abubasith86's picture
Update app.py
f3a8b9c verified
raw
history blame
4.76 kB
import streamlit as st
import json
import pandas as pd
import os
from uuid import uuid4
st.set_page_config(page_title="Dataset Builder", layout="wide")
st.title("πŸ“š 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):
keys = set()
for d in data:
keys.update(d.keys())
return sorted(list(keys))
def save_to_file():
with open(TMP_FILE, "w", encoding="utf-8") as f:
for row in st.session_state.data:
f.write(json.dumps(row, ensure_ascii=False) + "\n")
# --- 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 "editor_key" not in st.session_state:
st.session_state.editor_key = str(uuid4())
if "just_added" not in st.session_state:
st.session_state.just_added = False
# --- Load from TMP file ---
if os.path.exists(TMP_FILE) and not st.session_state.data:
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)
# --- File Upload ---
uploaded = st.file_uploader("Upload JSONL", type=["jsonl"])
if uploaded:
lines = uploaded.read().decode("utf-8").splitlines()
st.session_state.data = [json.loads(l) for l in lines]
st.session_state.all_fields = get_all_fields(st.session_state.data)
save_to_file()
st.session_state.editor_key = str(uuid4())
st.rerun()
# --- Add New Fields fallback ---
if not st.session_state.all_fields:
st.session_state.all_fields = ["context", "question", "answer"]
# --- Add Entry Form ---
with st.form("add_form"):
st.markdown("### βž• Add New Entry")
new_entry = {}
for field in st.session_state.all_fields:
new_entry[field] = st.text_area(field, key=f"add_{field}")
submit_add = st.form_submit_button("Add Entry")
if submit_add:
st.session_state.data.append(new_entry)
save_to_file()
st.session_state.editor_key = str(uuid4())
st.session_state.just_added = True
st.rerun()
# --- Wait for rerun before rendering editor ---
if st.session_state.just_added:
st.session_state.just_added = False
st.rerun()
# --- Display Editor ---
st.markdown("### ✏️ Edit Records")
df = pd.DataFrame(st.session_state.data)
df = df.reindex(columns=st.session_state.all_fields)
for field in st.session_state.all_fields:
df[field] = df[field].astype(str)
column_configs = {
field: st.column_config.TextColumn(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,
key=st.session_state.editor_key,
use_container_width=True,
num_rows="dynamic",
column_config=column_configs,
)
# --- Save edits ---
if edited_df is not None:
new_data = edited_df.fillna("").to_dict(orient="records")
if new_data != st.session_state.data:
st.session_state.data = new_data
save_to_file()
st.toast("βœ… Changes auto-saved!", icon="πŸ’Ύ")
# --- Add New Field ---
with st.expander("βž• 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 Dataset ---
st.markdown("### πŸ“€ Export")
export_path = st.text_input("Save path", value="./exports/exported_dataset.jsonl")
col1, col2, col3 = st.columns(3)
with col1:
if st.button("πŸ“ 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))
st.success("βœ… Exported!")
st.session_state.clear()
if os.path.exists(TMP_FILE):
os.remove(TMP_FILE)
st.rerun()
with col2:
if os.path.exists(TMP_FILE):
with open(TMP_FILE, "r", encoding="utf-8") as f:
tmp_data = f.read()
st.download_button("⬇️ Temp File", tmp_data, file_name="session_dataset.jsonl")
with col3:
if st.button("🧹 Clear Session"):
if os.path.exists(TMP_FILE):
os.remove(TMP_FILE)
st.session_state.clear()
st.rerun()