update endpoint

.gitignore

@@ -2,4 +2,6 @@ __pycache__/
 .devcontainer/
 .streamlit/
 .env
-private_data/
+private_data/
+app_script.py
+test.py

app.py

@@ -292,7 +292,7 @@ if submit_button_placeholder.button("AI Answer", type="primary"):
                 ]
 
                 raw_synthesis_answer = get_inference_endpoint_response(
-                    model="deepseek-ai/DeepSeek-R1-0528-Qwen3-8B",
+                    model="tgi",#"deepseek-ai/DeepSeek-R1-0528-Qwen3-8B",
                     messages=messages,
                     temperature=synthesis_deepseek_temperature,
                     top_p=synthesis_deepseek_top_p,

utils/endpoint_utils.py

@@ -5,6 +5,7 @@ import re
 import json
 
 from huggingface_hub import InferenceClient
+from openai import OpenAI
 
 # Helper function to parse the response
 def parse_thinking_response(response_text):
@@ -31,24 +32,23 @@ def get_inference_endpoint_response(
     """
     Serverless API (Pay-as-you-go)
     """
-    client = InferenceClient(
-        provider="auto",
-        api_key=os.getenv("HF_API_KEY")
+    client = OpenAI(
+        base_url = "https://w98rfd84portge8q.us-east-1.aws.endpoints.huggingface.cloud/v1/",
+        api_key =  os.getenv("HF_ENDPOINT_TOKEN")
     )
 
-    completion = client.chat.completions.create(
+    # print(messages)
+    chat_completion = client.chat.completions.create(
         model=model,
         messages=messages,
         temperature=temperature,
         top_p=top_p,
-        max_tokens=max_tokens
+        max_tokens=max_tokens,
+        stream=False
     )
 
     # Get the raw response content
-    raw_response = completion.choices[0].message.content
-
-    return raw_response
-
+    return chat_completion.choices[0].message.content
 
 def get_custom_inference_endpoint_response(
     messages: list,