* feat(config): make `performance_server` as default calculator
* deploy(docker): use osu-performance-server
* docs(readme): add ruleset download instructions
* chore(dev): update development environment
* feat(dev): update development environment setup and service startup order
* fix(deps): move `rosu-pp-py` to `project.optional-dependencies`
* feat(beatmap): handle deleted beatmaps
* feat(performance-server): add a long timeout for calculation
* feat(recalculate): enhance CLI arguments for performance, leaderboard, and rating recalculations with CSV output support
* fix(recalculate): resolve reviews
* Apply suggestions from code review
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* fix(beatmapsync): resolve too long line
---------
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
format: {time:YYYY-MM-DD HH:mm:ss} [{level}] | {name} | {message}
{name} is:
- Uvicorn: log from uvicorn server (#228B22)
- Service: log from class of `app.service` (blue)
- Fetcher: log from fetchers (magenta)
- Task: log from `app.tasks` (#FFD700)
- System: log from `system_logger` (red)
- Normal: log from `log(name)` (#FFC1C1)
- Default: the module name of caller
if you are writing services or tasks, you can just call `logger.`, we will pack it with name `Service` or `Task`
if you want to print fetcher logs, system-related logs, or normal logs, use `logger = (fetcher_logger / system_logger / log)(name)`
New API:
- DELETE /api/private/score/{score_id}: delete a score
- POST /api/private/beatmapsets/{beatmapset_id}/sync: request for syncing a beatmapset
New configuration:
- OLD_SCORE_PROCESSING_MODE
Introduces RateLimitError to manage 429 responses from the API, updating request_api to raise this error and adding handling in prefetch and warmup logic to skip or log when rate limits are hit. Also improves error handling for authentication failures and token expiration.
Simplified the BaseFetcher class by removing the max_retries parameter and all related retry and reauthorization logic. API requests now fail immediately on authentication errors, and token clearing is handled by a new _clear_tokens method. This change streamlines error handling and reduces complexity.
