[aaronbrethorst]

Part of #1977

Draft. Stacked on #1776. Draft until multi-feed phases B and C land.

This PR began as the whole import draft. It's now one layer of a stack:

PR	Change	Base
#1771	Parser	main
#1772	Validator	#1771
#1774	Apply	#1772
#1775	Remove consolidated stops while importing	#1677 plus cherry-picks of the three above
#1776	Apply the file on static import	#1775
#1773	Config option, plus the plan doc	#1776
#1777	Remove trips consolidation leaves too short	#1773

What it does

• gtfs-static-feed.stop-consolidation-url names a Java OneBusAway stop consolidation file, as an http(s) URL or a local path. It's the counterpart of STOP_CONSOLIDATION_URL in OneBusAway's Docker bundle builder.
• It's validated like gtfs-static-feed.url, which now shares the check: no file:// URLs and no path traversal.
• -dump-config shows it when set.
• Also adds docs/stop-consolidation-plan.md, which covers:
  • the settled behavior, and the production Puget Sound evidence behind it
  • what's done
  • the remaining PRs, in order

Root cause

Exposes #1776's Config.StopConsolidationURL to operators.

Testing

• Test failing on main before this change: n/a (new option).
• TestToGtfsConfigData_StopConsolidationURL, TestValidate_StopConsolidationURL and TestStopConsolidationURLReachesGTFSConfigAndDump.
• Manual check: maglev -f config.json -dump-config shows stop-consolidation-url, and the server applies the file (results in Apply a stop consolidation file on static import #1776).
• make test and go vet (both tag sets) pass.

Checklist

• This PR addresses one issue. (no issue filed yet)
• It is under ~200 changed lines, or the description says why it can't be split.
• This PR is not a bug fix, or the new or changed test fails on main and passes on this branch.
• Each commit is one logical change with a clear message.
• I did not hand-edit sqlc-generated files under gtfsdb/ or testdata/openapi.yml.
• This PR changes no handler.
• This PR doesn't change a Java-served response.
• This PR adds no endpoint.
• I read every line of this diff and can explain it if asked.

[coderabbitai]

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

• Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

Share

• X
• Mastodon
• Reddit
• LinkedIn

[github-actions]

The description follows the PR template.

[github-actions]

Performance Smoke Test Results

Status: PASSED

Metric	Value
p(95) latency	2.1 ms
Error rate	0.00%
Total requests	344
Req/sec	11.3

Smoke test config: 5 VUs x 30s. Thresholds: p(95) < 300ms, error rate < 1%.

Full results uploaded as workflow artifact: k6-smoke-summary.

[github-actions]

Performance Smoke Test Results

Status: PASSED

Metric	Value
p(95) latency	1.9 ms
Error rate	0.00%
Total requests	336
Req/sec	11.0

Smoke test config: 5 VUs x 30s. Thresholds: p(95) < 300ms, error rate < 1%.

Full results uploaded as workflow artifact: k6-smoke-summary.

[github-actions]

Performance Smoke Test Results

Status: PASSED

Metric	Value
p(95) latency	2.1 ms
Error rate	0.00%
Total requests	324
Req/sec	10.6

Smoke test config: 5 VUs x 30s. Thresholds: p(95) < 300ms, error rate < 1%.

Full results uploaded as workflow artifact: k6-smoke-summary.

[burma-shave]

Reviewed at 7faee60. The config plumbing is fine, and pulling the checks into validateFeedLocation is a real improvement: the old code returned early for an HTTP static URL, so without the refactor the new field would have skipped path-traversal validation. One scope comment inline about the plan doc.

Minor: there's no -stop-consolidation-url CLI flag to go with -gtfs-url. Fine if JSON-only is intended; worth a line in the description if so.

---

Generated by Claude Code

[burma-shave]

This plan doc is a separate concern from the config option and will go stale as the stack lands (#1777 and #1808 already edit it). The behaviour table is worth keeping, but I'd move it to the tracking issue #1977 or the wiki's Implementation Decisions (which the doc itself says should record these), and drop this commit from the PR.

---

Generated by Claude Code

[aaronbrethorst]

Agreed. The behavior table, production evidence, token-agency rule and open items now live in §9.4 of the multi-feed spec, and stop.md has an Implementation Decisions note for the absorbed-ID rule. I dropped the plan-doc commit here, and its follow-up edit in #1777, from the stack.

[aaronbrethorst]

Also added the -stop-consolidation-url flag next to -gtfs-url in 538aa41. It goes through the same validation as the JSON key, so file:// is still rejected.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[github-actions]

Performance Smoke Test Results

Status: PASSED

Metric	Value
p(95) latency	2.1 ms
Error rate	0.00%
Total requests	340
Req/sec	11.2

Smoke test config: 5 VUs x 30s. Thresholds: p(95) < 300ms, error rate < 1%.

Full results uploaded as workflow artifact: k6-smoke-summary.

[github-actions]

Performance Smoke Test Results

Status: PASSED

Metric	Value
p(95) latency	1.4 ms
Error rate	0.00%
Total requests	331
Req/sec	10.9

Smoke test config: 5 VUs x 30s. Thresholds: p(95) < 300ms, error rate < 1%.

Full results uploaded as workflow artifact: k6-smoke-summary.