import InputGuardrailTripwireTriggered
+from langsmith import traceable
+
+# Make Langfuse optional to avoid "Client will be disabled" errors
+if os.environ.get("LANGFUSE_PUBLIC_KEY"):
+    from langfuse import observe
+else:
+    # Dummy decorator if keys are missing
+    def observe(*args, **kwargs):
+        def decorator(func):
+            return func
+        return decorator
+
 
 # -----------------------------
 # Configuration & Utils
@@ -16,63 +30,14 @@ st.set_page_config(
     page_icon="🤖"
 )
 
-# ------------------------------------------------------------------------------
-# OpenTelemetry Setup
-# ------------------------------------------------------------------------------
-from opentelemetry import trace as trace_api
-# from opentelemetry.exporter.otlp.proto.grpc.trace_exporter import OTLPSpanExporter
-from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
-from opentelemetry.instrumentation.openai import OpenAIInstrumentor
-from opentelemetry.sdk.resources import Resource
-from opentelemetry.sdk.trace import TracerProvider
-from opentelemetry.sdk.trace.export import BatchSpanProcessor
-
-otel_endpoint = os.getenv("OTEL_EXPORTER_OTLP_ENDPOINT")
-
-if otel_endpoint:
-    enable_otel = True
-else:
-    # Local development default
-    otel_endpoint = "https://myotel.azurewebsites.net/v1/traces"
-    enable_otel = True
-
-if enable_otel:
-    try:
-        # Monkeypatch the internals of the OpenAI instrumentation to fix "Invalid type Omit" error
-        # This is necessary because the official instrumentation library doesn't handle 'Omit' types correctly
-        # and blindly tries to set them as attributes, causing console errors.
-        import opentelemetry.instrumentation.openai.shared
-        original_set_span_attribute = opentelemetry.instrumentation.openai.shared._set_span_attribute
-    
-        def patched_set_span_attribute(span, name, value):
-            # Check for Omit/NotGiven types by name to avoid importing internal types
-            if value is not None and type(value).__name__ in ["Omit", "NotGiven"]:
-                return
-            original_set_span_attribute(span, name, value)
-            
-        opentelemetry.instrumentation.openai.shared._set_span_attribute = patched_set_span_attribute
-
-        # Set up telemetry span exporter.
-        # otel_exporter = OTLPSpanExporter(endpoint=otel_endpoint, insecure=True)
-        otel_exporter = OTLPSpanExporter(endpoint=otel_endpoint)
-        span_processor = BatchSpanProcessor(otel_exporter)
-
-        # Set up telemetry trace provider.
-        tracer_provider = TracerProvider(resource=Resource({"service.name": "chatbot"}))
-        tracer_provider.add_span_processor(span_processor)
-        trace_api.set_tracer_provider(tracer_provider)
-
-        # Instrument the OpenAI Python library
-        OpenAIInstrumentor().instrument()
-        print(f"OpenTelemetry enabled with endpoint: {otel_endpoint}")
-    except Exception as e:
-        print(f"Failed to initialize OpenTelemetry: {e}")
-else:
-    print("OpenTelemetry disabled (Running in HF Space with no configured endpoint).")
-
-# Get a tracer (works even if OTEL is disabled, returning a NoOp tracer)
-tracer = trace_api.get_tracer("chatbot")
+# Load environment variables explicitly
+from dotenv import load_dotenv
+load_dotenv(override=True)
 
+# Configure debug logging for Langfuse only if enabled
+if os.environ.get("LANGFUSE_PUBLIC_KEY"):
+    logging.basicConfig()
+    logging.getLogger("langfuse").setLevel(logging.INFO)
 
 def load_prompts(folder="prompts"):
     prompts = []
@@ -83,7 +48,8 @@ def load_prompts(folder="prompts"):
                 content = f.read().strip()
                 if content:
                     prompts.append(content)
-                    prompt_labels.append(os.path.basename(file_path).replace("_", " ").replace(".txt", "").title())
+
+            prompt_labels.append(os.path.basename(file_path).replace("_", " ").replace(".txt", "").title())
     return prompts, prompt_labels
 
