Added Prompt Specification Features

app.py

@@ -2,45 +2,100 @@ import streamlit as st
 from openai import OpenAI
 import os
 
-# Set your API key here
+# Set up NVIDIA API client
 client = OpenAI(
     base_url="https://integrate.api.nvidia.com/v1",
     api_key=os.environ.get("NVIDIA_API_KEY")
 )
+"""
+Parameters for Response Specification Features:
+- model: The AI model to use for generating responses.
+- temperature: Controls the randomness of the response. Higher values result in more randomness.
+    Example Use Cases for this one:
+    - 0.0: Always the same response
+    - 0.1 - 0.3: Mostly Deterministic, Factual and repetitive siya.
+    - 0.4 - 0.7: Balanced between coherence and creative responses ni.
+    - 0.8 - 1.0: More creative and imaginative responses[less coherent].
+- max_tokens: The maximum number of tokens(words/subwords) to generate in the response.
+- top_p: Controls the probability of sampling from the top tokens. Higher values result in more creativity. [This is related to the temperature parameter]
+    -This is also known as nucleus sampling, determining the probability of nexty words the AI will consider
+        The higher the value, the more diverse the response will be.
+        For example bala:
+        top_p + low temp = more accurate and factual responses
+        top_p + high temp = more creative responses, unexpected responses siya bih.
+- num_responses: The number of responses to generate.
+- fact_check: If True, the AI will check the factual accuracy of the response.
+    If False, the AI will prioritize creativity over factual accuracy.
 
-def query_ai_model(prompt):
+IN SUMMARY:
+- temperature controls creativity vs accuracy.
+- max_tokens affects length.
+- top_p fine-tunes word diversity.
+- fact_check ensures factual correctness (but slightly limits fluency).
+- num_responses generates different variations of the same prompt.
+"""
+
+def query_ai_model(prompt, model="meta/llama-3.1-405b-instruct", temperature=0.7, max_tokens=512, top_p=0.9, fact_check=False, num_responses=1):
+    responses = []
+    
     try:
-        # Example of using NVIDIA's OpenAI integration
-        completion = client.chat.completions.create(
-            model="meta/llama-3.1-405b-instruct",
-            messages=[{"role": "user", "content": prompt}],
-            temperature=0.2,
-            top_p=0.7,
-            max_tokens=1024,
-            stream=True
-        )
-        response = ""
-        response_container = st.empty()
-
-        for chunk in completion:
-            if chunk.choices[0].delta.content is not None:
-                response += chunk.choices[0].delta.content
-                response_container.markdown(response)
+        if fact_check:
+            prompt = "Ensure factual accuracy. " + prompt
+
+        for _ in range(num_responses):  # Response Loop para sa number of responses
+            completion = client.chat.completions.create(
+                model=model,
+                messages=[{"role": "user", "content": prompt}],
+                temperature=temperature,
+                top_p=top_p,
+                max_tokens=max_tokens
+            )
+            response = completion.choices[0].message.content
+            responses.append(response)
+
     except Exception as e:
-        return f"An error occurred: {str(e)}"
+        st.error(f"An error occurred: {str(e)}")
+    
+    return responses  # Return a list of responses
 
-# Streamlit App
+# Simple streamlit UI palang this
 st.title("Mark's AI Chatbot")
-st.write("Enter a prompt below and click 'Ask AI' to get a response from the AI model.")
+st.write("Provide a topic and customize the response criteria.")
 
-# Input prompt
+# Input Fields
 user_input = st.text_area("Your Prompt:", placeholder="Type something...")
 
+# Dropdown Menus
+output_format = st.selectbox("Select Output Format:", ["Story", "Poem", "Article", "Code"])
+tone_style = st.selectbox("Select Tone/Style:", ["Formal", "Informal", "Humorous", "Technical"])
+
+# Sliders
+creativity_level = st.slider("Creativity Level:", min_value=0.0, max_value=1.0, value=0.7, step=0.1)
+max_length = st.slider("Max Length (tokens):", min_value=100, max_value=1024, value=512, step=50)
+
+#Numeric Inputs
+num_responses = st.number_input("Number of Responses:", min_value=1, max_value=5, value=1, step=1)
+
+# Checkboxes
+enable_creativity = st.checkbox("Enable Creative Mode", value=True)
+fact_checking = st.checkbox("Enable Fact-Checking")
+
 if st.button("Generate Answer"):
     if user_input.strip():
-        with st.spinner("Getting a response from the AI model..."):
-            ai_response = query_ai_model(user_input)
-        st.success("AI Response:")
-        st.write(ai_response)
+        with st.spinner("Generating response..."):
+            full_prompt = f"Format: {output_format}\nTone: {tone_style}\nPrompt: {user_input}"
+            ai_responses = query_ai_model(
+                full_prompt,
+                temperature=creativity_level if enable_creativity else 0.2,
+                max_tokens=max_length,
+                top_p=0.9 if enable_creativity else 0.7,
+                fact_check=fact_checking,
+                num_responses=num_responses
+            )
+
+        st.success("AI Responses:")
+        for i, response in enumerate(ai_responses, 1):
+            st.markdown(f"### Response {i}")
+            st.write(response)
     else:
-        st.warning("Please enter a prompt before clicking the Button.")
+        st.warning("Please enter a prompt before clicking the button.")