app.py

import gradio as gr
import polars as pl

# favourite_langs = {"English": "en", "Romanian": "ro", "German": "de", "-----": "-----"}
favourite_langs = {"English": "en", "Romanian": "ro", "German": "de"}
options = list(favourite_langs.keys())
models = ['ENRO', 'DERO']
# English, Romanian
def search_text(input_text, sselected_language, tselected_language, model_name, hits, toggle_case):
    # df = pl.read_csv('hf://datasets/TiberiuCristianLeon/2RO/ENRO/ENRO.tsv', separator='\t')
    # df = pl.read_parquet('hf://datasets/TiberiuCristianLeon/RSSNEWS/data/train-00000-of-00001.parquet')
    # df = pl.read_parquet('https://huggingface.co/datasets/TiberiuCristianLeon/2RO/resolve/refs%2Fconvert%2Fparquet/default/train/0000.parquet')
    path_to_model = f"https://huggingface.co/api/datasets/TiberiuCristianLeon/2RO/parquet/{model_name.lower()}/train/0.parquet"
    df = pl.read_parquet(path_to_model)
    # Filter rows
    # df.filter(pl.col(sselected_language).str.contains(input_text)).head(hits)
    # print(df.head(hits))
    if toggle_case:
        filtered = df.filter(pl.col(sselected_language).str.contains(input_text).alias("literal")) # case sensitive
    else:
        filtered = df.filter(pl.col(sselected_language).str.contains(f"(?i){input_text}").alias("literal")) # (?i) case insensitive
    # filtered = df.filter(pl.col(sselected_language).str.contains_any([input_text], ascii_case_insensitive=True).alias("contains_any"))
    print(toggle_case, filtered.head(hits))
    # print(filtered)
    # Extract rows
    list_of_arrays = filtered.select([sselected_language, tselected_language]).head(hits)
    # for dataframe type="numpy"
    # list_of_arrays = filtered.select([sselected_language, tselected_language]).head(hits).to_numpy()
    message_text = f'Done! Found {len(list_of_arrays)} entries'
    return list_of_arrays, message_text

# Define a function to swap dropdown values
def swap_languages(src_lang, tgt_lang):
    return tgt_lang, src_lang

def create_interface():
    with gr.Blocks() as interface:
        gr.Markdown("## Search Text in Dataset")

        with gr.Row():
            input_text = gr.Textbox(label="Enter text to search:", placeholder="Type your text here...", info="Press Enter key to start search")
        
        with gr.Row():
            sselected_language = gr.Dropdown(choices=options, value = options[0], label="Source language", interactive=True)
            tselected_language = gr.Dropdown(choices=options, value = options[1], label="Target language", interactive=True)
            swap_button = gr.Button("Swap Languages")
            swap_button.click(fn=swap_languages, inputs=[sselected_language, tselected_language], outputs=[sselected_language, tselected_language])
            toggle_case = gr.Checkbox(info="Case sensitive search", label="Toggle case sensitive search", value=True, interactive=True, visible=True)

        model_name = gr.Dropdown(choices=models, label="Select a dataset", value = models[0], interactive=True)
        search_button = gr.Button("Search")

        translated_text = gr.Dataframe(label="Returned entries:", interactive=False, headers=[options[0], options [1]], datatype=["str", "str"], col_count=(2, "fixed"),
        type="polars", wrap=True, show_row_numbers=False, show_copy_button=True)
        message_text = gr.Textbox(label="Messages:", placeholder="Display field for status and error messages", interactive=False)

        hits = gr.Slider(
            minimum=1,
            maximum=100,
            value=10,
            step=5,
            label="Number of returned hits")
        
        search_button.click(
            search_text, 
            inputs=[input_text, sselected_language, tselected_language, model_name, hits, toggle_case], 
            outputs=[translated_text, message_text]
        )

        # Submit the form when Enter is pressed in the input_text textbox
        input_text.submit(
            search_text, 
            inputs=[input_text, sselected_language, tselected_language, model_name, hits, toggle_case], 
            outputs=[translated_text, message_text]
        )

    return interface

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