chore: setup monorepo structure + indexer workspace

[bajtos]

Based on the layout of https://github.com/filecoin-station/spark-stats

Components:

• indexer: background service owning the persisted data, similar to spark-evaluate and spark-observer
• REST API server exposing that data - to be done later

Links:

• Spark v2: DDO, Allocators, Data Quality Improvements space-meridian/roadmap#115
• Lightweight DDO Compatibility: Spark Piece Indexer space-meridian/roadmap#143

[bajtos]

@juliangruber hindsight please 🙏🏻

[juliangruber]

Could we include lint in test? To me, when I run npm test, I expect it to tell me whether CI will pass. This is not the case here. Or do you have a reason for this?

[bajtos]

That makes sense.

However, your proposed change would mean that we execute the linter twice when you run npm test in the monorepo root.

I am using the setup we already have in other places:

• https://github.com/filecoin-station/spark-stats/blob/28fbe3a9b10c907f11052651d0f939f286397eec/observer/package.json#L9-L10
• https://github.com/filecoin-station/spark-api/blob/0d79f8b060aac01d2f8c04d622aed3aded3dc707/api/package.json#L11-L12

What you can do in the monorepo root

• npm test to run all CI checks in all workspaces (directories)
• npm test -w indexer to run only the tests only in the indexer directory
• npm lint -w indexer to run only the linters only in the indexer directory

[juliangruber]

What is ioredis used for here?

[bajtos]

Ah, it's not used here. I want to try using Redis instead of PG, as you did in spark-rewards, and I accidentally added ioredis to dependencies in the wrong PR.