changed dispatcher port to 5001

[pradyumna-rfai]

Changed Evals Dispatcher port to 5001 to avoid conflicts

---

Note

Standardizes service ports to 8851/8852/8853 and replaces hardcoded URLs with config classes, updating startup scripts, workflows, frontend proxies/tests, and documentation accordingly.

• Config & Backends:
  • Introduce/configure DispatcherConfig, MLFlowConfig, FrontendConfig across fit/evals; replace hardcoded URLs with *.URL.
  • Default ports unified: dispatcher 8851, mlflow 8852, frontend 8853; CORS origins updated accordingly.
  • Evals dispatcher defaults to 8851; adds host-aware port checks and minor CORS/formatting tweaks.
• Startup Scripts (setup/fit/start.sh, setup/fit/start_dev.sh):
  • Switch to new ports; add RF_LOG_PATH and route logs there; improve readiness checks/status; cleanup and environment handling.
• Frontend:
  • Dev server on 8853 (craco.config.js), Flask server updated; proxies (setupProxy.js, proxy_middleware.py) target new config URLs.
  • Tests updated to new URLs.
• Workflows:
  • Run version bump via setup/bump_version.sh; manual deploy outputs new_numeric_version and echoes version-specific PyPI URL.
• Library/CLI:
  • pyproject.toml: relax/adjust dependency versions; add project URLs and package data.
  • cli.install_packages: add faiss-gpu-cu12 for CUDA 12 evals.
• Docs & Notebooks:
  • Update all port references and image paths; README/troubleshooting commands reflect new ports.

^{Written by Cursor Bugbot for commit 1f8e174. This will update automatically on new commits. Configure here.}

[cursor]

This PR is being reviewed by Cursor Bugbot

Details

Your team is on the Bugbot Free tier. On this plan, Bugbot will review limited PRs each billing cycle for each member of your team.

To receive Bugbot reviews on all of your PRs, visit the Cursor dashboard to activate Pro and start your 14-day free trial.

[cursor]

Bug: Port Conflict: Dispatcher Ports Mismatched, Needs Alignment

Port conflict: The evals dispatcher is configured to use port 8851, which is the same port used by the fit dispatcher (defined in rapidfireai/fit/utils/constants.py line 48). This creates a port conflict when both systems run simultaneously. According to the PR description, the evals dispatcher should be changed to port 5001 to avoid conflicts, but it was incorrectly changed to 8851 instead.

 

[cursor]

Bug: Documentation Image Loading Failure

Typo in image URL path contains imagesmages instead of images, causing the documentation button image to fail loading. The URL should be docs/images/documentation-button.svg not docs/imagesmages/documentation-button.svg.