# Taipy — Build Data & AI Web Apps in Python

> Full-stack Python framework for data dashboards and AI pipelines. GUI builder + workflow orchestration in one package. 15K+ stars.

## Install

Copy the content below into your project:

# Taipy — Build Data & AI Web Apps in Python

## Quick Use

```bash
pip install taipy
```

```python
import taipy.gui.builder as tgb
from taipy.gui import Gui

value = "Hello World"

with tgb.Page() as page:
    tgb.text("# My AI App", mode="md")
    tgb.input("{value}", label="Enter text")
    tgb.text("You typed: {value}")

Gui(page=page).run()
```

This launches a web app at `http://localhost:5000`. For AI pipeline orchestration:

```python
from taipy import Config

# Define data nodes and tasks
input_cfg = Config.configure_data_node("raw_data", path="data.csv")
clean_cfg = Config.configure_data_node("clean_data")
model_cfg = Config.configure_data_node("predictions")

clean_task = Config.configure_task("clean", clean_function, [input_cfg], [clean_cfg])
predict_task = Config.configure_task("predict", predict_function, [clean_cfg], [model_cfg])

scenario_cfg = Config.configure_scenario("ml_pipeline", [clean_task, predict_task])
```

---

## Intro

Taipy is a full-stack Python framework with 15,000+ GitHub stars for building production data and AI web applications. It combines a **GUI builder** (create interactive dashboards without JavaScript) with a **workflow orchestrator** (manage data pipelines, ML training, scenario comparison). With Taipy, a single Python developer can build what normally requires a frontend engineer + backend engineer + DevOps. Used in production by enterprises for AI dashboards, what-if scenario analysis, and ML pipeline management.

Works with: Python, Pandas, scikit-learn, TensorFlow, PyTorch, any Python ML library. Best for data scientists and ML engineers who want to build full web apps without learning React/Vue. Setup time: under 3 minutes.

---

## Taipy Two-in-One Architecture

### Taipy GUI — Interactive Dashboards

Build rich web UIs with pure Python — no HTML, CSS, or JavaScript:

```python
import taipy.gui.builder as tgb
from taipy.gui import Gui
import pandas as pd

data = pd.DataFrame({"Month": ["Jan","Feb","Mar"], "Revenue": [100, 150, 200]})
selected_month = "All"

def filter_data(state):
    if state.selected_month == "All":
        state.filtered = data
    else:
        state.filtered = data[data["Month"] == state.selected_month]

with tgb.Page() as dashboard:
    tgb.text("# Revenue Dashboard", mode="md")
    tgb.selector("{selected_month}", lov=["All","Jan","Feb","Mar"],
                 on_change=filter_data)
    tgb.chart("{filtered}", x="Month", y="Revenue", type="bar")
    tgb.table("{filtered}")

Gui(page=dashboard).run()
```

Available components:
- **Charts** — Line, bar, scatter, pie, heatmap (Plotly-based)
- **Tables** — Sortable, filterable, editable data tables
- **Forms** — Input, slider, toggle, date picker, file upload
- **Layout** — Columns, expandable sections, tabs, dialog
- **Media** — Images, videos, maps

### Taipy Core — Pipeline Orchestration

Manage data pipelines with versioning and scenario comparison:

```python
from taipy import Config
import taipy as tp

# Define the pipeline
raw = Config.configure_data_node("raw_data", path="sales.csv")
forecast = Config.configure_data_node("forecast")

train_task = Config.configure_task("train_model", train_fn, [raw], [forecast])
scenario_cfg = Config.configure_scenario("sales_forecast", [train_task])

# Run scenarios
tp.Core().run()
scenario_1 = tp.create_scenario(scenario_cfg)
tp.submit(scenario_1)

# Compare scenarios (what-if analysis)
scenario_2 = tp.create_scenario(scenario_cfg)
scenario_2.raw_data.write(alternative_data)
tp.submit(scenario_2)

# Both scenarios tracked with full lineage
```

