Condition Tree Builder for Streamlit

Warnings

• gotcha In version 0.3.0, the default value for the `always_show_buttons` parameter was changed to `True`. This may alter the visual presentation of the component, showing rule addition/deletion buttons even when not hovered.
  Fix: To revert to the previous behavior (buttons only on hover), explicitly set `always_show_buttons=False` when calling `condition_tree()`.
• gotcha Version 0.2.0 introduced the capability to inject custom JavaScript code into the component's configuration. While powerful, this feature requires careful handling to avoid security vulnerabilities or unexpected rendering issues.
  Fix: Thoroughly review and sanitize any injected JavaScript code. Ensure it comes from trusted sources and does not expose sensitive information or introduce XSS vulnerabilities.
• gotcha When using Streamlit custom components, especially with `st.session_state`, ensure proper management of component keys. If a `key` is not used or updated correctly, the component might not re-render with new configurations or maintain its state across user interactions as expected.
  Fix: Always provide a unique `key` to `condition_tree` if you intend to programmatically control its state or access the generated tree via `st.session_state[key]`. If the component's `config` changes, ensure the `key` is stable, or use `st.session_state` to manage the initial `tree` parameter for persistence.
• gotcha To save and load a condition tree's state, you must use the `key` parameter to store the tree in `st.session_state`. The saved tree (a dictionary) can then be passed back into the `condition_tree` component via its `tree` parameter to pre-populate it with previous conditions.
  Fix: Assign a `key` to the `condition_tree` component. Access `st.session_state[key]` to get the current tree structure and pass this dictionary to the `tree` parameter in subsequent calls to load it.

Install

• pip install streamlit-condition-tree Install latest version

Imports

• condition_tree

  from streamlit_condition_tree import condition_tree

• config_from_dataframe

  from streamlit_condition_tree import config_from_dataframe

Quickstart

This quickstart demonstrates how to use `streamlit-condition-tree` to filter a Pandas DataFrame dynamically. It initializes a sample DataFrame, automatically generates a configuration for the condition tree, and then uses the `condition_tree` component to construct a query string. The DataFrame is then filtered using the generated query.

import streamlit as st
import pandas as pd
from streamlit_condition_tree import condition_tree, config_from_dataframe

st.set_page_config(layout="wide")
st.title("Streamlit Condition Tree Demo")

# Initial dataframe
df = pd.DataFrame({
    'First Name': ['Georges', 'Alfred', 'Maria', 'Sophie'],
    'Age': [45, 98, 30, 25],
    'Favorite Color': ['Green', 'Red', 'Blue', 'Green'],
    'Like Tomatoes': [True, False, True, True]
})

st.subheader("Original DataFrame")
st.dataframe(df)

# Basic field configuration from dataframe
config = config_from_dataframe(df)

# Condition tree component
query_string = condition_tree(
    config,
    always_show_buttons=True,
    placeholder="Build your filter conditions here..."
)

st.subheader("Generated Query String")
st.code(query_string)

# Filtered dataframe
if query_string:
    try:
        filtered_df = df.query(query_string)
        st.subheader("Filtered DataFrame")
        st.dataframe(filtered_df)
    except Exception as e:
        st.error(f"Error applying query: {e}")
else:
    st.info("No conditions applied yet. Displaying original DataFrame.")