Update app.py

app.py

@@ -1,20 +1,120 @@
 import streamlit as st
+from openai import OpenAI
+import os
 import pandas as pd
 import numpy as np
-from utils.constants import metadata_path, embeddings_path
-from question_handler import find_top_question, generate_detailed_prompt, extract_code_and_test_case
-from utils.openai_client import generate_response
-from code_executor import execute_code
+from sentence_transformers import SentenceTransformer
+from sklearn.metrics.pairwise import cosine_similarity
+import torch
+import re
+
+# Set up OpenAI client
+client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
+
+# Check if GPU is available
+device = "cuda" if torch.cuda.is_available() else "cpu"
+print(f"Using device: {device}")
+
+# Load metadata and embeddings (ensure these files are in your working directory or update paths)
+metadata_path = 'question_metadata.csv'  # Update this path if needed
+embeddings_path = 'question_dataset_embeddings.npy'  # Update this path if needed
 
-# Load metadata and embeddings once at app startup
 metadata = pd.read_csv(metadata_path)
 embeddings = np.load(embeddings_path)
 
+# Load the SentenceTransformer model
+model = SentenceTransformer("all-MiniLM-L6-v2").to(device)
+
+# Load prompts from files
+with open("technical_interviewer_prompt.txt", "r") as file:
+    technical_interviewer_prompt = file.read()
+
+with open("question_generation_prompt.txt", "r") as file:
+    question_generation_prompt = file.read()
+    
+st.title("Mock Interview: Real-World Programming")
+
+# Initialize session state variables
+if "messages" not in st.session_state:
+    st.session_state.messages = []
 
-# Streamlit UI components
-st.title("Real-World Programming Question Mock Interview")
+if "follow_up_mode" not in st.session_state:
+    st.session_state.follow_up_mode = False  # Tracks whether we're in follow-up mode
+
+if "generated_question" not in st.session_state:
+    st.session_state.generated_question = None  # Stores the generated question for persistence
+
+if "code_template" not in st.session_state:
+    st.session_state.code_template = ""  # Stores the code template
+
+if "sample_test_case" not in st.session_state:
+    st.session_state.sample_test_case = ""  # Stores the sample test case
+
+if "expected_output" not in st.session_state:
+    st.session_state.expected_output = ""  # Stores the expected output
+
+if "debug_logs" not in st.session_state:
+    st.session_state.debug_logs = None  # Stores debug logs for toggling
+
+# Function to find the top 1 most similar question based on user input
+def find_top_question(query):
+    # Generate embedding for the query
+    query_embedding = model.encode(query, convert_to_tensor=True, device=device).cpu().numpy()
+    
+    # Reshape query_embedding to ensure it is a 2D array
+    query_embedding = query_embedding.reshape(1, -1)  # Reshape to (1, n_features)
 
-# Sidebar form for generating questions
+    # Compute cosine similarity between query embedding and dataset embeddings
+    similarities = cosine_similarity(query_embedding, embeddings).flatten()  # Flatten to get a 1D array of similarities
+
+    # Get the index of the most similar result (top 1)
+    top_index = similarities.argsort()[-1]  # Index of highest similarity
+
+    # Retrieve metadata for the top result
+    top_result = metadata.iloc[top_index].copy()
+    top_result['similarity_score'] = similarities[top_index]
+
+    return top_result
+
+# Function to generate response using OpenAI API with debugging logs
+def generate_response(messages):
+    # For debug logs, store only the follow-up conversation history
+    st.session_state.debug_logs = st.session_state.messages  # Update debug logs with current conversation
+    
+    response = client.chat.completions.create(
+        model="o1-mini",
+        messages=messages,
+    )
+    
+    return response.choices[0].message.content
+
+# Function to extract code template and sample test case from the generated question
+def extract_code_and_test_case(generated_question):
+    code_template = ""
+    sample_test_case = ""
+    expected_output = ""
+    
+    # Extract code template
+    code_match = re.search(r'```python(.*?)```', generated_question, re.DOTALL)
+    if code_match:
+        code_template = code_match.group(1).strip()
+    else:
+        # Default code template if none is found
+        code_template = "# Write your code here\n"
+
+    # Extract sample test case and expected output
+    test_case_match = re.search(r'Sample Input:\s*(.*?)\n', generated_question, re.DOTALL)
+    expected_output_match = re.search(r'Expected Output:\s*(.*?)\n', generated_question, re.DOTALL)
+    if test_case_match and expected_output_match:
+        sample_test_case = test_case_match.group(1).strip()
+        expected_output = expected_output_match.group(1).strip()
+    else:
+        sample_test_case = ""
+        expected_output = ""
+
+    return code_template, sample_test_case, expected_output
+
+# Move the input form to the sidebar to make it always visible and more compact
 with st.sidebar.form(key="input_form"):
     st.markdown("## Generate a New Question")
     company = st.text_input("Company", value="Google")  # Default value: Google
@@ -29,14 +129,26 @@ if generate_button:
 