 prompts, prompt_labels = load_prompts()
@@ -256,19 +222,18 @@ st.markdown("""
 # -----------------------------
 # Logic
 # -----------------------------
+@observe()
+@traceable(name="Chatbot Interaction")
 async def get_ai_response(prompt: str) -> str:
     try:
         agent = orchestrator_agent
         # Ensure session is valid
         current_session = st.session_state.ai_session
         current_session = st.session_state.ai_session
-        with trace("Chatbot Agent Run"): # Keep existing custom trace wrapper if it exists, or just use new tracer
-             with tracer.start_as_current_span("get_ai_response") as span:
-                span.set_attribute("input.prompt", prompt)
-                # Run agent
-                result = await Runner.run(agent, prompt, session=current_session)
-                span.set_attribute("output.response", result.final_output)
-                return result.final_output
+        with trace("Chatbot Agent Run"): # Keep existing custom trace wrapper
+            # Run agent
+            result = await Runner.run(agent, prompt, session=current_session)
+            return result.final_output
     except InputGuardrailTripwireTriggered as e:
         reasoning = getattr(e, "reasoning", None) \
             or getattr(getattr(e, "output", None), "reasoning", None) \
@@ -333,6 +298,13 @@ if prompt := (st.chat_input("Type your message...") or selected_prompt):
         with st.spinner("Thinking..."):
             response_text = asyncio.run(get_ai_response(prompt))
             st.markdown(response_text)
+
+            # Ensure traces are sent before the script may stop/rerun
+            if os.environ.get("LANGFUSE_PUBLIC_KEY"):
+                try:
+                    langfuse.Langfuse().flush()
+                except:
+                    pass
             
     st.session_state.messages.append({"role": "assistant", "content": response_text})
     

src/chatbot/test_trace.py

@@ -0,0 +1,47 @@
+import os
+import time
+from dotenv import load_dotenv
+import logging
+
+# Load envs
+load_dotenv(override=True)
+
+# Enable debug logs
+import sys
+logging.basicConfig(stream=sys.stdout, level=logging.DEBUG)
+logging.getLogger("langfuse").setLevel(logging.DEBUG)
+
+
+import langfuse
+from langfuse import observe
+
+# 1. Print config to verify keys
+print(f"Host: {os.environ.get('LANGFUSE_HOST')}")
+print(f"Public Key: {os.environ.get('LANGFUSE_PUBLIC_KEY')}")
+
+# 2. Define observed function
+@observe(name="test-trace-script")
+def run_test():
+    print("Executing observed function...")
+    time.sleep(0.1)
+    return "Test successful"
+
+# 3. Run
+run_test()
+
+# 4. Flush / Wait
+print("Waiting for background upload...")
+time.sleep(3)
+
+try:
+    from langfuse import Langfuse
+    # Try to flush using a new client instance (hoping for shared state or just to test connection)
+    client = Langfuse()
+    client.flush()
+    print("Flush called on client instance.")
+except Exception as e:
+    print(f"Flush failed: {e}")
+
+print("Script finished.")
+
+

