Feat/av visualizer - #1791
Draft
DAVVVVVVVVVVVID wants to merge 23 commits into
Draft
Conversation
* feat: add per-filter clear button to FilterPill dropdown Each FilterPill dropdown now shows a Clear button at the bottom of the options list. The button is disabled when no options are selected, and clears only that specific filter when clicked without affecting other active filters or closing the dropdown. * feat: replace filter button with clear * feat: replace custom tabler x with mantine X --------- Co-authored-by: Fribbels <fribbelsgithub@gmail.com>
- Timeline rendering with action markers, multi-action badges, and click-to-snap AV selection with hover preview - Intervention engine (simulateTimeline) supporting before/after timing bound to specific character action indices, spd buffs with gauge conservation, and av_advance/av_delay with same-AV tiebreaking - Intervention list panel with structured per-action grouping (during/after zones) and a two-column list+form layout - Memory of Chaos mode: optional 150 AV first row via toggle - 41 unit tests covering the simulation engine Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- New avVisualizerTab i18n namespace (zh_CN + en_US) registered in src/lib/i18n/i18n.ts, with resources.d.ts regenerated via npm run update-resources - Replace all hardcoded strings in AvVisualizerTab, Timeline, ActionMarker, InterventionMarker, InterventionListPanel, and CharacterSlotCard with useTranslation() calls Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Translate all inline/block comments across the tabAvVisualizer module (types, engine, store, controller, timeline components, intervention panel) to English. No behavior changes; vitest (62/62 in this module) and oxlint both pass. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace the dev-time scratch docs (docs/av_dev/: TASK-*.md breakdowns, phase1/phase2 plans, and the original pre-implementation design doc) with a single docs/dev/av-visualizer.md describing the feature as actually shipped: file layout, simulation engine semantics (timing model, gauge conservation, same-AV ordering), Memory of Chaos mode, intervention panel UI, persistence, and i18n — matching the style of the other docs/dev/*.md files. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
New docs/guides/en/av-visualizer.md covering the tab's overview, layout, and a 3-step usage walkthrough (importing/selecting characters, manually adding interventions, and simulating same-AV multi-action / self-advance effects), matching the style of the existing relics-tab.md / warp-planner.md guides. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This flag was accidentally dropped (unrelated to AV visualizer work), which let Node's experimental built-in localStorage partially activate without --localstorage-file, leaving globalThis.localStorage undefined and breaking saveState.test.ts (4 failures). Restoring it fixes all 4; full suite is 924/924 passing. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Redesign the intervention editing workflow from a click-to-open modal into a persistent, video-editor-like layout: - Added: a Playhead that lives on the timeline at all times — click, drag (with magnetic snap to nearby character markers), or left/right arrow keys (snaps to whole AV ticks) move it - Added: ActionDisplayPanel, a always-on panel that shows the action order + during-action / end-of-action-instant intervention lists for whatever AV the Playhead is on, replacing the left half of the old modal - Added: EditPanel, an always-mounted add/edit form driven by a lifted EditRequest, replacing the right half of the old modal; shows a centered hint when idle - Removed: InterventionListPanel modal and MarkerClickContext (clicking a marker/ruler now just seeks the Playhead, nothing else) - UI: active add-zone buttons and the intervention being edited are now highlighted (filled blue / blue border) so it's clear what the EditPanel corresponds to; both panels use Mantine ScrollArea with a fixed height (matching the 2x2 character slot grid) instead of growing with content; layout width matches OptimizerTab's - Fix: dragging the Playhead no longer triggers the browser's native drag-to-select (img draggable=false, preventDefault + user-select: none on the ruler) - Fix: switching the Playhead now clears any in-flight add/edit request, since it's bound to whatever AV was active when created - Persistence: interventions, rowCount, and mocFirstRow now live in AVVisualizerTabSavedSession and persist across reloads like the character slots already did (playheadAv intentionally stays session-only); persistenceService.ts falls back to defaults for saves written before these fields existed - i18n: added AvLabel key for the two "AV" header labels that were previously hardcoded Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add SP/战技点 system: TeamBattleState, sp_gain/sp_loss/sp_cap_up/sp_cap_down intervention types, SpCapChange tracking, simulateBattle.ts with full SP logic - Add SpBar component (segmented horizontal bar) shown above the context panel - Add battle configs registry (Sparkle, Seele, RuanMei, JingYuan, Huohuo) with energy_gain, sp_gain/loss, av_advance, stat_buff, spd_up, spCapBonus templates - Add intervention panel components: ActionConfigPanel, AddBranchPanel, CharacterStatePanel, InterventionEditPanel, UltCasterPanel, UltEffectsPanel - Overhaul AvVisualizerTab layout: right column is now flex-column with SpBar + context panel; playhead-based SP state read from simEvents - Replace EditPanel with InterventionEditPanel; remove EditPanel.tsx - Extend types.ts: InterventionTemplate discriminated union, CharacterBattleConfig, BattleEntity, BattleEvent, UltInsertion, ActiveIntervention, CharacterBattleState, TeamBattleState, TargetType, TurnKind, ActionChoice, ActionNodeOverride - Add simulateBattle.ts and simulateBattle.test.ts (10 SP tracking tests) - Update i18n: avVisualizerTab.yaml with new keys; regenerate resources.d.ts Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- docs/av_dev/: existing AV visualizer dev notes - docs/sparkle_dev/dev_plan.md: 5-step plan for Sparkle B1 config refinement (data-first ordering: optimizer review → boundary analysis → traces → eidolons → full impl) - docs/sparkle_dev/step1.md: deep dive into optimizer's eidolon/trace implementation; AbilityEidolon pattern, trace stat data flow, dynamicConditionals, borrowable patterns - docs/sparkle_dev/step2.md: AV Config system reference + optimizer-to-AV mapping table; identifies 7 errors in current Sparkle.ts and 2 missing system capabilities (battleStart field, factory function + eidolon param) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…ergy bar UI overhaul Engine: - AbilityResolver: abilities can be a function computed from live caster state (energy/buffs), not just static templates — needed for branching kits like Saber's skill. - passiveTrigger (array, function-effect support) and extraAttack: continuously re-checked conditions independent of any specific action, for turn-less one-off bursts (Gilgamesh's extra attack) and self-syncing buffs. - GlobalListener: effect can be an array or a function of the listener context (incl. the new hitCount field), condition/effect both widened for Gilgamesh's hit-count-driven reactions. - clear_buff template type, including its spdBuffs cleanup; buffStacks template condition; energy_set_minimum; everyNOccurrences on listeners. - actionLock: per-character UI choice lock (e.g. Gilgamesh forced to Basic vs Skill depending on Interest stacks). - Fixed: initial energy now keyed off ultThreshold (not maxEnergy) for overflow-style energy caps; SimulationResult.initialActiveInterventions for onBattleStart buffs to show up before any action runs. Characters: - Saber.ts: Reactor Core energy conversion, Mana Burst self-pull, Crown of the Stars, eidolons. - Gilgamesh.ts: Interest/Interest-Piqued action lock, Hero's Solitude, extraAttack shared with Saber's hit count, eidolons. - TrailblazerRemembrance.ts split out, Huohuo/JingYuan eidolon fixes. UI: - EnergyBar/EnergyDisplay/characterEnergyBars: shared energy-bar component with per-character custom variants (Saber's segmented overflow bar, Huohuo's Rangming turn pips). - ActionMarker: Ult/extraAttack get their own half-size icon + a circular- segment overlay badge when they land on the same AV as their own caster's action, using dedicated FUA/ULT artwork. - ActionConfigPanel/ActionDisplayPanel: dynamic-ability effect preview, basicVariants/actionLock-aware choice locking, UltCasterPanel prior-ult state fix when stacking insertions. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
… energy panel drill-down - Extract shared ActionMarker icon/overlay rendering into SpecialActionIcon.tsx - Action-order preview strip now includes Ult/extraAttack as their own slots (avatar + overlay badge), never merged with the character's normal-action slot - Ult/extraAttack cards in the action display panel: border now uses the character's own element color, icon now uses the custom Ult/FUA artwork, and extraAttack cards drop the add/delete affordances it has no use for - CharacterBattleConfig.actionLock now actually drives the engine's resolved action choice (previously UI-only, so a locked Skill silently still resolved as Basic and never cleared its own buffs) - Energy overview panel avatars are clickable to drill into that character's live state, with a back button and auto-revert on Playhead movement - CharacterStatePanel's buff list is now its own scroll region, grouped into turn-based vs stack-based buffs Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
