cwprep
ai generate tableau prep file
cwprep - Text-to-PrepFlow Engine
cwprep is a Python-based engine that enables Text-to-PrepFlow generation.
By reverse-engineering the .tfl JSON structure and providing a built-in MCP (Model Context Protocol) server, cwprep acts as a bridge between LLMs (like Claude, Gemini) and Tableau Prep. You can now generate, modify, and build data cleaning flows simply through natural language conversations or Python scripts, without ever opening the GUI!
Author: Cooper Wenhua <[email protected]>
Installation
pip install cwprep
Quick Start
from cwprep import TFLBuilder, TFLPackager
# Create builder
builder = TFLBuilder(flow_name="My Flow")
# Add database connection
conn_id = builder.add_connection(
host="localhost",
username="root",
dbname="mydb"
)
# Add input tables
orders = builder.add_input_table("orders", "orders", conn_id)
customers = builder.add_input_table("customers", "customers", conn_id)
# Join tables
joined = builder.add_join(
name="Orders + Customers",
left_id=orders,
right_id=customers,
left_col="customer_id",
right_col="customer_id",
join_type="left"
)
# Add output
builder.add_output_server("Output", joined, "My_Datasource")
# Build and save
flow, display, meta = builder.build()
TFLPackager.save_tfl("./my_flow.tfl", flow, display, meta)
By default, both the SDK and MCP output only the final .tfl/.tflx archive. Use save_to_folder() only when you explicitly want the exploded folder for inspection.
Features
| Feature | Method | Description |
|---|---|---|
| Database Connection | add_connection() | Connect to MySQL/PostgreSQL/SQL Server |
| File Connection | add_file_connection() | Connect to Excel (.xlsx/.xls) or CSV files |
| SQL Input | add_input_sql() | Custom SQL query input |
| Table Input | add_input_table() | Direct table connection |
| Excel Input | add_input_excel() | Read from Excel worksheet |
| CSV Input | add_input_csv() | Read from CSV file |
| CSV Union | add_input_csv_union() | Merge multiple CSV files |
| Join | add_join() | left/right/inner/full joins (single or multi-column) |
| Union | add_union() | Merge multiple tables |
| Filter | add_filter() | Expression-based filter |
| Value Filter | add_value_filter() | Keep/exclude by values |
| Keep Only | add_keep_only() | Select columns |
| Remove Columns | add_remove_columns() | Drop columns |
| Rename | add_rename() | Rename columns |
| Calculation | add_calculation() | Tableau formula fields |
| Quick Calc | add_quick_calc() | Quick clean (lowercase/uppercase/trim/remove) |
| Change Type | add_change_type() | Change column data types |
| Duplicate Column | add_duplicate_column() | Duplicate (copy) a column |
| Aggregate | add_aggregate() | GROUP BY with SUM/AVG/COUNT |
| Pivot | add_pivot() | Rows to columns |
| Unpivot | add_unpivot() | Columns to rows |
| Output | add_output_server() | Publish to Tableau Server |
| TFLX Packaging | build(is_packaged=True) | Generate .tflx with embedded data files |
| SQL Translation | SQLTranslator | Translate TFL flows to equivalent ANSI SQL |
Examples
See the examples/ directory for complete demos:
demo_basic.py- Input, Join, Outputdemo_cleaning.py- Filter, Calculate, Renamedemo_field_operations.py- Quick Calc, Change Type, Duplicate Columndemo_aggregation.py- Union, Aggregate, Pivotdemo_comprehensive.py- All features combinedprompts.md- 8 ready-to-use MCP prompt templates for AI-driven flow generation
MCP Server
cwprep includes a built-in Model Context Protocol server, enabling AI clients (Claude Desktop, Cursor, Gemini CLI, etc.) to generate TFL files directly.
Prerequisites
| Method | Requirement |
|---|---|
uvx (recommended) | Install uv — it auto-downloads cwprep[mcp] in an isolated env |
pip install | Python ≥ 3.8 + pip install cwprep[mcp] |
Quick Start
# Local (stdio)
cwprep-mcp
# Remote (Streamable HTTP)
cwprep-mcp --transport streamable-http --port 8000
[!TIP] Upgrading? If you previously used
uvxwith an older version, clear the cache to pick up the latest release:uv cache clean cwprep
Client Configuration
All clients below use the uvx method (recommended). Replace uvx with cwprep-mcp if you prefer a local pip install.
