Changes in app.py

app.py

@@ -2,6 +2,8 @@ import gradio as gr
 from collinear import Collinear
 import os
 import json
+from openai import AsyncOpenAI
+from jinja2 import Template
 collinear = Collinear(access_token=os.getenv('COLLINEAR_API_KEY'))
 
 def update_inputs(input_style):
@@ -13,6 +15,53 @@ def update_inputs(input_style):
         return gr.update(visible=True), gr.update(visible=False), gr.update(visible=False), gr.update(visible=True), gr.update(visible=True)
 
 
+async def lynx(input_style_dropdown,document_input,question_input,answer_input):
+    if input_style_dropdown=='QA format':
+        client = AsyncOpenAI(
+        base_url="https://s6mipt5j797e6fql.us-east-1.aws.endpoints.huggingface.cloud/v1/", 
+        api_key=os.getenv("HF_TOKEN") 
+        )
+        PROMPT = Template("""
+Given the following QUESTION, DOCUMENT and ANSWER you must analyze the provided answer and determine whether it is faithful to the contents of the DOCUMENT. The ANSWER must not offer new information beyond the context provided in the DOCUMENT. The ANSWER also must not contradict information provided in the DOCUMENT. Output your final verdict by strictly following this format: "PASS" if the answer is faithful to the DOCUMENT and "FAIL" if the answer is not faithful to the DOCUMENT. Show your reasoning.
+
+--
+QUESTION (THIS DOES NOT COUNT AS BACKGROUND INFORMATION):
+{question}
+
+--
+DOCUMENT:
+{context}
+
+--
+ANSWER:
+{answer}
+
+--
+
+Your output should be in JSON FORMAT with the keys "REASONING" and "SCORE":
+{{"REASONING": <your reasoning as bullet points>, "SCORE": <your final score>}}
+""")
+        chat_completion = await client.chat.completions.create(
+            model="tgi",
+            messages=[
+            {
+                "role": "user",
+                "content": PROMPT.render(question=question_input,context=document_input,answer=answer_input)
+            }
+        ],
+            top_p=None,
+            temperature=None,
+            max_tokens=150,
+            stream=False,
+            seed=None,
+            frequency_penalty=None,
+            presence_penalty=None
+        )
+        print(chat_completion)
+        return chat_completion.choices.pop().message.content
+    else:
+        return 'NA'
+
 # Function to judge reliability based on the selected input format
 async def judge_reliability(input_style, document, conversation, claim, question, answer):
     if input_style == "Conv":
@@ -41,7 +90,9 @@ with gr.Blocks() as demo:
         answer_input = gr.Textbox(label="Answer", lines=5, visible=False, value="CV was born in Iowa")
 
     with gr.Row():
-        result_output = gr.Textbox(label="Results")
+        result_output = gr.Textbox(label="Veritas Model")
+
+        lynx_output = gr.Textbox(label="Lynx Model")
 
 
     # Set the visibility of inputs based on the selected input style
@@ -56,6 +107,10 @@ with gr.Blocks() as demo:
         fn=judge_reliability, 
         inputs=[input_style_dropdown, document_input, conversation_input, claim_input, question_input, answer_input], 
         outputs=result_output
+    ).then(
+        fn=lynx,
+        inputs=[input_style_dropdown,document_input,question_input,answer_input],
+        outputs=lynx_output
     )
 
 # Launch the demo