mirror of
https://github.com/we-promise/sure.git
synced 2026-09-09 08:34:26 +00:00
210de793c4eacd7f9d8d33664a191ed330c2eb7f
6
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 Nate Mendes
|
ced8d0ccb6 |
fix(insights): project month-to-date spend over the whole month (#3424)
Period.current_month_for returns a range ending today, so Period#days was the same arithmetic as elapsed_days and pace_factor was always 1.0. Month-to-date spend was compared against full baseline months, so an on-pace category read as a large shortfall for most of the month. Derive the month length from the period start instead, which is also correct for families using a custom month start. Adds the generator's first test coverage. |
||
|
Thomas SteiblandJohns
|
06dd14124f |
fix(i18n): localize maintained reserve insight (U5) (#3396)
Co-authored-by: Johns <19662585+Rowdy@users.noreply.github.com> |
||
|
Brandon
|
686205c0ff |
feat(bills): schema and domain core for the bills subsystem (#3201)
* feat(bills): schema and domain core for the bills subsystem First of three chunks carved out of #3083. This one carries the schema and the domain layer: no bills pages, no calendar feed, no assistant tools. Nothing here is reachable from the UI yet, so it changes no user-visible behavior on its own. Schema, in a single migration with a full down: - recurrence_rules, recurring_occurrences, recurring_allocations, recurring_price_changes and recurring_match_rejections - bill columns on recurring_transactions (bill_type, payment_url, autopay, notes, anchor and end conditions, weekend adjustment, dedup scope) - the four data backfills, in their original order Domain layer: - Schedule, the pure date PORO every cadence resolves through, and FrequencyPreset for the labels - OccurrenceGenerator, Matcher, Allocator, PriceChangeDetector, Classifier, DeclaredBill, HistoryBackfiller and PaycheckPlanner - Pipeline, tying detection to generation, plus the nightly job and rake task Existing detection code changed in three places, each a bug this schema exposes: - Cleaner used a flat two-month staleness threshold, which silently retired every quarterly and annual series - SubscriptionAuditGenerator used a flat 45-day overdue threshold, meaningless at both ends of the frequency range - CashFlowWarningGenerator read one projected entry per series, which only equalled the monthly amount because every series was monthly; weekly bills were under-counted fourfold in its 30-day projection The JSON API travels with the model rather than the UI, because the status enum widens here. The API accepts only active and inactive on write; suggested, paused and ended are lifecycle states owned by detection, so the documented enum stays truthful. Uniqueness keys gain dedup_scope alongside amount, never instead of it: a series that is not price-forked carries a blank scope, so amount is what keeps two different prices apart. Suite 7,550 runs, 0 failures. Rubocop and brakeman clean. Eager loading verified, and the migration reverses and re-applies. Includes the first review round: orphan repair matches income and refuses coincidental twins, session imports persist occurrence mappings across chunks, semimonthly anchors canonicalize, classifier keywords match whole words, and the down refuses rather than failing when price-forked rows exist. * Address second review round Bound the cross-currency default allocation by the entry leftover and the occurrence remainder, matching the same-currency path. Let keyword stems carry a suffix again after the word-boundary fix silenced them. Skip an incoherent recurrence rule row instead of rolling back the whole import. Check rollback collisions per restored index so a refusal cannot land after the bills tables are dropped. Replay the closed_at test through a real second import. Preload the orphan repair associations and move the allocator errors to locale keys. * Match index NULL semantics in the rollback collision checks GROUP BY treats NULLs as equal but the restored unique indexes do not: account_id is nullable and indexed, so two accountless rows can never collide under any of them. Excluding NULL accounts keeps the guard from refusing a rollback PostgreSQL can perform. Verified live both ways: accountless duplicates roll back, a real collision still refuses. * Address maintainer review Scope the payable debt-destination subquery to the row and its family instead of scanning every account in the installation. Batch the cash flow generator remaining-amount sums into one grouped query, matching the two sibling sites. Enforce both window bounds in the after_count branch so a future-anchored plan cannot leak past the requested end date. Skip the explicit regeneration when the day column change will fire the model callback anyway. Add the missing locale entry for the allocation currency validation. |
