app.py

# clean_app.py

# -----------------------------------
# Copy_right CC _developed By HNM
# -----------------------------------
import re
import gradio as gr
import tempfile

# STEP A
mappings = {
    "\x04ٲ": "ٲ",
    "\x04ُ": "ُ",
    "\x04ٚ": "ٚ",
    "\x04ٕ": "ٕ",
    "\x04ٛ": "ٛ",
    "\x04ٔ": "ٔ",
    "\x04ں": "ں",
    "\x04":  "",   # if stray \x04 alone, remove or replace as needed

    # ASCII symbol mappings
    ">": "ٲ",
    "<": "ُ",
    ";": "ٚ",
    "=": "ٕ",
    ":": "ٛ",
    ".": "ٔ",
    ",": "ں",
    "/":  ""       # remove slash
}

def basic_replacements(text: str) -> str:
    """
    Apply the dictionary-based .replace() calls for
    the \x04 combos and ASCII symbols.
    """
    for old, new in mappings.items():
        text = text.replace(old, new)
    return text

# STEP B
def fix_alif_combo(text: str) -> str:
    """ Replace any occurrence of 'اٲ' with 'ٲ'. """
    return text.replace("اٲ", "ٲ")

# STEP C
def fix_question_mark(text: str) -> str:
    """
    For each occurrence of (.)?(.) => remove '?', add "یٕ" to the 2nd letter.
    E.g. "س?ت" => "ستیٕ".
    """
    def _repl(m):
        first_char = m.group(1)  # the character before '?'
        second_char = m.group(2) # the character after '?'
        return f"{first_char}{second_char}یٕ"

    pattern = r"(.)\?(.)"
    return re.sub(pattern, _repl, text)

# STEP D
def clean_line(line: str) -> str:
    """
    Cleans a single line using:
     1) basic replacements (\x04 combos, ASCII symbols),
     2) fix_alif_combo (اٲ -> ٲ),
     3) fix_question_mark (س?ت -> ستیٕ)
    """
    line = basic_replacements(line)
    line = fix_alif_combo(line)
    line = fix_question_mark(line)
    return line

# Helper: cleans the entire string (multiple lines).
def clean_text(input_text: str) -> str:
    # Split into lines, clean each, then join
    lines = input_text.splitlines()
    cleaned_lines = [clean_line(line) for line in lines]
    return "\n".join(cleaned_lines)


# -----------------------------------
# Gradio Interface
# -----------------------------------
def process_text(raw_text):
    """
    This function is called by Gradio when the user clicks the button.
    It returns two outputs:
     1) The cleaned text (for display)
     2) A temporary file path with the cleaned text (for download)
    """
    cleaned = clean_text(raw_text)

    # Write the cleaned text to a temporary file for download
    tmp = tempfile.NamedTemporaryFile(delete=False, suffix=".txt", mode="w", encoding="utf-8")
    tmp.write(cleaned)
    tmp.flush()
    tmp.close()

    return cleaned, tmp.name


# Build the interface
with gr.Blocks() as demo:
    gr.Markdown("## Clean Text Tool")
    gr.Markdown(
        "Paste your raw/unprocessed text below, then click 'Clean Text' to get the cleaned result."
    )

    with gr.Row():
        with gr.Column():
            raw_text = gr.Textbox(
                label="Input (Paste uncleaned text)",
                lines=15,
                placeholder="Paste any length of text here...",
            )
        with gr.Column():
            cleaned_output = gr.Textbox(
                label="Output (Cleaned text)",
                lines=15,
                interactive=False
            )

    # Button to trigger cleaning
    button = gr.Button("Clean Text")

    # We'll show the file download output in a second row
    download_file = gr.File(label="Download Cleaned .txt File")

    # Connect the function to the button
    button.click(
        fn=process_text,
        inputs=raw_text,
        outputs=[cleaned_output, download_file]
    )

# Run the app
if __name__ == "__main__":
    demo.launch()