app.py

import gradio as gr
import fitz  # PyMuPDF

def extract_text(pdf_file):
    # 打開 PDF
    doc = fitz.open(pdf_file.name)
    full_text = ""

    # 純文字抽取
    for page in doc:
        full_text += page.get_text()

    # 過濾非 ASCII 字元（只保留英數、標點）
    filtered = full_text.encode("ascii", errors="ignore").decode()

    # 輸出成 txt
    out_path = "output.txt"
    with open(out_path, "w", encoding="utf-8") as f:
        f.write(filtered)

    return out_path

# 建立 Gradio 介面
demo = gr.Interface(
    fn=extract_text,
    inputs=gr.File(label="Upload PDF (.pdf)"),
    outputs=gr.File(label="Download TXT"),
    title="PDF → TXT (English only)",
    description="Extract English text from PDF (純文字抽取) and download as .txt"
)

if __name__ == "__main__":
    demo.launch()