Spec 24 sprints 3-5: console forecast-sharing UI (clean)#18
Open
ilabra-axo wants to merge 3 commits into
Open
Conversation
apply_publish_shares used `let _ = api...await`, so a failed share/invite during commit (e.g. a transient Railway 502) left the forecast published but the grants silently dropped — the "I shared and nothing happened with no error" failure mode. Now track per-target outcomes, log failures, set publish_status to "Published, but N/M share(s) failed: …", and emit a toast. Success path toasts "Shared with N". Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
There was a problem hiding this comment.
Hi @ilabra-axo! 👋
Your private repo does not have access to Sourcery.
Please upgrade to continue using Sourcery ✨
![ai-coding-guardrails[bot]](https://avatars.githubusercontent.com/in/1202546?s=60&v=4){kind=small}
There was a problem hiding this comment.
Nice work! 😎
I didn't find anything of concern
Skipped files due to being too large for us to review
If you don't need these files reviewed, consider excluding them from the review.
If you do need them reviewed, consider breaking them into smaller files.
Risk: 🟢 Low
Risk analysis
No code changes are present in this pull request, as indicated by zero files changed and zero lines of code changed. Therefore, there is no forward-looking impact on any of the rubric dimensions.
Did you know we can integrate this feedback directly into 50+ IDEs? Get setup in just one command
ai-coding-guardrails
Bot
added
the
zenable/risk:low
ABW is shared substrate, so GET /api/teams returns every vertical's teams — the console was showing rabble swarms, kask workspaces, etc. Scope the Teams panel to fermi_forecast: - console: filter the teams list via is_fermi_team(); tag console-created teams with origin="fermi_forecast" (create_team_handler already accepts it). The filter is strict on `origin` when present, with a slug/description heuristic fallback so it also cleans up against API builds that don't yet return origin. - fermi-auth: get_user_teams / get_team / create_team now SELECT and return teams.origin (added to the Team struct, serde-default). Takes effect for the strict filter on the next ABW deploy; the console heuristic covers the interim. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…asts The cockpit had no view of a forecast's server-side status, so an eliminated/resolved forecast (e.g. a World-Cup team knocked out) still looked editable — you could run a sim and hit save, only for the write to silently fail server-side. The operator shouldn't have to carry "this team is out" in their head. - CockpitState tracks authoritative forecast_status / outcome / resolution note, populated on open_forecast. - reconcile_forecast(): re-GETs the forecast and adopts the settled state + resolved probability. Runs on cockpit re-entry, on a stale-save 409, and via a manual ↻ Reconcile button. - run_simulation / publish_forecast refuse when locked (resolved/void) with the authoritative reason instead of failing opaquely. - A locked banner shows the outcome, resolution note, settled probability, and the Reconcile button. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
ilabra-axo
pushed a commit
that referenced
this pull request
Jun 30, 2026
…cher trajectory events Console-only fixes layered on the WC recompose feature (so they live on the main line, not the isolated sharing PR which lacks that base). #1 sim/save value jump: a just-run sim's raw mean is provisional until the server recomposes mutex-group eliminations. Mark `recomposing`, block save until it settles (so save can't disagree with the sim), and post a "Recomposed: standalone X% → Y% (eliminations priced in)" message. #4 Sobol indices: the sensitivity bars rendered raw p95−p5 spread (≈uniform for a factor model), never the computed Sobol indices. Store per-driver total_order_index each sim and drive the bars from it (relabeled "Driver influence (Sobol total-order)"); spread is now only the pre-sim fallback. #5 trajectory event richness: BayesOps fit events now surface effective-n and 90% CI (collected but previously hidden), so a staged vs auto-accepted decision shows its statistical rationale. Also carries the cockpit reconcile/lock work (already on PR #18) since it shares the same working tree. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Clean replacement for #17, which got polluted: the parallel World-Cup agent pushed its forecast revisions + a
recompose mutex groupfeature onto that shared branch. This branch contains only the console forecast-sharing work — two commits on top ofmain:spec 24 sprints 3-5: console forecast-sharing UI— client wrappers, Teams panel, cockpit + portfolio Access panels, commit-sheet rewrite (drops the dishonest "team" tile), row badges, and the Inbox.fix: surface commit-sheet share failures—apply_publish_sharespreviously swallowed errors (let _ = …await), so a share failing mid-commit (e.g. a transient Railway 502) left the forecast published but grants silently dropped. Now logs failures, sets publish status to "Published, but N/M share(s) failed: …", and toasts.Validation
cargo build -p fermi-consoleclean; client unit tests pass.forecast_acl); manual share create→list→revoke round-trip through live Railway (201/200).🤖 Generated with Claude Code