… panel fixes - Intervention/UltInsertion gain afterItemId for exact insertion ordering relative to a specific other item sharing an anchor (not just the old coarse "all interventions before any Ult" default), shared via a new chainOrder.ts helper used by both the engine and ActionDisplayPanel - runPassiveTriggers now also runs right before a normal action's/Ult's own snapshot, so a clear_buff inside that same action can't outrun the passive-synced buff tied to it (previously visible as a stale value in the panel for one checkpoint) - Fix the Ult-card "+" buttons added during the ordering rework dropping afterUltId, which silently broke the UltCasterPanel stale-state guard - Add Wave support (混沌回忆换面): savedSession now holds an array of Waves, each with its own rows/interventions/overrides/Ult insertions; cutting the current wave at the Playhead truncates it, records where via cutoffAv (shown on the timeline as a fixed marker + greyed-out overlay), and seeds a new wave with carried-over energy/buffs/team SP (simulateBattle skips onBattleStart when seeded). Old flat-shape saves migrate automatically. - GlobalActionsPanel: wave indicator/switcher, cut/remove-last-wave actions gated to only make sense from the latest wave Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…ber Reactor Core - simulateBattle now records a per-item state snapshot (chainSnapshots) for every chained Ult AND Intervention, not just Ults — a plain Intervention never produced its own BattleEvent, so the "+" right after one had nothing to look up and silently fell back to this AV's very first state instead of "after everything chained before it" - Fix ActionDisplayPanel's add-button active-highlight key colliding between the leading and trailing "+" in the same before/after chain (both fell back to the same coarse per-action key whenever the clicked button had no preceding chain item to reference) - Saber's enhanced Basic now grants its base-kit +2 Reactor Core unconditionally (previously gated entirely behind E1, and only +1 there) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…s with no ultThreshold/ultEnergyCost set Mirrors simulateBattle.ts's own processUlt cost resolution (customMaxEnergy -> real max_sp -> ultThreshold -> ultEnergyCost) instead of stopping at ultEnergyCost ?? ultThreshold ?? 0 — a character like Trailblazer-Remembrance, which sets neither, still has a real energy cost (their actual max_sp), so the refund was silently computing to 0 instead of 30% of that. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…low maxEnergy
EnergyState's maxEnergy (the hard cap, e.g. Saber's 480/560 overflow mechanic)
was also doubling as the "100%" reference for percent-based effects like
Huohuo's Ult ("restore 20% energy"), inflating the restored amount for any
overflow-style character. Added a separate percentBasis (ultThreshold ??
maxEnergy, same baseline the starting-energy calc already uses) for the
percent math specifically; maxEnergy still governs the actual clamp.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
… the hit counter - onBattleStart now also grants +1 stack of the shared 本王允许你进攻 hit counter (previously only granted Interest), matching its real kit - Add GilgameshEnergyBar (mirrors HuohuoEnergyBar's slot-row pattern): 8 fixed slots below the energy bar, lit per held stack of that same counter Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
… config fixes UI & Layout: - New landing page (ShowcaseTab) with background image, slogan system, scramble text, easter egg glitch effect - Header: custom avatar, updated title, #home routing - Sidebar: zh-only optimization/tools groups, donate entry replacing Ko-fi - Language selector: zh_CN + en_US only, default Chinese - Import tab: simplified to IceDynamix Reliquary Archiver only - Donate tab: WeChat QR + intro text - Favicon and page title updated Battle configs: - Saber E6: globalListener now only counts Saber's own ults (not any ally) toward everyNOccurrences:3 marker - Saber E6: ult always clears E6_ENERGY_MARKER after firing (was missing clear_buff) - Gilgamesh extraAttack: removed Saber's +5 ERR-scaled energy grant Co-Authored-By: Claude Sonnet 4.6 <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.
Pull Request
Description
This is an experimental/draft PR. It hasn't been discussed with the team yet — opening it as a draft so
people can check out the branch and try the feature, to gauge interest before deciding whether to keep
developing it. Not requesting a merge at this stage; will follow up in the
#devDiscord channel.Adds a new AV Visualizer tab for visualizing the relationship between speed and Action Value (AV), to help
plan turn order for up to 4 characters.
Added:
AV = 10000 / speed)simulateTimeline.ts) for speed up/down and advance/delay "interventions", with two timingmodes — during a character's action, and right after it ends — each bound to a specific action index, plus
gauge-conservation math for mid-cycle speed changes and same-AV ordering
preview on the timeline ruler
avVisualizerTabnamespace,zh_CN+en_US)docs/dev/av-visualizer.md) and user guide (docs/guides/en/av-visualizer.md)Fixed:
--no-webstoragetestexecArgvflag invite.config.ts(unrelatedpre-existing regression found while testing this PR; was breaking
saveState.test.ts)Related Issue
deciding whether to pursue it further.
Checklist
Screenshots
![04472ed9b8edf25bda50fc79012c0159]()
-d512-4b47-ba6a-7f03cc8af6eb" />