Files
interactive-story/frontend/src/pages/Play/index.jsx
T
JesseMarkowitzandClaude Opus 5 1ce9972760 M8: the browser becomes the storyteller
The interface was AI-DnD's with this product's features bolted into it. The
navigation read Home · Adventures · Scenarios · Settings · AI Chat; starting a
story meant first picking a *world*, and making a world meant a JSON stat-schema
form, a story-card table and an art picker. The play screen had a Branches tab.
The input had three modes. Sixteen of the sixteen controls on a two-turn story
had no accessible name — they were single glyphs with a tooltip.

All of that was measured in a real browser before anything was changed, and the
measurements are in planning/reports/M8-IMPLEMENTATION-REPORT.md §C. Almost
nothing underneath was wrong: the play loop, the history controls, the takes,
the Save Points, the state correction and the knowledge library all worked. What
was wrong was what a reader was asked to understand in order to use them.

So the shape now is one entry point and one screen:

  Campaigns -> Campaign -> Story
                           State · Knowledge · Context · Save Points · Settings

Everything that is not the story lives in a panel that starts closed. The
top navigation bar is hidden on the story screen entirely, because on that one
screen the story is the interface.

Play is one natural-language field. An action and a piece of quoted dialogue are
both just what the reader wrote, and B01/B02 confirmed against a real narrator
that the model reads the quotes without being told which kind of turn it is.
What survives from the old Story mode is a Story direction toggle, which is not
a fourth mode: it changes who is being spoken to, not what kind of action is
taken, and the box is visibly marked while it is on.

Branch, fork, node, merge and head appear nowhere a reader can see them. The
branch panel and the tree overlay are gone from the browser. The mechanism is
untouched — takes, divergence, retained futures and Save Points all still work,
and their endpoints are still tested. This is a decision about what a reader is
asked to understand, not a reduction of what the product can do.

The two defects worth the space:

A player action is stored with AI Dungeon's "> You " prefix. That was right when
the Do mode asked for a bare verb phrase. With one field the spec tells the
reader to write "I enter the tavern", and the result was "> You I enter the
tavern." — in the transcript, in the replayed history, and therefore in the
narration, where a small model imitates it and writes "You I thank her". M8's
own design surfaced it, so M8 fixed it: the prefix is added only when the reader
has not already written a subject. The ">" marker, which is what actually
identifies a player turn in the prompt, is unchanged in every case.

And a stale `.input-bar { display: flex }` in play.css overrode the new
composer, because that sheet is imported after the new one. The direction row
and the input row laid out side by side and the box was unusably narrow. Found
by opening the product in a browser, not by reading the CSS — which is the
argument for having done that first.

Failures now have the taxonomy the spec asked for rather than one toast: model,
generation, state, knowledge, server, each with the thing to do about it. A
failed turn leaves the reader's words in the box and says so. The classification
reads backend strings, so it is a fallback ladder rather than a lookup — an
unrecognised message still classifies, still shows the server's own words and
still offers Retry.

`Settings.model` could be empty with nothing saying so until the first turn
failed with a provider error. The header now reports Ollama in five states, and
an unconfigured or missing model offers the models actually installed on the
endpoint, from the connection test that already knew them. Nothing is chosen
automatically: an endpoint's first model may be an embedding model, which cannot
narrate at all.

Narrator prose is rendered as safe Markdown — headings, emphasis, lists,
blockquotes, code. The safety is structural rather than filtered: every node is
a React element built from parsed text, and there is no dangerouslySetInnerHTML
in the file. A sanitizer is not needed to make markup safe if markup is never
produced from input. Link schemes are checked with the URL parser rather than a
pattern, because the bypasses are all in the parsing. A remote image is a
placeholder naming the blocked address; the knowledge and context panels
deliberately do not use this renderer at all, because they exist to show a
reader exactly what is in their file.

Backend, and only what the browser could not otherwise reach:

  AdventureCreate.opening   a start action could only come from a Scenario, so
                            every campaign made in the new setup flow opened on
                            a blank page. Same node, same code path.
  canon_rules               campaign_canon has been the highest authority in a
                            campaign since M5, read by the prompt builder and
                            the state validator, and had no API at all — a
                            fixture had to write it with SQL.
  a 401 and a 429 message   the last user-facing text describing a hosted
                            deployment. One told the reader to check an API key
                            that has not existed since M2.

No schema change and no migration: proved by building a database with a server
running the M7 commit's own code and opening it with this one.

The project had no frontend tests. It has 132 now, across ten files, running
in about six seconds — the enabled state of every history control, the take
selector, the confirmations, the panels, the five model states, the failure
taxonomy, the focus trap, accessibility, and that the reserved dictation control
never touches the microphone. Writing them found a real defect: the focus trap
filtered candidates with offsetParent, which is null inside the fixed-position
ancestor the dialog has and which jsdom never computes — it would have behaved
differently in the tests from the browser.

They do not replace the real-browser runs, and both kinds of evidence are in the
report. The browser suites drive the production build served by the real backend
with a real local narrator, including a genuine process restart.

