DataContext for VS Code

---

Features

📊 Database Explorer

Browse your database schema directly in VS Code. See tables, columns, types, and primary keys at a glance.

Supported Databases:

• PostgreSQL
• MySQL/MariaDB
• SQLite
• MongoDB (NEW!) — Schema inferred from sample documents

💡 SQL Autocomplete

Intelligent code completion based on your database schema:

• Table names after FROM, JOIN, INTO
• Column names with table. notation
• SQL keywords with snippets
• Context-aware suggestions

📝 Inline Descriptions

Hover over table or column names in SQL files to see business descriptions, data types, and constraints.

🚀 SQL Execution

Execute SQL queries directly from VS Code with Cmd+Enter:

• Results displayed in a formatted table panel
• Copy results as CSV or JSON
• Execution time and row count metrics

🤖 AI-Powered SQL Generation

Generate, fix, and explain SQL using LLM providers:

• Generate SQL — Describe what you want in natural language
• Fix SQL — Let AI correct your query based on context
• Explain SQL — Get human-readable explanations

📚 Knowledge Panel

View and manage stored knowledge:

• Relationships — Table relationships and foreign keys ✨
• Business Glossary — Business terms with SQL expressions
• Table Descriptions — What each table represents
• Business Rules — Domain constraints and conventions
• Query Examples — Verified SQL patterns

🕸️ Knowledge Graph ✨ NEW

Automatically discover and visualize table relationships:

• Show Relationships — View all FK relationships in a table
• Find Join Path — Discover how two tables connect (multi-hop)
• Graph Summary — Overview of your schema connections

🌾 Auto-Harvesting

Automatically extract metadata from your database:

• Table/column comments
• Primary keys and indexes
• Foreign key relationships

🛡️ Safety & Cost Estimation

Before running expensive queries:

• Estimate Cost — Preview query impact
• Warnings — Get alerts for potentially slow queries

📊 Feedback & Learning

Record query corrections to improve AI suggestions over time:

• Record Feedback — Track successful queries and corrections
• View Metrics — See success rates and correction patterns

---

Quick Start

1. Start DataContext Server

npx @kiyeonjeon21/datacontext serve postgres://user:pass@localhost:5432/mydb --port 3000

2. Connect in VS Code

1. Open Command Palette (Cmd+Shift+P)
2. Run "DataContext: Connect to Database"
3. Enter API URL: http://localhost:3000

3. Explore

• Click the DataContext icon in the Activity Bar (left sidebar)
• Browse Tables → expand to see columns
• View Knowledge → see descriptions, rules, examples
• Create a .sql file to test autocomplete

---

Commands

Connection & Schema

Command	Keybinding	Description
DataContext: Connect to Database	—	Connect to DataContext API server
DataContext: Disconnect	—	Disconnect from server
DataContext: Switch Connection	—	Switch between saved connections
DataContext: Remove Connection	—	Remove a saved connection
DataContext: Clear All Connections	—	Clear all connections

SQL Execution

Command	Keybinding	Description
DataContext: Execute Query	Cmd+Enter	Execute SQL in editor
DataContext: Estimate Query Cost	—	Analyze query before running

AI Features

Command	Keybinding	Description
DataContext: Generate SQL with AI	Cmd+Shift+G	Generate SQL from natural language
DataContext: Fix SQL with AI	—	Fix selected SQL with AI
DataContext: Explain SQL with AI	—	Get explanation of SQL query
DataContext: Set API Key (Secure)	—	Store API key securely
DataContext: Clear API Key	—	Remove stored API key

Knowledge Graph ✨

Command	Keybinding	Description
DataContext: Show Table Relationships	—	View all FK relationships
DataContext: Find Join Path	—	Find path between two tables
DataContext: Show Knowledge Graph Summary	—	View graph statistics

Knowledge Management

Command	Keybinding	Description
DataContext: Harvest Metadata	—	Auto-extract database metadata
DataContext: Describe Table/Column	—	Add description to table or column
DataContext: Add Business Rule	—	Add a business rule
DataContext: Add Query Example	—	Add a verified query example
DataContext: Refresh Knowledge	—	Refresh knowledge panel