uv.lock

@@ -35,6 +35,7 @@ dependencies = [
     { name = "langchain-ollama" },
     { name = "langchain-openai" },
     { name = "langchain-text-splitters" },
+    { name = "langfuse" },
     { name = "langgraph" },
     { name = "langgraph-checkpoint-sqlite" },
     { name = "langsmith" },
@@ -45,10 +46,6 @@ dependencies = [
     { name = "openai" },
     { name = "openai-agents" },
     { name = "openai-whisper" },
-    { name = "opentelemetry-api" },
-    { name = "opentelemetry-exporter-otlp" },
-    { name = "opentelemetry-instrumentation-openai" },
-    { name = "opentelemetry-sdk" },
     { name = "playwright" },
     { name = "plotly" },
     { name = "polygon-api-client" },
@@ -108,6 +105,7 @@ requires-dist = [
     { name = "langchain-ollama", specifier = ">=1.0.0" },
     { name = "langchain-openai", specifier = ">=1.0.3" },
     { name = "langchain-text-splitters", specifier = ">=1.0.0" },
+    { name = "langfuse", specifier = ">=2.0.0" },
     { name = "langgraph", specifier = ">=1.0.3" },
     { name = "langgraph-checkpoint-sqlite", specifier = ">=3.0.0" },
     { name = "langsmith", specifier = ">=0.4.43" },
@@ -118,10 +116,6 @@ requires-dist = [
     { name = "openai", specifier = ">=2.8.1" },
     { name = "openai-agents", specifier = ">=0.5.1" },
     { name = "openai-whisper", specifier = ">=1.0.0" },
-    { name = "opentelemetry-api" },
-    { name = "opentelemetry-exporter-otlp" },
-    { name = "opentelemetry-instrumentation-openai" },
-    { name = "opentelemetry-sdk" },
     { name = "playwright", specifier = ">=1.51.0" },
     { name = "plotly", specifier = ">=6.5.0" },
     { name = "polygon-api-client", specifier = ">=1.16.3" },
@@ -1797,6 +1791,27 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/1e/97/d362353ab04f865af6f81d4d46e7aa428734aa032de0017934b771fc34b7/langchain_text_splitters-1.0.0-py3-none-any.whl", hash = "sha256:f00c8219d3468f2c5bd951b708b6a7dd9bc3c62d0cfb83124c377f7170f33b2e", size = 33851, upload-time = "2025-10-17T14:33:40.46Z" },
 ]
 
+[[package]]
+name = "langfuse"
+version = "3.10.6"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "backoff" },
+    { name = "httpx" },
+    { name = "openai" },
+    { name = "opentelemetry-api" },
+    { name = "opentelemetry-exporter-otlp-proto-http" },
+    { name = "opentelemetry-sdk" },
+    { name = "packaging" },
+    { name = "pydantic" },
+    { name = "requests" },
+    { name = "wrapt" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/e6/70/4ff19dd1085bb4d5007f008a696c8cf989a0ad76eabc512a5cd19ee4a0b7/langfuse-3.10.6.tar.gz", hash = "sha256:fced9ca0416ba7499afa45fbedf831afc0ec824cb283719b9cf429bf5713f205", size = 223656, upload-time = "2025-12-12T13:29:24.048Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/ce/f0/fac7d56ce1136afbbebaddd1dc119fb1b94b5a7489944d0b4c2dcee99ed7/langfuse-3.10.6-py3-none-any.whl", hash = "sha256:36ca490cd64e372b1b94c28063b3fea39b1a8446cabd20172b524d01011a34e1", size = 399347, upload-time = "2025-12-12T13:29:22.462Z" },
+]
+
 [[package]]
 name = "langgraph"
 version = "1.0.3"
@@ -2486,19 +2501,6 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/ae/a2/d86e01c28300bd41bab8f18afd613676e2bd63515417b77636fc1add426f/opentelemetry_api-1.38.0-py3-none-any.whl", hash = "sha256:2891b0197f47124454ab9f0cf58f3be33faca394457ac3e09daba13ff50aa582", size = 65947, upload-time = "2025-10-16T08:35:30.23Z" },
 ]
 