### What-If Scenario Analysis

Taipy's killer feature — compare multiple pipeline runs:

```
Scenario A: "Base forecast with 2025 data"
  └─ Revenue prediction: $2.1M

Scenario B: "Optimistic — 10% growth assumption"
  └─ Revenue prediction: $2.3M

Scenario C: "Conservative — recession adjustment"
  └─ Revenue prediction: $1.8M

→ Side-by-side comparison in the dashboard
```

### REST API (Built-in)

Every Taipy app automatically exposes a REST API:

```bash
# Get scenario results
GET /api/v1/scenarios/{id}

# Submit a new scenario
POST /api/v1/scenarios/{id}/submit

# Read data node
GET /api/v1/datanodes/{id}/read
```

### Deployment

```bash
# Development
taipy run app.py

# Production with Gunicorn
gunicorn -w 4 -b 0.0.0.0:5000 app:gui

# Docker
docker build -t my-taipy-app .
docker run -p 5000:5000 my-taipy-app
```

---

## FAQ

**Q: What is Taipy?**
A: Taipy is a full-stack Python framework with 15,000+ GitHub stars that combines a GUI builder (create dashboards without JavaScript) and a workflow orchestrator (manage ML pipelines and scenarios) in one package.

**Q: How is Taipy different from Streamlit or Gradio?**
A: Streamlit/Gradio are great for quick demos but struggle with production needs. Taipy adds pipeline orchestration, scenario management, multi-page apps, built-in REST API, and enterprise deployment — it's designed for production, not just prototyping.

**Q: Is Taipy free?**
A: The Community Edition is free and open-source under Apache-2.0. Taipy also offers an Enterprise Edition with additional features.

---

## Source & Thanks

> Created by [Avaiga](https://github.com/Avaiga). Licensed under Apache-2.0.
>
> [taipy](https://github.com/Avaiga/taipy) — ⭐ 15,000+

Thanks to the Avaiga team for making full-stack AI app development accessible to Python developers.

---

<!-- ZH -->

## 快速使用

```bash
pip install taipy
```

```python
import taipy.gui.builder as tgb
from taipy.gui import Gui

value = "你好世界"

with tgb.Page() as page:
    tgb.text("# 我的 AI 应用", mode="md")
    tgb.input("{value}", label="输入文本")
    tgb.text("你输入了：{value}")

Gui(page=page).run()
```

---

## 简介

Taipy 是一个拥有 15,000+ GitHub stars 的全栈 Python 框架，用于构建生产级数据和 AI Web 应用。它将 **GUI 构建器**（无需 JavaScript 创建交互式仪表板）和**工作流编排器**（管理数据管线、ML 训练、场景对比）合二为一。一个 Python 开发者就能构建通常需要前端+后端+运维三个角色的应用。

适用于：Python、Pandas、scikit-learn、TensorFlow、PyTorch。适合不想学 React/Vue 就能构建完整 Web 应用的数据科学家和 ML 工程师。

---

## 核心特性

### GUI 构建器
纯 Python 构建丰富 Web UI：图表、表格、表单、布局，不需要 HTML/CSS/JavaScript。

### 管线编排
定义数据节点和任务，自动管理依赖关系和版本控制。

### 场景对比分析
Taipy 的杀手级功能 — 运行多个管线场景并排对比结果。

### 内置 REST API
每个 Taipy 应用自动暴露 REST API。

### 生产部署
支持 Gunicorn、Docker、企业级部署。

---

## 来源与感谢

> Created by [Avaiga](https://github.com/Avaiga). Licensed under Apache-2.0.
>
> [taipy](https://github.com/Avaiga/taipy) — ⭐ 15,000+


---
Source: https://tokrepo.com/en/workflows/c926c5d3-35bc-463f-8450-c1e6e8449050
Author: TokRepo精选