Separate channeldata generation into a separate process.

[dholth]

Checklist

• I added a descriptive title
• I searched open requests and couldn't find a duplicate

What is the idea?

It should be possible to generate channeldata from an output tree by reading in all the existing */repodata.json, and combining them. Instead of having to generate them at the same time as repodata.json.

We may or may not currently save some time parsing repodata.json into channeldata by starting with the already-parsed version, but it would be nice to be able to generate the not-often-used channeldata at a different frequency than the always-used repodata.

Why is this needed?

This would give us more flexibility when generating channeldata and repodata.

What should happen?

No response

Additional Context

No response