-[[package]]
-name = "opentelemetry-exporter-otlp"
-version = "1.38.0"
-source = { registry = "https://pypi.org/simple" }
-dependencies = [
-    { name = "opentelemetry-exporter-otlp-proto-grpc" },
-    { name = "opentelemetry-exporter-otlp-proto-http" },
-]
-sdist = { url = "https://files.pythonhosted.org/packages/c2/2d/16e3487ddde2dee702bd746dd41950a8789b846d22a1c7e64824aac5ebea/opentelemetry_exporter_otlp-1.38.0.tar.gz", hash = "sha256:2f55acdd475e4136117eff20fbf1b9488b1b0b665ab64407516e1ac06f9c3f9d", size = 6147, upload-time = "2025-10-16T08:35:52.53Z" }
-wheels = [
-    { url = "https://files.pythonhosted.org/packages/fd/8a/81cd252b16b7d95ec1147982b6af81c7932d23918b4c3b15372531242ddd/opentelemetry_exporter_otlp-1.38.0-py3-none-any.whl", hash = "sha256:bc6562cef229fac8887ed7109fc5abc52315f39d9c03fd487bb8b4ef8fbbc231", size = 7018, upload-time = "2025-10-16T08:35:32.995Z" },
-]
-
 [[package]]
 name = "opentelemetry-exporter-otlp-proto-common"
 version = "1.38.0"
@@ -2562,21 +2564,6 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/10/f5/7a40ff3f62bfe715dad2f633d7f1174ba1a7dd74254c15b2558b3401262a/opentelemetry_instrumentation-0.59b0-py3-none-any.whl", hash = "sha256:44082cc8fe56b0186e87ee8f7c17c327c4c2ce93bdbe86496e600985d74368ee", size = 33020, upload-time = "2025-10-16T08:38:31.463Z" },
 ]
 
-[[package]]
-name = "opentelemetry-instrumentation-openai"
-version = "0.49.8"
-source = { registry = "https://pypi.org/simple" }
-dependencies = [
-    { name = "opentelemetry-api" },
-    { name = "opentelemetry-instrumentation" },
-    { name = "opentelemetry-semantic-conventions" },
-    { name = "opentelemetry-semantic-conventions-ai" },
-]
-sdist = { url = "https://files.pythonhosted.org/packages/44/03/a04b74790ae3c5ea80aa257fae07698a9111ad1c58714ef78eb40f070414/opentelemetry_instrumentation_openai-0.49.8.tar.gz", hash = "sha256:2efe4efea59f2708ef3fc470a10d6db11eb7c48328a2729383d9adef89b6b2da", size = 32254, upload-time = "2025-12-11T20:32:53.415Z" }
-wheels = [
-    { url = "https://files.pythonhosted.org/packages/85/e7/36e0d15a1dfb94faf5fcc70721c6706ccbcf58323b31395b857884c0eb91/opentelemetry_instrumentation_openai-0.49.8-py3-none-any.whl", hash = "sha256:2555694d0f009b2d43776d718a7467229d49e04bb2ab78e2a9880d52674b8393", size = 43003, upload-time = "2025-12-11T20:32:20.844Z" },
-]
-
 [[package]]
 name = "opentelemetry-proto"
 version = "1.38.0"
@@ -2616,15 +2603,6 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/24/7d/c88d7b15ba8fe5c6b8f93be50fc11795e9fc05386c44afaf6b76fe191f9b/opentelemetry_semantic_conventions-0.59b0-py3-none-any.whl", hash = "sha256:35d3b8833ef97d614136e253c1da9342b4c3c083bbaf29ce31d572a1c3825eed", size = 207954, upload-time = "2025-10-16T08:35:48.054Z" },
 ]
 
-[[package]]
-name = "opentelemetry-semantic-conventions-ai"
-version = "0.4.13"
-source = { registry = "https://pypi.org/simple" }
-sdist = { url = "https://files.pythonhosted.org/packages/ba/e6/40b59eda51ac47009fb47afcdf37c6938594a0bd7f3b9fadcbc6058248e3/opentelemetry_semantic_conventions_ai-0.4.13.tar.gz", hash = "sha256:94efa9fb4ffac18c45f54a3a338ffeb7eedb7e1bb4d147786e77202e159f0036", size = 5368, upload-time = "2025-08-22T10:14:17.387Z" }
-wheels = [
-    { url = "https://files.pythonhosted.org/packages/35/b5/cf25da2218910f0d6cdf7f876a06bed118c4969eacaf60a887cbaef44f44/opentelemetry_semantic_conventions_ai-0.4.13-py3-none-any.whl", hash = "sha256:883a30a6bb5deaec0d646912b5f9f6dcbb9f6f72557b73d0f2560bf25d13e2d5", size = 6080, upload-time = "2025-08-22T10:14:16.477Z" },
-]
-
 [[package]]
 name = "orjson"
 version = "3.11.4"