What's Changed

• fix: polars test case by @longxiaofei in https://github.com/Kanaries/pygwalker/pull/588
• feat: add pygwalker web api tips in streamlit by @longxiaofei in https://github.com/Kanaries/pygwalker/pull/586
• chore: update duckdb version by @longxiaofei in https://github.com/Kanaries/pygwalker/pull/591
• fix: error spec of ChartPreviewApp by @longxiaofei in https://github.com/Kanaries/pygwalker/pull/592
• feat: add component api by @longxiaofei in https://github.com/Kanaries/pygwalker/pull/593
• feat: new tips for data limit by @longxiaofei in https://github.com/Kanaries/pygwalker/pull/594

Component api

We're excited to announce an experimental new feature in pygwalker:

The component chaining API that enables intuitive, step-by-step chart construction. This feature currently supports rendering to static HTML.

Key Points:

• Introduces a chainable API for creating visualizations
• Allows for more intuitive and readable code
• Currently an experimental feature
• Limited to static HTML output at this time

Example usage:

import pygwalker as pyg
import pandas as pd

df = pd.read_csv("xxx")

(
pyg.component(df)
    .rect()
    .encode(x='bin("feeling_temp", 6)', y='bin("temperature", 6)', color="MEAN(humidity)")
    .layout(height=400, width=460)
)

For more examples, please refer to the pygwalker/examples directory in our repository.

Upcoming Improvements:

1. Optimize the size of pure chart HTML output
2. Implement two-way communication for Jupyter and Streamlit environments

We welcome feedback from our community as we continue to develop and refine this feature.

Please note that as an experimental feature, syntax and functionality may change in future releases

Full Changelog: https://github.com/Kanaries/pygwalker/compare/0.4.9...0.4.9.3