Use the local project id when the cloud project is unavailable

[AlexCatarino]

Problem

A local run sends the project's cloud-id as its project id even when that cloud project can no longer be read. It surfaced through the Auth0 authorize URL of lean live deploy, which pointed at a project that was no longer there:

https://www.quantconnect.com/api/v2/live/auth0/authorize?brokerage=alpaca&projectId=35623870

Fix

Look the project up before using its cloud-id, and fall back to the negative local id when it cannot be reached, as a project that was never pushed already does.

project state	projects/read	resolved id
exists	200	positive cloud-id
purged, or not reachable by this user	RequestFailedError	-local_id

A project sitting in the Recycle Bin still reads back normally, so it keeps its cloud-id. That is deliberate: it can be moved back out, and the id stays valid.

Notes

• This adds one projects/read call to every local run of a pushed project: backtest, research, optimize, local live deploy and private_cloud start. data download is unaffected, it has no project directory and returns -1 before reaching the lookup.
• Offline, a pushed project resolves to -local_id instead of failing, since the fallback covers connection errors along with everything else. The reason is logged at debug level.
• Nothing is cached, so the id follows the project's current state on every run.

Testing

• tests/components/config: 101 passing, with new cases for a project that exists, one that cannot be read, and one that was never pushed.
• Verified against the live API: an unreachable id raised Project not found and fell back to the local id.

🤖 Generated with Claude Code