app.py

# import os
# from langchain.text_splitter import RecursiveCharacterTextSplitter
# import gradio as gr

# # Ensure data directory exists
# os.makedirs("data", exist_ok=True)

# # Function to process uploaded files
# def process_files(italy_file, france_file):
#     docs = []
#     text_splitter = RecursiveCharacterTextSplitter(chunk_size=300, chunk_overlap=50)

#     # Save uploaded files
#     file_map = {
#         "data/italy.txt": italy_file,
#         "data/france.txt": france_file
#     }

#     for save_path, file_obj in file_map.items():
#         if file_obj is not None:
#             with open(save_path, "wb") as f:
#                 f.write(file_obj.read())

#     # Process files
#     for filename in file_map:
#         try:
#             with open(filename, "r", encoding="utf-8") as f:
#                 content = f.read()
#                 split_docs = text_splitter.create_documents([content])
#                 docs.extend(split_docs)
#         except FileNotFoundError:
#             return f"❌ File not found: {filename}", []
#         except Exception as e:
#             return f"⚠️ Error: {e}", []

#     output = [doc.page_content for doc in docs]
#     return f"✅ Processed {len(output)} chunks.", output[:10]  # Show first 10 chunks

# # Gradio Interface
# interface = gr.Interface(
#     fn=process_files,
#     inputs=[
#         gr.File(label="Upload italy.txt"),
#         gr.File(label="Upload france.txt")
#     ],
#     outputs=[
#         gr.Textbox(label="Status"),
#         gr.Textbox(label="First 10 Chunks (Preview)", lines=20)
#     ],
#     title="LangChain Text Splitter App",
#     description="Upload italy.txt and france.txt to split their contents into 300-character chunks using LangChain."
# )

# if __name__ == "__main__":
#     interface.launch()

import os
import gradio as gr
from langchain.text_splitter import RecursiveCharacterTextSplitter
import shutil

# Ensure data folder exists
os.makedirs("data", exist_ok=True)

def process_files(italy_file, france_file):
    docs = []
    text_splitter = RecursiveCharacterTextSplitter(chunk_size=300, chunk_overlap=50)

    try:
        file_paths = {}
        
        # Move italy file
        if italy_file is not None:
            italy_path = "data/italy.txt"
            shutil.copy(italy_file.name, italy_path)
            file_paths["Italy"] = italy_path

        # Move france file
        if france_file is not None:
            france_path = "data/france.txt"
            shutil.copy(france_file.name, france_path)
            file_paths["France"] = france_path

    except Exception as e:
        return f"❌ Error while saving files: {e}", []

    # Process the files
    try:
        for country, path in file_paths.items():
            with open(path, "r", encoding="utf-8") as f:
                content = f.read()
                chunks = text_splitter.create_documents([content])
                docs.extend(chunks)
    except Exception as e:
        return f"❌ Error while reading files: {e}", []

    return f"✅ Successfully split into {len(docs)} chunks.", [doc.page_content for doc in docs[:10]]

# Gradio Interface
gr.Interface(
    fn=process_files,
    inputs=[
        gr.File(label="Upload italy.txt"),
        gr.File(label="Upload france.txt")
    ],
    outputs=[
        gr.Textbox(label="Status"),
        gr.Textbox(label="First 10 Chunks (Preview)", lines=20)
    ],
    title="Text Chunking App (LangChain)",
    description="Upload two .txt files (Italy & France). They will be split into 300-character chunks."
).launch()