||
|
buzzromainandClaude Opus 5
|
90203139bf |
feat(goals): tell the user when a reserve has dropped below its level (#3175)
* feat(goals): reserves you maintain, not goals you finish An emergency fund is not a goal you reach and close — it is a level you hold, and every withdrawal is a shortfall to make good. Sure treated it like anything else: at 100% it offered to close it, which would release the very money being set aside; a withdrawal dropped the bar with no sign that anything was owed. `kind` (added by the lifecycle lot without behavior) now means something. A maintained goal is `funded` or `depleted`, never `reached` — sitting at its floor is a steady state, not an achievement to file away. `complete` is refused by an AASM guard rather than merely hidden, so no path can release a reserve's earmark. Two ordering traps, both of which would have made a drained reserve invisible: `ACTIVE_DISPLAY_STATUS_RANK` falls back to 4 for any status it does not know, so an unranked `:depleted` would sort a drained emergency fund below everything else — the exact opposite of what it means. It ranks alongside `:behind` now, and `:funded` sorts near the end with the goals that need nothing. `behind_pace?` excludes reserves. `monthly_target_amount` and `pace` both derive from `target_date`, which a reserve does not have, so "save X/month to catch up" would be advice about a deadline that does not exist. The form leads with the choice, since it changes what the rest of it means, and hides the target date for a reserve rather than disabling it — a hidden field cannot submit a stale value that would then drive a pace. The card states the shortfall, which is exactly `remaining_amount`. The panel that offers a one-off its closing action tells a reserve it is intact and offers nothing, because there is nothing to do. Scope: fixed targets only. Targets expressed in months of expenses, the monthly refresh job, and the depletion insight are the next two PRs. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DJ1npaGEHr6t2HW1rYZdt4 * fix(goals): let a reserve behave like one everywhere it is shown Addresses review feedback on #3167. The kind selector never hid the target date. `data-controller="goal-kind"` sat on the selector div while its `dateField` target is a sibling, so `dateFieldTargets` came back empty and picking "Reserve to maintain" left the deadline on screen and submittable. The controller moves to the form wrapper, which encloses both. Hiding a field is not enforcement, so the model now clears `target_date` for a maintained goal. Normalising rather than rejecting: the field is hidden, and an error about something the user cannot see is not actionable. A date could only arrive through a conversion or a crafted request, and either way a stored deadline would drive a pace the reserve does not have. A completed goal could be switched to `maintained` from the edit form. It then sat in a released state — one that has handed its earmark back — while the show page promised its money stays reserved, and `complete` for reserves is refused precisely to prevent that state. `kind` is now locked while released: reopen first. Reserves counted against the "goals on track" tile. Their statuses are `funded`/`depleted`, which match none of the exclusions in `tracked_total`, so they could never reach the numerator and a family with one reserve read "0 of 1 on track" for a goal working exactly as intended. Two more places still spoke of pace to something that has none. `pace_line` is suppressed for reserves on the card, and a depleted reserve gets its own panel before the projection card — the projection's summary, catch-up line and colour are all built from a deadline. What a drained reserve needs is the number the projection cannot show: how much is missing from the floor. The French celebration copy read "Votre réserve est à son niveau", which never says which level. Each guard was confirmed load-bearing by removing it and watching its test fail. bin/rails test: 6962 runs, 28003 assertions, 0 failures. RuboCop, erb_lint and Brakeman clean. Left open deliberately: extracting the show page's lifecycle panel into a ViewComponent. The guideline behind it is right, but the refactor is wider than this round of fixes and belongs on its own. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016GTNba5qE5NwzaHzbp27ye * fix(goals): finish teaching the status consumers about reserves Second round of review feedback on #3167: two consumers still had no branch for the reserve statuses. `ProgressRingComponent#percent_text_class` styled only `:reached` as success, so a funded reserve — a floor the user is holding exactly as intended — fell back to the neutral colour and read as unfinished. `status_callout_context` had no `:depleted` branch, so a drained reserve showed no callout at all: the one status that most deserves a line of explanation was the only one saying nothing. It now names the shortfall. `:funded` deliberately keeps no callout — a reserve at its level has nothing to report, and the celebration panel already says so. A test pins that, so the silence reads as a decision rather than another missing branch. bin/rails test: 6964 runs, 28008 assertions, 0 failures. RuboCop and Brakeman clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016GTNba5qE5NwzaHzbp27ye * fix(goals): lock the kind on the state the goal is actually in Addresses review feedback on #3167, on the guard that landed in c1c7f4f7. `kind_locked_while_released` read the in-memory `state`, so a single write setting `state: "active"` alongside the new kind saw the goal as already reopened and waved it through. The end state looks legitimate — active and maintained — which is why the hole is easy to miss. It is not: the direct write skipped the `reopen` transition, and with it `thaw_completed_amount!`. `completed_amount` survived, so `current_balance` returned that frozen snapshot forever on a live reserve. Reopening has to be its own gesture, because it is the gesture that thaws. Now reads `state_in_database`, with a regression test on the combined write asserting both that it is refused and that the frozen amount is untouched. Confirmed load-bearing by reading the attribute again and watching it fail. bin/rails test: 6969 runs, 28019 assertions, 0 failures. RuboCop and Brakeman clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016GTNba5qE5NwzaHzbp27ye * fix(goals): send an empty reserve to the shortfall panel, not the empty state Addresses the last review thread on #3167. The `maintained?` branch sat after the zero-balance/zero-pace one, so a brand-new reserve matched the generic "make your first transfer" card. I had put it there on purpose, thinking a reserve with nothing in it wanted the first-transfer nudge. The review is right that it does not: it is still a reserve short of its floor, and the shortfall panel says so with the saved, target and missing amounts, where the generic card says none of them. Ordering it after also meant evaluating `pace` on a goal that has no pace to evaluate. bin/rails test: 6965 runs, 0 failures. RuboCop and erb_lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016GTNba5qE5NwzaHzbp27ye * fix(goals): stop a paused goal outranking a reserve that is whole Review on #3175. `:funded` and paused both ranked 3 in `active_display_sort`, so the tie broke on name and a paused goal called "Alpha" sat above a reserve called "Zeta" that was fully funded — the list saying the paused one wanted attention more. Paused now ranks behind every status, which is what the comment above the table already claimed. The seven panels on the goal page were hand-rolled repetitions of `DS::Card`'s exact shell, two of them adjacent and identical. They render through the primitive now, so their surface styling cannot drift apart. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016GTNba5qE5NwzaHzbp27ye * refactor(goals): move the lifecycle panel decision out of the template Review on #3167 and #3180. Which panel a goal gets is a lifecycle question with five answers, and the template worked it out inline from `completed?`, `maintained?`, `one_off?`, `status` and `may_complete?` — five predicates deep in ERB where the ordering between them was load-bearing and nothing said so. `Goals::LifecyclePanelComponent` answers it in Ruby and the template renders the answer. The markup moves across unchanged, keys made absolute because a relative `t(".x")` in a component resolves against the component's own path rather than the page these strings belong to. The order is now stated once, where it can be read and tested: `:reserve_shortfall` before `:empty`, because a brand-new reserve sits at zero balance and zero pace and the generic "make your first transfer" card would otherwise swallow it. Closing from the panel now confirms, as the header menu already did. Completing releases the goal's earmarked money, and the panel offered that in one click. Both go through `goal_complete_confirm` rather than building the wording twice — two copies drifting apart is how one ends up describing the wrong consequence. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016GTNba5qE5NwzaHzbp27ye * fix(goals): refresh the pace suggestion when the deadline is cleared Assigning `input.value = ""` fires no event, so `goal-form#suggestedChanged` never ran: selecting "Reserve to maintain" cleared the date but left the monthly pace suggestion on screen, derived from a deadline the goal no longer has. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016GTNba5qE5NwzaHzbp27ye * fix(goals): let a depleted reserve look as urgent as it is Review on #3179 and #3180. `Goal#needs_attention?` names the pair of statuses that mean "this one wants looking at" — a goal off its pace and a reserve below its floor. Three places were spelling that out and the Plan hub's progress bar had fallen behind, so a depleted reserve got a neutral bar an inch from its own amber status pill: the same goal reported as needing attention and not. `projection_summary` told a funded reserve it had "hit the target, no projection needed". A reserve holds a level; there is no finish line to project toward and no target to have hit. It does not reach that panel today — the shortfall and celebration panels catch it first — but the method reads as the single source of truth for that subtitle and should not hand a caller a one-off's wording. The legend swatches are bordered spans now rather than inline SVG, and the label takes `text-xs` instead of an arbitrary 11px. The projection swatch keeps the chart's own colour variables in an inline style rather than `border-success` / `border-warning`: the chart hard-codes green-600 and yellow-600, and a legend whose colour does not match the line it describes is worse than the markup it would save. The Plan-card test counts warning bars rather than matching one. A fixture goal is already off its pace, so the markup is on the page either way and a presence check passed without the fix. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016GTNba5qE5NwzaHzbp27ye * feat(goals): tell the user when a reserve has dropped below its level A maintained reserve exists to hold a floor. Nothing said when it stopped holding it: the goal page shows the shortfall, but only to someone who thought to look, and a reserve is precisely the thing you stop looking at once it is full. That silence is the gap this closes. High priority, unlike most of this feed. `IdleCashGenerator` is a nudge about money doing nothing; this is the opposite — a floor the user deliberately set is no longer there. It is also the one signal a reserve can produce that a one-off goal cannot, which is what makes it worth a generator of its own. Three deliberate limits: - **Active reserves only.** A paused one is shelved on purpose, and `behind_pace?` already excludes paused goals for the same reason. Nagging about a goal someone put down is noise. - **The dedup key rotates monthly.** A reserve can sit short for weeks while it is rebuilt, and re-raising the same shortfall every night trains people to dismiss the feed. - **Two at a time, worst shortfall first.** A family running four drained reserves has one problem, not four. Loaded through `Goal.prepared_for` so the family-wide pooled allocations are read once: asking each reserve for its status reaches `current_balance`, and without that injection every one of them would re-read the whole pool. bin/rails test: 7079 runs, 28491 assertions, 0 failures. RuboCop and Brakeman clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016GTNba5qE5NwzaHzbp27ye * fix(insights): stop an insight naming something that has been renamed Review on #3175: renaming a depleted reserve leaves the stored title naming the old one for the rest of the month, because the name is not part of the metadata that drives a refresh. Rather than making the name material, the same-signal branch now refreshes the title alongside the facts. The title is built from I18n and the generator's own data — the model writes the body, not this — so keeping it current costs nothing, and a rename is not a reason to resurface an insight the user has already read. The body still says the old name until the numbers move. Forcing an LLM rewrite on every rename is the wrong trade, and making the name material would do that *and* re-nag the user. This is generic to every generator whose title embeds a name, not only the reserve one. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016GTNba5qE5NwzaHzbp27ye * fix(insights): refresh the title on reactivation too A gap in the previous commit: the expired-and-returned branch resurfaces an insight without touching its title, so a subject renamed while the insight was expired came back naming the old one. Same reasoning as the same-signal branch — the title is I18n plus the generator's own data, not the model's prose, so keeping it current costs nothing. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016GTNba5qE5NwzaHzbp27ye --------- Co-authored-by: Claude Opus 5 <noreply@anthropic.com> |
||
|
Guillem Arias Fauste
|
59852ca0f3 |