Claude Desktop
Edit config file:
- Windows:
%APPDATA%\Claude\claude_desktop_config.json - macOS:
~/Library/Application Support/Claude/claude_desktop_config.json
{
"mcpServers": {
"cwprep": {
"command": "uvx",
"args": ["--from", "cwprep[mcp]", "cwprep-mcp"]
}
}
}
Cursor
Settings → MCP → Add new MCP server, or edit ~/.cursor/mcp.json:
{
"mcpServers": {
"cwprep": {
"command": "uvx",
"args": ["--from", "cwprep[mcp]", "cwprep-mcp"]
}
}
}
VS Code (Copilot)
Create .vscode/mcp.json in project root:
{
"servers": {
"cwprep": {
"command": "uvx",
"args": ["--from", "cwprep[mcp]", "cwprep-mcp"]
}
}
}
Windsurf (Codeium)
Edit ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"cwprep": {
"command": "uvx",
"args": ["--from", "cwprep[mcp]", "cwprep-mcp"]
}
}
}
Claude Code (CLI)
claude mcp add cwprep -- uvx --from "cwprep[mcp]" cwprep-mcp
Gemini CLI
Edit ~/.gemini/settings.json:
{
"mcpServers": {
"cwprep": {
"command": "uvx",
"args": ["--from", "cwprep[mcp]", "cwprep-mcp"]
}
}
}
Continue (VS Code / JetBrains)
Edit ~/.continue/config.yaml:
mcpServers:
- name: cwprep
command: uvx
args:
- --from
- cwprep[mcp]
- cwprep-mcp
Remote HTTP Mode (any client)
Start the server:
cwprep-mcp --transport streamable-http --port 8000
Then configure your client with the endpoint: http://your-server-ip:8000/mcp
Available MCP Capabilities
| Type | Name | Description |
|---|---|---|
| 🔧 Tool | generate_tfl | Generate .tfl/.tflx file from flow definition |
| 🔧 Tool | translate_to_sql | Translate flow definition or .tfl file to ANSI SQL |
| 🔧 Tool | list_supported_operations | List all supported node types |
| 🔧 Tool | validate_flow_definition | Validate flow definition before generating |
| 📖 Resource | cwprep://docs/api-reference | SDK API reference |
| 📖 Resource | cwprep://docs/calculation-syntax | Tableau Prep calculation syntax |
| 📖 Resource | cwprep://docs/best-practices | Common pitfalls and flow design rules |
| 💬 Prompt | design_data_flow | Interactive flow design assistant |
| 💬 Prompt | explain_tfl_structure | TFL file structure explanation |
AI Skill Support
This project includes a specialized AI Skill for assistants like Claude or Gemini to help you build flows.
- Location:
.agents/skills/tfl-generator/ - Features: MCP server index with fallback SDK usage guide. Detailed API and syntax references are served via MCP Resources from
src/cwprep/references/.
Directory Structure
cwprep/
├── .agents/skills/ # AI Agent skills (MCP index)
├── src/cwprep/ # SDK source code
│ ├── builder.py # TFLBuilder class
│ ├── packager.py # TFLPackager class
│ ├── translator.py # SQLTranslator class
│ ├── expression_translator.py # ExpressionTranslator class
│ ├── config.py # Configuration utilities
│ ├── mcp_server.py # MCP Server (Tools, Resources, Prompts)
│ └── references/ # MCP Resource documents (.md)
├── examples/ # Demo scripts
├── docs/ # Documentation
└── tests/ # Unit tests
Configuration
Create config.yaml for default settings:
# MySQL (default)
database:
host: localhost
port: 3306
dbname: mydb
type: mysql
# SQL Server (Windows Authentication)
# database:
# host: localhost
# type: sqlserver
# authentication: sspi
# schema: dbo
# PostgreSQL
# database:
# host: localhost
# port: 5432
# dbname: mydb
# type: postgres
tableau_server:
url: http://your-server
default_project: Default
Changelog
See changelog.md for version history.
License
AGPL-3.0 License
Похожие серверы
WSLSnapit-MCP
Capture screenshots and read the clipboard on Windows from within a WSL environment.
Context Savvy MCP
Transforms Claude Desktop into a memory-enabled AI assistant with persistent context, secure command execution, and intelligent workflow automation.
Kumbify MCP
Tools that boost your productivity, from sending emails, scheduling to news updates—everything you need for your productivity.
Jira Weekly Reporter
Connects to a Jira instance to generate weekly reports based on issue activity.
Obsidian MCP Server
Interact with Obsidian vaults using the Local REST API plugin.
NPX-MCP
A TypeScript server for automation and integration, featuring Microsoft OAuth, browser control, and basic utilities.
Targetprocess
Enables AI assistants to interact with Targetprocess data using semantic operations.
JiraMCP
Full Jira & Confluence CRUD
GitBook
Access and manage GitBook spaces and content using the GitBook API.
unstructured-mcp
Provides unstructured document processing capabilities using the Unstructured API.