Skip to main content

Documentation index: llms.txt. This page is also available as markdown: append .md to this URL or send Accept: text/markdown.

Castor Extractor

The castor-extractor package is a set of command-line tools that connect to your data sources, read metadata only, and write local JSON or CSV files. You upload those files to Catalog when an integration is client managed: Catalog does not connect to your BI tool or warehouse directly. You control credentials, scheduling, and network access on your side.

Developers is the documentation area for engineers automating Catalog; castor-extractor is the installable Python package documented in this section.

When to Use the Extractor​

Choose client-managed extraction when Catalog cannot reach your source over the network, or when your security policy requires metadata to leave your environment only as files you send.

Typical cases include:

  • BI tools listed as client managed in Quick Set Up (for example during a trial one-time load, or after trial when you schedule daily pushes to a Catalog GCP bucket).
  • Warehouses or other sources where you run extraction on your own hosts instead of granting Catalog a persistent connection.
  • Custom or generic warehouse CSV flows where you build files yourself and validate them before upload.

If Catalog manages the connection for your technology, configure credentials in Settings > Integrations and Catalog runs extraction. You do not need the extractor for that path.

How Extraction Fits in Catalog Setup​

The extractor sits between your systems and Catalog ingestion. Catalog never reads your warehouse or BI tool during client-managed sync; it ingests the files you upload.

In practice, follow this order:

  1. Install the castor-extractor package and the extra for your source (for example [tableau] or [snowflake]). See Installation below.
  2. Extract metadata with the CLI for your technology (for example castor-extract-tableau). Output lands in a directory you choose.
  3. Optionally validate warehouse CSV files with File Checker before upload.
  4. Upload files with castor-upload using your Catalog token and source ID.
  5. Wait for ingestion in Catalog after the upload completes.

Technology-specific credentials, flags, and environment variables are in the BI Tools extraction guides and Warehouse extraction guides.

What Gets Extracted​

The package writes structured metadata files to disk. Asset types depend on the source.

Warehouse Assets​

Warehouse metadata typically includes:

  • databases
  • schemas
  • tables
  • columns
  • queries

Visualization Assets​

Visualization metadata typically includes:

  • dashboards
  • users
  • folders

Knowledge Assets​

Knowledge content can come from tools such as Confluence and Notion when you use those extractors.

The package also includes utilities to push metadata into Catalog:

  • File Checker to validate your Warehouse Importer CSV files before upload
  • castor-upload to push extracted files to Google Cloud Storage (GCS) for Catalog ingestion

Before You Begin​

  • Castor Extractor requires Python 3.10, 3.11, 3.12, or 3.13.
  • A source ID from Catalog for the integration you are loading. The upload CLI flag is --source_id.
  • A Catalog API token for upload. Catalog administrators create tokens in Settings > API. See Getting Your Catalog API Keys for generation and rotation steps.

Installation​

Create an Isolated Environment​

We recommend creating an isolated Python environment. The following example uses pyenv:

brew install pyenv pyenv-virtualenv

pyenv install -v 3.11.9
pyenv virtualenv 3.11.9 castor-env
pyenv shell castor-env
python --version   # should print 3.11.9

If pyenv shell doesn't work, add the following lines to your shell profile and restart your terminal:

eval "$(pyenv init -)"
eval "$(pyenv init --path)"
eval "$(pyenv virtualenv-init -)"

Install the Package​

pip install --upgrade pip
pip install castor-extractor

Most sources need an extra. Install only the ones you use:

pip install castor-extractor[all]
# or only one integration, for example:
pip install castor-extractor[bigquery]
pip install castor-extractor[count]
pip install castor-extractor[databricks]
pip install castor-extractor[glue-athena]
pip install castor-extractor[looker]
pip install castor-extractor[lookerstudio]
pip install castor-extractor[metabase]
pip install castor-extractor[mysql]
pip install castor-extractor[omni]
pip install castor-extractor[powerbi]
pip install castor-extractor[qlik]
pip install castor-extractor[postgres]
pip install castor-extractor[redshift]
pip install castor-extractor[snowflake]
pip install castor-extractor[sqlserver]
pip install castor-extractor[strategy]
pip install castor-extractor[tableau]

Create an Output Directory​

mkdir -p /tmp/castor

Example: Extract Tableau Metadata​

After you install castor-extractor[tableau], run the Tableau extractor from a terminal. The command writes JSON files under your output directory and prints progress in the log.

castor-extract-tableau \
  --username YOUR_TABLEAU_USER \
  --password YOUR_TABLEAU_PASSWORD \
  --server-url https://YOUR_SITE.online.tableau.com \
  --site-id YOUR_SITE_ID \
  --output /tmp/castor

Example log output:

INFO - Logging in using user and password authentication
INFO - Signed into https://eu-west-1a.online.tableau.com as user with id ****
INFO - Extracting USER from Tableau API
INFO - Fetching USER
INFO - Querying all users on site
INFO - Wrote output file: /tmp/castor/1649078755-custom_sql_queries.json
INFO - Wrote output file: /tmp/castor/1649078755-summary.json

For PAT sign-in, flags, and environment variables, see Tableau extraction.

1. Extract Metadata​

You can either use an extractor or add the files to the directory.

Using an Extractor Package​

Using one of the packages, for example, pip install castor-extractor[snowflake] you can extract the data directly and load it into the output directory created during setup to upload to Catalog.

castor-extract-snowflake \
  --account xy12345.eu-west-1 \
  --user svc_castor \
  --password secret \
  --output /tmp/castor

See the full list in Castor Extractor Reference.

Using a Folder​

Place your generic CSV files in the output directory.

2. Validate Data (Optional)​

This step is optional. Run castor-file-check to validate your generic warehouse CSV files before uploading.

castor-file-check --directory /tmp/castor

See Castor Extractor Reference for all castor-file-check flags.

3. Upload to Catalog​

You need a source ID from Catalog and a Catalog API token from Settings > API. See Getting Your Catalog API Keys for token creation and rotation.

castor-upload \
  --token YOUR_CATALOG_TOKEN \
  --source_id YOUR_SOURCE_ID \
  --file_type WAREHOUSE \
  --zone EU \
  --directory_path /tmp/castor

See Castor Extractor Reference for all castor-upload flags and zone selection.

Using Environment Variables​

You can set environment variables for extract and upload commands instead of passing every flag on the command line.

Snowflake extractor example
export CASTOR_SNOWFLAKE_ACCOUNT="xy12345.eu-west-1"
export CASTOR_SNOWFLAKE_USER="svc_castor"
export CASTOR_SNOWFLAKE_PASSWORD="secret"
export CASTOR_OUTPUT_DIRECTORY="/tmp/castor"

castor-extract-snowflake
Castor uploader example
export CASTOR_UPLOADER_TOKEN="your-token"
export CASTOR_UPLOADER_SOURCE_ID="your-source-id"
export CASTOR_UPLOADER_FILE_TYPE="WAREHOUSE"
export CASTOR_UPLOADER_ZONE="US"
export CASTOR_UPLOADER_DIRECTORY_PATH="/tmp/castor"

castor-upload

Troubleshooting​

If you have problems uploading your files, you can increase the timeout or configure retries with environment variables:

  • CASTOR_TIMEOUT_OVERRIDE: seconds before timeout (default 60)
  • CASTOR_RETRY_OVERRIDE: number of retries (default 1)

What's Next?​