Metadata-Version: 2.4
Name: jedify-lens
Version: 0.1.0
Summary: MCP server for data warehouse schema context generation. Works with any connected database MCP server (Snowflake, BigQuery, Postgres) to generate semantic YAML context.
License: MIT
Keywords: mcp,schema,warehouse,snowflake,bigquery,postgres,context,sql
Author: Jedify
Author-email: adi@jedify.com
Requires-Python: >=3.11,<4.0
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Programming Language :: Python :: 3.14
Requires-Dist: httpx (>=0.27,<0.28)
Requires-Dist: mcp[cli] (>=1.0,<2.0)
Requires-Dist: pydantic (>=2.0,<3.0)
Requires-Dist: pyyaml (>=6.0,<7.0)
Project-URL: Homepage, https://jedify.com/schema-context
Project-URL: Repository, https://github.com/jedify-com/skills
Description-Content-Type: text/markdown

# jedify-lens

MCP server that connects Claude to your data warehouse and generates rich **semantic context YAML** for every table and column — business labels, descriptions, semantic types, and example questions. Powers the [`schema-context`](https://github.com/jedify-com/skills) Claude skill. Standalone — no dependency on Jedify's backend.

## Install

```bash
pip install jedify-lens
```

Or, recommended for MCP servers, run it on demand with uvx:

```bash
uvx jedify-lens
```

## How it works

`jedify-lens` does **not** connect to your warehouse itself, so it needs no warehouse credentials. It works through whatever **database MCP server you already have connected** (Snowflake, BigQuery, PostgreSQL/Redshift): Claude reads your schema and sample rows through that database server, generates the semantic enrichment, and `jedify-lens` writes the structured YAML to disk.

## Tools

- `check_registration_tool` — check sign-in state (call this first)
- `login_tool` — open the Descope sign-up / sign-in page in the browser
- `save_company_context_tool(context)` — save optional company/dataset context to improve enrichment
- `export_context_yaml_tool(enriched_context, output_path, warehouse_type)` — write the schema-context YAML file

## Authentication

Sign-in uses Jedify's Descope inbound app via a public-client OAuth Authorization Code + PKCE flow — there is **no client secret**. The production project is baked in, so end users need no auth configuration. Developers can point at a non-prod project with the `DESCOPE_BASE_URL` and `DESCOPE_CLIENT_ID` environment variables (both are public identifiers).

## License

MIT — [jedify.com](https://jedify.com)

