migrate-meltano-connector

Context: This plugin is for the Fivetran Connector SDK (CSDK). "CSDK" is shorthand for "Connector SDK".

Migrate a Meltano Extractor or Singer Tap to CSDK

FIRST: Read sdk-reference.md from the plugin directory to load SDK rules, operation names, schema constraints, configuration rules, and testing requirements.

Use this workflow to convert Meltano extractor logic into a Connector SDK project. Meltano migration references:

• Meltano project/config docs: https://docs.meltano.com/concepts/project/
• Meltano plugin docs: https://docs.meltano.com/concepts/plugins/
• Singer spec reference: https://hub.meltano.com/singer/spec/
• Meltano Singer SDK docs: https://sdk.meltano.com/
• Fivetran CLI for pipeline resources: https://pypi.org/project/fivetran-cli/

When validating the migration workflow itself, use representative public projects such as Matatika/example-github-analytics, which exercises:

• A Meltano project with tap-github, target-postgres, dbt transforms, a utility plugin, a job, and a schedule.
• Selector rules like *!.* followed by explicitly selected streams.
• Array settings such as repositories.
• A third-party Singer SDK tap whose source code must be consulted for schemas, primary keys, parent-child stream context, pagination, and state behavior.

Step 1: Locate the Source and Target

If the user did not provide a source path, ask for one of:

• A Meltano project directory containing meltano.yml.
• A custom Singer tap package or local tap source directory.
• Singer artifacts: config.json, catalog.json or properties.json, state.json, and sample output.

Find or create the target CSDK project:

• If the current directory already has connector.py, configuration.json, and requirements.txt, migrate in place.
• If no CSDK project exists, ask for the target directory name, then scaffold it with fivetran init before editing files.
• Do not overwrite unrelated files.

Scope the migration:

• Migrate Meltano extractors/Singer taps only.
• Do not port Meltano loaders/targets, dbt transforms, schedules, orchestration, or environments into CSDK connector code.
• For pipeline resources that correspond to Fivetran platform objects, create a follow-up plan using fivetran-cli instead of embedding that behavior in connector.py.
• Use loaders, transforms, schedules, and environments as context for what the extractor produced, how it was configured, and which Fivetran CLI resources may need to be created or updated.