Improved Documentation using expanders and streamlit text formatting features.

app.py

@@ -7,33 +7,46 @@ 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.
-
-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.
-"""
+st.markdown("## 🛠️ Response Specification Features")
+st.markdown("**The expanders below are parameters that you can adjust to customize the AI response.**")
+with st.expander("📌 **Model Selection**"):
+    st.write("Choose the AI model to generate responses.")
+
+with st.expander("🎨 **Temperature (Creativity Control)**"):
+    st.write("""
+    - **0.0**: Always the same response (deterministic).  
+    - **0.1 - 0.3**: Mostly factual and repetitive.  
+    - **0.4 - 0.7**: Balanced between coherence and creativity.  
+    - **0.8 - 1.0**: Highly creative but less predictable.
+    """)
+
+with st.expander("📏 **Max Tokens (Response Length)**"):
+    st.write("Defines the maximum number of words/subwords in the response.")
+
+with st.expander("🎯 **Top-p (Nucleus Sampling)**"):
+    st.write("""
+    Controls word diversity by sampling from top-probability tokens:
+    - **High `top_p` + Low `temperature`** → More factual, structured responses.
+    - **High `top_p` + High `temperature`** → More diverse, unexpected responses.
+    """)
+
+with st.expander("🔄 **Number of Responses**"):
+    st.write("Specifies how many response variations the AI should generate.")
+
+with st.expander("✅ **Fact-Checking**"):
+    st.write("""
+    - If **enabled**, AI prioritizes factual accuracy.  
+    - If **disabled**, AI prioritizes creativity.
+    """)
+
+st.markdown("""
+### 🔎 **Summary**  
+- `temperature` → Adjusts **creativity vs accuracy**.  
+- `max_tokens` → Defines **response length**.  
+- `top_p` → Fine-tunes **word diversity**.  
+- `fact_check` → Ensures **factual correctness** (but may reduce 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 = []