app.py

import streamlit as st
import json
import pandas as pd
import os

st.set_page_config(page_title="Dataset Builder", layout="wide")
st.title("📚 JSONL Dataset Editor")

TMP_DIR = "temp"
TMP_FILE = os.path.join(TMP_DIR, "session_dataset.jsonl")
os.makedirs(TMP_DIR, exist_ok=True)

# --- Helpers ---
def get_all_fields(data):
    all_keys = set()
    for record in data:
        all_keys.update(record.keys())
    return sorted(all_keys)

# --- Session Initialization ---
if "data" not in st.session_state:
    st.session_state.data = []
if "all_fields" not in st.session_state:
    st.session_state.all_fields = []
if "prev_data" not in st.session_state:
    st.session_state.prev_data = []

# --- Load from temp if needed ---
if not st.session_state.data and os.path.exists(TMP_FILE):
    with open(TMP_FILE, "r", encoding="utf-8") as f:
        st.session_state.data = [json.loads(line) for line in f]
    st.session_state.all_fields = get_all_fields(st.session_state.data)
    st.session_state.prev_data = st.session_state.data.copy()

# --- Upload JSONL ---
uploaded_file = st.file_uploader("Upload a JSONL file", type=["jsonl"])
if uploaded_file:
    content = uploaded_file.read().decode("utf-8")
    st.session_state.data = [json.loads(line) for line in content.strip().splitlines()]
    st.session_state.all_fields = get_all_fields(st.session_state.data)
    st.session_state.prev_data = st.session_state.data.copy()
    with open(TMP_FILE, "w", encoding="utf-8") as f:
        for item in st.session_state.data:
            f.write(json.dumps(item, ensure_ascii=False) + "\n")
    st.rerun()

# --- Fallback fields if none ---
if not st.session_state.all_fields:
    st.session_state.all_fields = ["context", "question", "answer"]

# --- Edit Records ---
st.markdown("### ✏️ Edit Records")
df = pd.DataFrame(st.session_state.data)
df = df.reindex(columns=st.session_state.all_fields)

# Ensure fields are strings for editor
for field in st.session_state.all_fields:
    if field.lower() in ["context", "question", "answer"]:
        df[field] = df[field].astype(str)

# TextAreas for longer fields
column_configs = {
    field: (
        st.column_config.TextColumn(label=field, width="large")
        if field.lower() in ["context", "question", "answer"]
        else None
    )
    for field in st.session_state.all_fields
}

edited_df = st.data_editor(
    df,
    use_container_width=True,
    num_rows="dynamic",
    column_config=column_configs,
    key="editable_table",
)

# Save if changed
new_data = edited_df.fillna("").to_dict(orient="records")
if new_data != st.session_state.prev_data:
    st.session_state.data = new_data
    st.session_state.prev_data = new_data.copy()
    with open(TMP_FILE, "w", encoding="utf-8") as f:
        for item in new_data:
            f.write(json.dumps(item, ensure_ascii=False) + "\n")
    st.toast("✅ Auto-saved!", icon="💾")

# --- Add New Entry ---
st.markdown("### ➕ Add New Entry")
with st.form("new_entry_form"):
    new_record = {}
    for field in st.session_state.all_fields:
        new_record[field] = st.text_area(f"{field}", key=f"input_{field}")
    submitted = st.form_submit_button("Add Entry")
    if submitted:
        st.session_state.data.append(new_record)
        st.session_state.prev_data = st.session_state.data.copy()
        with open(TMP_FILE, "w", encoding="utf-8") as f:
            for item in st.session_state.data:
                f.write(json.dumps(item, ensure_ascii=False) + "\n")
        st.success("✅ New entry added!")
        st.rerun()

# --- Add New Field ---
with st.expander("➕ Add New Field"):
    new_field = st.text_input("New field name", key="new_field_name")
    if st.button("Add Field"):
        if new_field and new_field not in st.session_state.all_fields:
            st.session_state.all_fields.append(new_field)
            st.rerun()

# --- Export Section ---
st.markdown("### 📤 Export Dataset")
export_path = st.text_input("Save path", value="./exports/exported_dataset.jsonl")

col1, col2, col3 = st.columns(3)

# Export
with col1:
    if st.button("📁 Export JSONL"):
        os.makedirs(os.path.dirname(export_path), exist_ok=True)
        with open(export_path, "w", encoding="utf-8") as f:
            for row in st.session_state.data:
                f.write(json.dumps(row, ensure_ascii=False) + "\n")
        with open(export_path, "r", encoding="utf-8") as f:
            content = f.read()
        st.download_button("⬇️ Download JSONL", content, file_name=os.path.basename(export_path))
        if os.path.exists(TMP_FILE):
            os.remove(TMP_FILE)
        st.session_state.clear()
        st.rerun()

# Download temp
with col2:
    if os.path.exists(TMP_FILE):
        with open(TMP_FILE, "r", encoding="utf-8") as f:
            tmp_content = f.read()
        st.download_button(
            "⬇️ Download Temp File",
            tmp_content,
            file_name="session_dataset.jsonl",
            mime="application/json",
        )
    else:
        st.warning("⚠️ No temp file found.")

# Clear session
with col3:
    if st.button("🗑️ Clear Session"):
        if os.path.exists(TMP_FILE):
            os.remove(TMP_FILE)
        st.session_state.clear()
        st.rerun()