import streamlit as st import json import os from Chunker import CodeChunker from utils import count_tokens # Function to load JSON data def load_json_file(file_path): with open(file_path, 'r') as file: return json.load(file) # Setup Streamlit page st.set_page_config(page_title="Cintra Code Chunker", layout="wide") # Assuming app.py and mock_codefiles.json are in the same directory json_file_path = os.path.join(os.path.dirname(__file__), 'mock_codefiles.json') code_files_data = load_json_file(json_file_path) # Extract filenames and contents code_files = list(code_files_data.keys()) # UI Elements st.title('Cintra Code Chunker') # File selection selected_file_name = st.selectbox("Select a code file", code_files) # Assuming you have the content as a string in the JSON, extract directly code_content = code_files_data[selected_file_name] file_extension = selected_file_name.split('.')[-1] # Determine the language for syntax highlighting def get_language_by_extension(file_extension): if file_extension in ['py', 'python']: return 'python' elif file_extension in ['js', 'jsx', 'javascript']: return 'javascript' elif file_extension == 'css': return 'css' else: return None # Default to no syntax highlighting if extension is not recognized language = get_language_by_extension(file_extension) col1, col2 = st.columns(2) with col1: st.subheader('Original File') st.code(code_content, language=language) with col2: token_chunk_size = st.sidebar.slider('Token Chunk Size Target', min_value=5, max_value=50, value=25) if st.sidebar.button("Chunk Code"): # Initialize the code chunker code_chunker = CodeChunker(file_extension=file_extension) # Chunk the code content chunked_code_dict = code_chunker.chunk(code_content, token_chunk_size) # Select a chunk to display chunk_keys = list(chunked_code_dict.keys()) selected_chunk_key = st.selectbox("Select Chunk", options=chunk_keys) st.subheader('Chunked Code') st.code(chunked_code_dict[selected_chunk_key], language=language)