GitHub Search should return similar output to HTTP request (body w/ object)

[akeller]

Feature description

Currently GitHub Search returns .ion files which make it difficult to access the returned JSON object for quick and easy use (need to convert to .json, get a .jsonl, then need to parse the file, etc.). Returning the object in a body (like we do with HTTP request) would give me the ability to use pebble expressions to immediately work through what GitHub returns.

Returning a .ion file might be a great option, but by default we should have a similar experience to HTTP request.

[wrussell1999]

Did some more digging into these tasks. Perhaps we should introduce the fetchType property we often use with DB tasks or other data fetching tasks where we can define if we want it as a file in local storage or as an output. This could be good as sometimes the response body can be large so processing it through outputs could potentially cause performance impacts when using the Outputs tab.

See this Kubernetes example: https://kestra.io/plugins/plugin-kubernetes/kubectl/io.kestra.plugin.kubernetes.kubectl.get#properties_fetchType-body

or DuckDB: https://kestra.io/plugins/plugin-jdbc-duckdb/io.kestra.plugin.jdbc.duckdb.query#properties_fetchType-body