app.py

import io
import os
from pathlib import Path

import pandas as pd
import streamlit as st

from business_export_job import JobConfig, run_job

# On HF Spaces /app/output doesn't exist; default to a local output folder.
if "SWISSPARL_OUTPUT_DIR" not in os.environ:
    os.environ["SWISSPARL_OUTPUT_DIR"] = str(Path(__file__).parent / "output")

OUTPUT_DIR = Path(os.environ["SWISSPARL_OUTPUT_DIR"])
OUTPUT_DIR.mkdir(parents=True, exist_ok=True)

st.set_page_config(page_title="SwissParl Topic Tagger – Results", layout="wide")
st.title("SwissParl Topic Tagger – Results")

with st.sidebar:
    st.header("Actions")
    if st.button("Trigger Topic Modelling", help="Run the topic modelling job now."):
        with st.spinner("Running topic modelling... this may take a while."):
            try:
                config = JobConfig.from_env()
                # Ensure the filename template includes the timestamp if we want it to be unique
                if "{timestamp}" not in config.output_filename_template:
                    config = JobConfig(
                        output_dir=config.output_dir,
                        output_filename_template=config.output_filename_template.replace(".xlsx", "-{timestamp}.xlsx"),
                        days_past=config.days_past,
                        start_date=config.start_date,
                        end_date=config.end_date,
                        languages=config.languages,
                        language_priority=config.language_priority,
                        topic_score_threshold=config.topic_score_threshold,
                        min_chunk_days=config.min_chunk_days,
                        excel_sheet_name=config.excel_sheet_name,
                    )
                output_path = run_job(config)
                st.success(f"Job completed! Result saved to `{output_path.name}`")
                st.rerun()
            except Exception as e:
                st.error(f"Job failed: {e}")

excel_files = sorted(OUTPUT_DIR.glob("*.xlsx"), reverse=True)

if not excel_files:
    st.warning(f"No Excel files found in `{OUTPUT_DIR}`. Use the sidebar button to run the topic modelling job.")
    st.stop()

selected = st.selectbox(
    "Select a result file",
    excel_files,
    format_func=lambda p: p.name,
)


@st.cache_data
def load_excel(path: str) -> pd.DataFrame:
    return pd.read_excel(path, engine="openpyxl")


df = load_excel(str(selected))

# Download button (above the table so it's always visible)
buf = io.BytesIO()
with pd.ExcelWriter(buf, engine="openpyxl") as writer:
    df.to_excel(writer, index=False)
buf.seek(0)

col1, col2 = st.columns([3, 1])
with col1:
    st.caption(f"{len(df):,} rows · {len(df.columns)} columns · `{selected.name}`")
with col2:
    st.download_button(
        label="Download Excel",
        data=buf,
        file_name=selected.name,
        mime="application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
    )

# Row filter
search = st.text_input("Filter rows (searches all text columns)", "")
if search:
    mask = df.apply(
        lambda col: col.astype(str).str.contains(search, case=False, na=False)
    ).any(axis=1)
    df = df[mask]
    st.caption(f"{len(df):,} rows match the filter")

st.dataframe(df, use_container_width=True, height=600)