Business Glossary ✨

Command	Keybinding	Description
DataContext: Add Term	—	Add a business term manually
DataContext: Generate Glossary with AI	—	Generate terms from natural language
DataContext: Search Terms	—	Search business glossary
DataContext: Delete Term	—	Delete a business term

Feedback & Metrics

Command	Keybinding	Description
DataContext: Record Feedback	—	Record query feedback for learning
DataContext: View Metrics	—	View query success metrics
DataContext: Show Context Panel	—	Show context for current SQL

---

Configuration

Basic Settings

{
  "datacontext.connections": [
    {
      "name": "Local Dev",
      "connectionString": "http://localhost:3000",
      "schema": "public"
    }
  ],
  "datacontext.autoHarvest": true,
  "datacontext.showInlineDescriptions": true
}

LLM Settings (for AI features)

Configure an LLM provider to enable AI-powered SQL generation:

{
  "datacontext.llm.provider": "openai",
  "datacontext.llm.apiKey": "sk-your-api-key",
  "datacontext.llm.model": "gpt-4o-mini"
}

Supported Providers

Provider	Config Value	Models
OpenAI	"openai"	gpt-4o-mini, gpt-4o, gpt-3.5-turbo
Anthropic	"anthropic"	claude-3-haiku-20240307, claude-3-sonnet-20240229
Mock (testing)	"mock"	—
None (disabled)	"none"	—

Custom Endpoint (Local LLMs)

For local LLMs or proxies:

{
  "datacontext.llm.provider": "openai",
  "datacontext.llm.endpoint": "http://localhost:11434/v1/chat/completions",
  "datacontext.llm.model": "llama2"
}

---

Context Menus

Right-click in SQL files for quick access:

• Execute Query — Run the current SQL
• Generate SQL with AI — Create SQL from description
• Fix SQL with AI — Fix selected SQL (when text selected)
• Explain SQL with AI — Get explanation

---

Requirements

• VS Code 1.85.0 or higher
• DataContext API server running (npx @kiyeonjeon21/datacontext serve ...)

---

How It Works

This extension connects to a DataContext REST API server (not directly to your database). This architecture:

1. Separates concerns — Extension handles UI, server handles database
2. Enables sharing — Multiple developers can connect to the same server
3. Improves security — Database credentials stay on the server

┌─────────────────┐     HTTP      ┌─────────────────┐     SQL      ┌──────────┐
│  VS Code Ext    │ ───────────▶ │  DataContext    │ ──────────▶ │ Database │
│  (This)         │              │  API Server     │              │          │
└─────────────────┘              └─────────────────┘              └──────────┘

---

API Endpoints Used

This extension uses the following REST API endpoints from the DataContext server:

Schema & Query

Endpoint	Method	Purpose
/health	GET	Connection health check
/api/tables	GET	List all tables
/api/tables/:name	GET	Get table details with context
/api/tables/:name/description	PUT	Set table description
/api/tables/:table/columns/:column/description	PUT	Set column description
/api/query	POST	Execute SQL query
/api/query/estimate	POST	Estimate query cost

Knowledge Graph ✨

Endpoint	Method	Purpose
/api/relationships	GET	Get all relationships
/api/relationships?table=orders	GET	Get relationships for a table
/api/relationships/path?from=X&to=Y	GET	Find join path between tables
/api/graph/summary	GET	Get graph statistics

Knowledge

Endpoint	Method	Purpose
/api/harvest	POST	Run auto-harvesting
/api/knowledge	GET	Get all knowledge (incl. relationships)
/api/feedback	POST	Record query feedback
/api/metrics	GET	Get query metrics
/api/examples	POST	Add query example
/api/rules	POST	Add business rule

Business Glossary ✨

Endpoint	Method	Purpose
/api/terms	GET	List all business terms
/api/terms	POST	Add a business term
/api/terms/search?q=query	GET	Search terms
/api/terms/generate	POST	Generate glossary with AI
/api/terms/:id	DELETE	Delete a term

See API Reference for full documentation.

---

Related

• @kiyeonjeon21/datacontext — CLI & MCP Server
• DataContext Docs — Full documentation

---

Changelog

See CHANGELOG.md for version history.

---

License

MIT © DataContext