add config file + HF provider support

app.py

@@ -6,11 +6,12 @@ import pandas as pd
 from llama_index.core import PromptTemplate
 from llama_index.core.workflow import Context
 from llama_index.core.agent.workflow import ReActAgent, AgentStream, ToolCallResult
-from llama_index.llms.huggingface_api import HuggingFaceInferenceAPI
+from llama_index.llms.huggingface_api import HuggingFaceInferenceAPI # customized to support different provider
 from llama_index.tools.wikipedia import WikipediaToolSpec
 from llama_index.tools.duckduckgo import DuckDuckGoSearchToolSpec
 
 from prompt import custom_react_system_header_str
+from config import HF_MODEL_NAME, HF_PROVIDER
 
 # (Keep Constants as is)
 # --- Constants ---
@@ -30,16 +31,18 @@ class BasicAgent:
 class LLamaIndexAgent:
     def __init__(self,
                  model_name="Qwen/Qwen2.5-Coder-32B-Instruct",
+                 provider="hf-inference",
                  show_tools_desc=True,
                  show_prompt=True):
 
         # LLM definition
-        llm = HuggingFaceInferenceAPI(model_name=model_name)
+        llm = HuggingFaceInferenceAPI(model_name=model_name,
+                                      provider=provider)
         print(f"LLamaIndexAgent initialized with model \"{model_name}\"")
 
         # tools definition
         tool_spec_list = []
-        tool_spec_list += WikipediaToolSpec().to_tool_list()
+        # tool_spec_list += WikipediaToolSpec().to_tool_list()
         tool_spec_list += DuckDuckGoSearchToolSpec().to_tool_list()
 
         # agent definition
@@ -63,7 +66,6 @@ class LLamaIndexAgent:
                 print("\n" + "="*30 + f" Prompt: {k} " + "="*30)
                 print(v.template)
 
-
     async def __call__(self, question: str) -> str:
         print(f"Agent received question (first 50 chars): {question[:50]}...")
 
@@ -101,7 +103,8 @@ async def run_and_submit_all(profile: gr.OAuthProfile | None):
     # 1. Instantiate Agent (modify this part to create your agent)
     try:
         # agent = BasicAgent()
-        agent = LLamaIndexAgent()
+        agent = LLamaIndexAgent(model_name=HF_MODEL_NAME,
+                                provider=HF_PROVIDER)
     except Exception as e:
         print(f"Error instantiating agent: {e}")
         return f"Error initializing agent: {e}", None
@@ -143,8 +146,10 @@ async def run_and_submit_all(profile: gr.OAuthProfile | None):
         try:
             # submitted_answer = agent(question_text)
             submitted_answer = await agent(question_text)
+            submitted_answer = str(submitted_answer) # cast AgentOutput to str
             answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
             results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
+            agent.ctx.clear() # clear context for next question
         except Exception as e:
              print(f"Error running agent on task {task_id}: {e}")
              results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"})

config.py

@@ -0,0 +1,2 @@
+HF_MODEL_NAME = "google/gemma-3-27b-it"
+HF_PROVIDER = "nebius"

requirements.txt

@@ -1,6 +1,6 @@
 gradio
 requests
 llama-index
-llama-index-llms-huggingface-api
+llama-index-llms-huggingface-api @ git+https://github.com/guillaumefrd/llama_index.git@add-provider-HF-API#subdirectory=llama-index-integrations/llms"
 llama_index.tools.wikipedia
 llama_index.tools.duckduckgo