     # Create a query from user inputs and find the most relevant question
     query = f"{company} {difficulty} {topic}"
-    top_question = find_top_question(query, metadata, embeddings)
+    top_question = find_top_question(query)
     
     # Prepare a detailed prompt for GPT using the top question's details
-    detailed_prompt = generate_detailed_prompt(top_question)
+    detailed_prompt = (
+        f"Transform this LeetCode question into a real-world interview scenario.\n\n"
+        f"**Company**: {top_question['company']}\n"
+        f"**Question Name**: {top_question['questionName']}\n"
+        f"**Difficulty Level**: {top_question['difficulty level']}\n"
+        f"**Tags**: {top_question['Tags']}\n"
+        f"**Content**: {top_question['Content']}\n"
+        f"\nPlease create a real-world interview question based on this information. "
+        f"Include the following sections:\n\n"
+        f"- Problem Description\n"
+        f"- Code Template (in a Python code block)\n"
+        f"- Sample Input and Expected Output (clearly separated)\n"
+    )
     
     # Generate response using OpenAI API with detailed prompt and debugging logs
-    response = generate_response([{"role": "user", "content": detailed_prompt}])
-    
+    response = generate_response([{"role": "user", "content": detailed_prompt}])  # Question generation prompt excluded here
+
     # Store generated question in session state for persistence in sidebar and follow-up conversation state
     st.session_state.generated_question = response
 
@@ -64,14 +176,17 @@ if st.session_state.follow_up_mode:
         st.session_state.messages.append({"role": "user", "content": user_input})
 
         # Prepare messages to send to the assistant
+        # Include the technical interviewer prompt and generated question, but do not display them
+        # Add an instruction for the assistant to reply as a real-world interviewer would
         assistant_instruction = (
             "As a real-world interviewer, please reply to the candidate's follow-up questions "
             "specific to the generated interview question, to the point, and in a natural, human-sounding way."
         )
 
         messages_to_send = [
-            {"role": "user", "content": assistant_instruction},
-            {"role": "assistant", "content": st.session_state.generated_question}
+            {"role": "user", "content": technical_interviewer_prompt},
+            {"role": "assistant", "content": st.session_state.generated_question},
+            {"role": "user", "content": assistant_instruction}
         ] + st.session_state.messages
 
         assistant_response = generate_response(messages_to_send)
@@ -98,9 +213,41 @@ else:
     code_input = st.sidebar.text_area("Write your Python code here:", height=300)
 
 if st.sidebar.button("Run Code"):
-    execute_code(code_input)
-
-# Right sidebar toggleable debug logs section
+    try:
+        # Prepare the code for execution
+        exec_globals = {}
+        # Create a function wrapper to execute the user's code
+        exec(f"def user_solution():\n{code_input}", exec_globals)
+        user_solution = exec_globals.get('user_solution', None)
+        
+        # Prepare sample test case execution
+        if st.session_state.sample_test_case:
+            # Assume the sample test case is in the format of arguments to the function
+            test_case = st.session_state.sample_test_case
+            # Evaluate the test case safely
+            test_args = eval(test_case)
+            if not isinstance(test_args, tuple):
+                test_args = (test_args,)
+            # Capture the output
+            returned_output = user_solution(*test_args)
+        else:
+            returned_output = user_solution()
+        
+        # Display the expected output and returned output
+        st.sidebar.markdown("### Sample Test Case Result:")
+        st.sidebar.markdown(f"**Sample Input:** {st.session_state.sample_test_case}")
+        st.sidebar.markdown(f"**Expected Output:** {st.session_state.expected_output}")
+        st.sidebar.markdown(f"**Your Output:** {returned_output}")
+        
+        # Compare outputs
+        if str(returned_output) == st.session_state.expected_output:
+            st.sidebar.success("Your output matches the expected output!")
+        else:
+            st.sidebar.error("Your output does not match the expected output.")
+    except Exception as e:
+        st.sidebar.error(f"Error: {e}")
+
+# Right sidebar toggleable debug logs and code interpreter section
 with st.expander("Debug Logs (Toggle On/Off)", expanded=False):
-    if "debug_logs" in st.session_state:
-        st.write(st.session_state.debug_logs)
+    if st.session_state.debug_logs:
+        st.write(st.session_state.debug_logs)