pages/HF_Dataset.py

import streamlit as st
import pandas as pd
from huggingface_hub import hf_hub_download
import os
import pydeck as pdk
from datetime import datetime

# Set the Streamlit layout to wide
st.set_page_config(layout="wide")

# Set your Hugging Face token from environment variable
hf_token = os.getenv("pegelonline_dataset_read_only")

if hf_token is None:
    st.error("Hugging Face token not found. Please set the HF_TOKEN environment variable.")
else:
    # Download the dataset file
    dataset_path_stations = hf_hub_download(
        repo_id="DSSG-Wasserwacht/pegelonline-dataset",
        filename="processed/stations.parquet",
        repo_type="dataset",
        use_auth_token=hf_token
    )
    dataset_path_water_level = hf_hub_download(
        repo_id="DSSG-Wasserwacht/pegelonline-dataset",
        filename="processed/current_water_level.parquet",
        repo_type="dataset",
        use_auth_token=hf_token
    )
    dataset_path_timeseries = hf_hub_download(
        repo_id="DSSG-Wasserwacht/pegelonline-dataset",
        filename="processed/timeseries.parquet",
        repo_type="dataset",
        use_auth_token=hf_token
    )

    # Load the dataset
    df_stations = pd.read_parquet(dataset_path_stations)
    df_water_level = pd.read_parquet(dataset_path_water_level)
    df_timeseries = pd.read_parquet(dataset_path_timeseries).groupby("uuid").mean("value").round(1).reset_index().rename(columns={"value": "mean_value"})
    df = df_stations.merge(df_water_level, how="left", on="uuid")
    df = df.merge(df_timeseries, how="left", on="uuid")

    # Format the timestamp nicely
    df["formatted_timestamp"] = df["timestamp"].apply(
        lambda x: datetime.fromisoformat(x).strftime("%Y-%m-%d %H:%M:%S") if pd.notna(x) else None
        )

    # Add arrows based on the result of mean_value - value
    def add_arrow(row):
        if pd.isna(row["value"]):
            return None
        difference = row["mean_value"] - row["value"]
        if abs(difference) <= 0.01 * row["mean_value"]:
            return "➡️"
        elif difference > 0:
            return "⬇️"
        else:
            return "⬆️"

    df["arrow"] = df.apply(add_arrow, axis=1)

    # Add traffic light column
    def add_traffic_light(row):
        if pd.isna(row["value"]):
            return None
        difference = abs(row["mean_value"] - row["value"])
        if difference > 0.15 * row["mean_value"]:
            return "🔴"
        elif difference > 0.10 * row["mean_value"]:
            return "🟡"
        else:
            return "🟢"

    df["traffic_light"] = df.apply(add_traffic_light, axis=1)

    # Define a color mapping for the traffic lights
    color_mapping = {
        "🟢": [0, 255, 0, 140],  # Green
        "🟡": [255, 255, 0, 140],  # Yellow
        "🔴": [255, 0, 0, 140],   # Red
        None: [128, 128, 128, 140]  # Grey for None
    }

    # Map the traffic_light column to colors
    df["color"] = df["traffic_light"].map(color_mapping)

    # Streamlit app
    st.title("Pegelonline Dataset Viewer")
    st.write("This app displays data from the Pegelonline dataset.")

    # PyDeck Layer für Kartenanzeige
    layer = pdk.Layer(
        "ScatterplotLayer",
        data=df,
        get_position=["longitude", "latitude"],
        get_radius=2000,
        get_color="color",
        pickable=True,
    )

    # Deck.gl Map
    view_state = pdk.ViewState(
        latitude=df["latitude"].mean(),
        longitude=df["longitude"].mean(),
        zoom=6,
        pitch=0,
    )
    r = pdk.Deck(
        layers=[layer], 
        initial_view_state=view_state, 
        tooltip={"text": "{shortname}, {value} cm, {mean_value} cm,\n {arrow} {traffic_light} {formatted_timestamp}"}
    )
    st.pydeck_chart(r)

    # Rohdaten anzeigen
    st.write("### Rohdaten der Pegelstationen")
    
    # Display the dataframe
    st.dataframe(df[["shortname", "km", "value", "mean_value", "arrow", "traffic_light", "formatted_timestamp"]])