backup2.app.py

from typing import List, Dict
import httpx
import gradio as gr
import pandas as pd
import json

async def get_splits(dataset_name: str) -> Dict[str, List[Dict]]:
    URL = f"https://huggingface.co/api/datasets/{dataset_name}"
    async with httpx.AsyncClient() as session:
        response = await session.get(URL)
        dataset_info = response.json()
        return {
            "splits": [
                {"split": split_name, "config": config_name}
                for config_name, config_info in dataset_info.get("config", {}).items()
                for split_name in config_info.get("splits", [])
            ]
        }

async def get_valid_datasets() -> List[str]:
    URL = f"https://huggingface.co/api/datasets"
    async with httpx.AsyncClient() as session:
        response = await session.get(URL)
        try:
            datasets_raw = response.json()
            datasets = []
            for dataset_raw in datasets_raw:
                if isinstance(dataset_raw, str) and "/" in dataset_raw:
                    datasets.append(dataset_raw)
                elif isinstance(dataset_raw, dict) and "/" in dataset_raw.get("id", ""):
                    datasets.append(dataset_raw["id"])
        except (KeyError, json.JSONDecodeError):
            datasets = []  # Set a default value if the response is not in the expected format
        return datasets

async def get_first_rows(dataset: str, config: str, split: str) -> Dict[str, Dict[str, List[Dict]]]:
    URL = f"https://huggingface.co/datasets/{dataset}/resolve/main/dataset_info.json"
    async with httpx.AsyncClient() as session:
        response = await session.get(URL)
        dataset_info = response.json()
        split_info = dataset_info["splits"][split]
        first_rows = {
            "rows": [
                {"row": row} for row in split_info["examples"][:10]
            ]
        }
        return first_rows

def get_df_from_rows(api_output):
    dfFromSort = pd.DataFrame([row["row"] for row in api_output["rows"]])
    try:
        dfFromSort.sort_values(by=1, axis=1, ascending=True, inplace=False, kind='mergesort', na_position='last', ignore_index=False, key=None)
    except:
        print("Exception sorting due to keyerror?")
    return dfFromSort

async def update_configs(dataset_name: str):
    try:
        splits = await get_splits(dataset_name)
        all_configs = sorted(set([s["config"] for s in splits["splits"]]))
        return (all_configs, all_configs[0] if all_configs else None), splits
    except json.JSONDecodeError:
        return [], None, {"splits": []}

async def update_splits(config_name: str, state: gr.State):
    if state["splits"] is None:
        return [], None, gr.DataFrame.update()

    splits_for_config = sorted(set([s["split"] for s in state["splits"] if s["config"] == config_name]))
    dataset_name = state["splits"][0]["dataset"] if state["splits"] else None
    dataset = await update_dataset(splits_for_config[0], config_name, dataset_name) if splits_for_config and dataset_name else gr.DataFrame.update()
    return splits_for_config, splits_for_config[0] if splits_for_config else None, dataset

async def update_dataset(split_name: str, config_name: str, dataset_name: str):
    rows = await get_first_rows(dataset_name, config_name, split_name)
    df = get_df_from_rows(rows)
    return df

async def update_URL(dataset: str, config: str, split: str) -> str:
    URL = f"https://huggingface.co/datasets/{dataset}/tree/main/{config}/{split}"
    return URL
   
async def openurl(URL: str) -> str:
    html = f"<a href={URL} target=_blank>{URL}</a>"
    return html

with gr.Blocks() as demo:
    gr.Markdown("<h1><center>🥫Datasetter📊 Datasets Analyzer and Transformer</center></h1>")
    gr.Markdown("""<div align="center">Curated Datasets: <a href = "https://www.kaggle.com/datasets">Kaggle</a>. <a href="https://www.nlm.nih.gov/research/umls/index.html">NLM UMLS</a>.  <a href="https://loinc.org/downloads/">LOINC</a>. <a href="https://www.cms.gov/medicare/icd-10/2022-icd-10-cm">ICD10 Diagnosis</a>. <a href="https://icd.who.int/dev11/downloads">ICD11</a>.  <a href="https://paperswithcode.com/datasets?q=medical&v=lst&o=newest">Papers,Code,Datasets for SOTA in Medicine</a>.   <a href="https://paperswithcode.com/datasets?q=mental&v=lst&o=newest">Mental</a>.  <a href="https://paperswithcode.com/datasets?q=behavior&v=lst&o=newest">Behavior</a>. <a href="https://www.cms.gov/medicare-coverage-database/downloads/downloads.aspx">CMS Downloads</a>.  <a href="https://www.cms.gov/medicare/fraud-and-abuse/physicianselfreferral/list_of_codes">CMS CPT and HCPCS Procedures and Services</a>  """)

    splits_data = gr.State()
    
    with gr.Row():
        dataset_name = gr.Dropdown(label="Dataset", interactive=True, allow_custom_value=True)
        config = gr.Dropdown(label="Subset", interactive=True, allow_custom_value=True)
        split = gr.Dropdown(label="Split", interactive=True, allow_custom_value=True)
    
    with gr.Row():
        URLcenter = gr.Textbox(label="Dataset URL", placeholder="URL")
        btn = gr.Button("Use Dataset")
        URLoutput = gr.HTML(label="Output")

    with gr.Row():
        dataset = gr.DataFrame(wrap=True, interactive=True)
    
    demo.load(get_valid_datasets, inputs=None, outputs=[dataset_name])
    
    dataset_name.change(update_configs, inputs=[dataset_name], outputs=[config, splits_data])
    config.change(update_splits, inputs=[config, splits_data], outputs=[split, dataset])
    split.change(update_dataset, inputs=[split, config, dataset_name], outputs=[dataset])

    dataset_name.change(update_URL, inputs=[split, config, dataset_name], outputs=[URLcenter])

    btn.click(openurl, [URLcenter], URLoutput)

demo.launch(debug=True)