fix(insights): respect recurring_transactions_disabled in subscription_audit (#2831)
* fix(insights): respect recurring_transactions_disabled in subscription_audit SubscriptionAuditGenerator queried family.recurring_transactions directly, so disabling recurring-transaction detection (Settings -> Recurring Transactions) never stopped already-identified rows from surfacing "recurring charge overdue" insights on the Insights feed — the family-wide flag was already checked at both call sites in IdentifyRecurringTransactionsJob, just not here. Returning [] early is enough for existing insights to self-clean up: produced_types is a class-level declaration, so GenerateInsightsJob still counts subscription_audit as a succeeded type and expires any insight whose dedup_key wasn't regenerated on the next nightly run. * docs(insights): document why cash_flow_warning skips the recurring-disabled guard Answers jjmata's open review question. Unlike SubscriptionAuditGenerator, recurring transactions here are one input into a broader cash-flow projection, not the insight's entire subject — so it intentionally keeps using the last-known identified set rather than gating on family.recurring_transactions_disabled?. No behavior change. |
||
|
|
66cf9e7f0b |
feat(insights): proactive financial intelligence feed (#2550)
* feat(insights): proactive financial intelligence feed Adds a nightly job that analyzes each family's finances in pure Ruby and surfaces typed, stateful insights on the dashboard and a new /insights page. - Insight model (active/read/dismissed) with a per-family dedup_key unique index so nightly re-runs refresh rows instead of duplicating them - Seven generators (spending anomaly, cash-flow warning, net worth milestone, subscription audit, savings rate change, idle cash, budget health) built on IncomeStatement, BalanceSheet, RecurringTransaction, and BudgetCategory - LLM used as a writer, not a reasoner: Insight::BodyWriter narrates pre-computed facts via the configured provider, with an i18n template fallback so self-hosted installs without API keys work identically; bodies are only (re)written when an insight is new or its numbers changed - GenerateInsightsJob: cron fan-out per family, per-family advisory lock, metadata-diff upsert that preserves read/dismissed state for unchanged signals and reactivates on material change - Dashboard insights_feed section (top 3, collapsible/reorderable) and an /insights feed page with turbo-stream dismissal; viewing the feed marks insights read - Tests for the model, job upsert semantics, and controller flows Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Y6ZRA6cgCRM4UdFKct3wm4 * fix(insights): address Codex review — expiry, budget metadata, LLM usage - Expire visible insights whose condition cleared: generators declare the insight types they produce, the registry reports which generators ran to completion, and the job expires visible insights of those types whose dedup_key was not regenerated. A crashing generator can't wipe out its healthy insights, and an expired insight reactivates when its condition returns — unlike a user-dismissed one, which stays dismissed. - Include a bucketed budget-spent percent in budget_at_risk metadata so the body refreshes when overall usage moves >=10 points even if the same categories remain flagged. - Pass family to chat_response so LLM narration is recorded in llm_usages. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Y6ZRA6cgCRM4UdFKct3wm4 * fix(insights): address CodeRabbit review - Skip the mark-as-read write for Turbo hover-prefetch requests (X-Sec-Purpose) so unread badges don't clear before a real visit - Show the New pill on the dashboard feed (active = unread there; the feed never marks insights read) - Filter idle accounts in SQL instead of a per-account exists? loop - Eager-load merchants in the subscription audit query - Widen the advisory-lock key to the signed-bigint range and log when acquisition fails so a skipped nightly run is observable - Mirror the dedup_key unique index as a model validation - Assert the refresh action enqueues for the signed-in family Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Y6ZRA6cgCRM4UdFKct3wm4 * fix(insights): clear stale lifecycle timestamps on reactivation When an insight resurfaces, the row now leaves no contradictory state behind: the material-change path clears both read_at and dismissed_at, and the expired-recovery path clears read_at. Tests assert the contract, and the dashboard feed test now also locks in the unread "New" badge. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Y6ZRA6cgCRM4UdFKct3wm4 * fix(insights): drop redundant standalone family_id index Every composite index on insights already leads with family_id, so the auto-created single-column index was pure write overhead. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Y6ZRA6cgCRM4UdFKct3wm4 * fix(insights): stop nightly drift from resurrecting dismissed insights Three generators stored continuously drifting values (projected amounts, starting balance, current net worth) in metadata, and any metadata diff counts as a material change: the body is rewritten (an LLM call when a provider is configured) and read/dismissed state is cleared. Dismissing a cash-flow warning was undone at the next 6:00 run. Metadata now carries only the signal identity plus a coarse bucket, the same damping the budget generators already use; display values live in facts. Also orders the idle-cash pick by balance so the selected accounts don't flip between runs, and documents the dollar-scale threshold assumption on the relevant constants. * fix(ds): map DS::Button aria_label into the aria hash A bare aria_label: option reaches the tag helpers as a literal aria_label attribute (Rails only dasherizes the nested aria: hash), so the icon-only fallback overrode it and screen readers announced the insight dismiss button and the popover trigger as "X". * fix(insights): gate LLM narration behind AI consent The nightly job runs unprompted for every family, so narration now requires someone in the family to have AI enabled: consent to share financial facts with the provider, and a cost cap in managed mode. The template fallback keeps behavior identical otherwise. Narration failures are captured via DebugLogEntry so support can see them. Adds BodyWriter coverage, including a template-interpolation test for every generator template key. * feat(insights): rework the dashboard feed The feed rendered full insight cards inside the section shell — the only widget nesting card-on-card — and appended below the fold for every family with a saved section order. It now mirrors the outflows and balance-sheet list idiom (inset well with an uppercase mini-header, white row block, 28px sentiment-tinted icon circles via color-mix on the DS CSS variables, right-aligned key figures) and leads the dashboard for saved orders that predate it. Icon color comes from sentiment, not priority — a savings-rate improvement is high priority AND good news, and must not render red; red is reserved for a projected-negative balance. Rows have no hover wash (cursor plus a gentle icon scale, like the sibling widgets), links to /insights disable Turbo prefetch so the mark-as-read actually fires for mouse users, and the standard widget-size popover offers Half/Full. Full stays the default: the feed is far shorter than any other single-width widget, so a half default leaves a grid hole the masonry cannot backfill. * feat(insights): actionable cards, dismiss undo, live refresh Each row now persists its display facts (new jsonb column) alongside the change-detection metadata. Facts refresh every run without touching the body or user state, which is exactly why they are not part of the material-change comparison. The card gains what the stored data always supported: a type-and-period meta line replacing "x minutes ago", a right-aligned key figure (green only for good news), and a contextual link resolved from the subject ids in metadata — category, account, recurring transaction, budget month — omitted when the subject no longer exists. Dismiss is forgiving: a toast in the notification tray offers undo, and undismissing restores the row as read rather than re-badging it. Milestone insights can never regenerate once dismissed, so this closes a real loss path. Upsert failures are captured via DebugLogEntry. Manual refresh gets feedback: the button swaps to a disabled checking state, the page subscribes to a family-scoped stream, and the job broadcasts the refreshed list and the idle button when it finishes (also after a lock-skipped run, so the button cannot stay stuck). Savings copy is sign-aware: a negative rate reads "you spent more than you earned" with true minus signs. The empty state swaps Lucide sparkles for the brand assistant glyph (DS::EmptyState learns icon_custom:). * feat(insights): top-bar entry with unread count The dashboard feed hides at zero insights and nothing else linked to /insights, so the manual refresh (and the empty state) were unreachable for exactly the families who need them: fresh setups before the first 6:00 run. The sticky top bar travels to every screen and its right cluster had room, so insights get an icon entry there with a monochrome unread badge. Prefetch is disabled on the link for the same mark-as-read reason as the feed. --------- Signed-off-by: Juan José Mata <juanjo.mata@gmail.com> Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: Guillem Arias <accounts@gariasf.com> |