A verification pass over all of it then found three more, each by driving the
product rather than reading it:

Stepping between alternate takes did nothing. The pager asked whether a take
lived on another line by comparing `target.branch_id !== action.branch_id`, and
`ActionOut` has never carried `branch_id` — so the comparison was permanently
`number !== undefined`, always true, and every step took the branch-switch path.
For two takes of an ordinary retry, which share a line until one is written
below, that meant switching to the line already being read: the same window came
back and nothing moved. D07 is a required v1 acceptance test. The fix needed no
new field — the variants list already carries every attempt's branch and marks
the live one.

The first regression test for that passed against the broken code, because its
fixture gave the action a `branch_id` the real payload never sends. That is the
exact failure M7's review was about, so the fixture was corrected, the tests were
re-run against the reverted code and failed for the right reason, and the
fixture now carries a docstring saying why the field must never come back.

And the knowledge panel pointed readers at an "embedding model" while the
setting is called "Model for meaning-based search" — a reader sent looking for a
field that does not exist by that name.

Campaign canon was measured rather than assumed. Editing it after play is a
configuration change: every turn already played keeps the canon it was actually
given, in its own context snapshot, and the accepted story, the state document
and the state audit log are byte-identical across an edit. It is not routed
through M5's state audit, because canon is not narrative state and doing so
would create the second representation the spec forbids. What the editor does
now is say so, once a campaign has moments.

`BROWSER-UX-SPEC.md` §38 asked for a "Show Hidden Story State" toggle. There is
no hidden story state — a secret lives in a narrator-only knowledge source and
never enters the state document. The section is rewritten to require what it
actually meant: ordinary surfaces must not carry narrator-only information,
advanced inspection must withhold it by default behind an explicit warned
choice, and no second store may be invented to give a toggle something to
reveal. The protection is stricter than before, not weaker.

Closeout. An independent review returned M8 IMPLEMENTATION: PASS subject to
evidence and documentation cleanup, and this commit carries that cleanup:

The report named two frontend bundles as the artifact behind its acceptance
evidence. The saved run logs settle it. index-Ii-lARp9.js, built at 18:53:02
from this tree, is the one final frozen artifact behind all 157 browser checks;
index-C6E5Uvtu.js is superseded — it predates the D09 fix and its acceptance
suite ended 54/55 on exactly that defect. No tracked file under backend/app or
frontend/src has a modification time after the freeze, so the whole final
campaign describes one build. §P sets the two side by side.

Finding 14 — the app budgets 16,384 prompt tokens while an Ollama that sees no
VRAM enforces 4,096 — is resolved operationally, with no application change.
The OpenAI-compatible endpoint this app speaks accepts num_ctx and ignores it,
and reloads the model at its own default, so a native call cannot prime it
either. A model derived with POST /api/create carries the parameter, is honoured
through the app's own OpenAI-compatible path, and appears in /v1/models — which
is the listing the Settings model picker already reads. Measured end to end.
The procedure is in DEVELOPMENT.md; nothing in the repository depends on any
particular derived model existing. Adding provider code to work around this was
declined deliberately: it would mean either a second native request path,
against ADR 011, or a parameter the endpoint provably ignores.

The §38 rewrite is ratified as a requirement clarification aligned with the
implemented architecture, and the spec gains the clause finding 3 was really
about: withheld material must be absent from the rendered DOM, not merely
collapsed in it.

The report's §U carries the M9 handoff — what a portable campaign has to include,
whether historical context snapshots belong in the bundle, what happens to
inherited story cards, and that a restored campaign may meet a different context
window than the one that wrote it. None of it is implemented here.

Final: backend 950 passed / 14 skipped; frontend 132 passed; lint, production
build and Docker build clean; 157 browser checks across six suites, zero
failures. M8 is implemented, verified, reviewed and accepted (2026-09-06).
M9 has not been started.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017HdaXiFbscatQaLS7dJk6b
2026-09-06 23:31:45 -04:00

688 lines
25 KiB
React

/* The story screen.
*
* `BROWSER-UX-SPEC.md` §98 sets the priority this file is arranged around:
* transcript, input, Undo/Redo/Retry, Save Point, current state, then
* diagnostics. The transcript and the composer are the page; everything else
* lives in a panel that starts closed and can be closed again.
*
* The session state below is inherited and largely unchanged, because it was
* right: takes, previews, the pinned take, the paging window and the
* server-reported history flags all work the way M3-M7 established, and M8 is
* not a rewrite of them. What changed is what the reader is shown and what they
* are allowed to do. The substantive behavioural changes:
*
* - **Failures are classified rather than toasted** (§12). A failed turn
* leaves the typed text in the box, drops any partial prose from the
* transcript, and explains which of §71's five failures happened.
* - **Editing explains itself** (§16, §17). Changing an earlier turn with
* story after it now says what will happen — a different continuation, the
* existing one retained — before it happens, rather than silently forking.
* - **Branch controls are gone from the reader's view** (§27). The panel that
* drew the tree is not reachable from here. The mechanism underneath is
* untouched: takes, divergence and retained futures all still work, and the
* endpoints are still tested.
* - **The RPG stat drawer only opens for a campaign that has stats** (§18),
* which a campaign created in M8's setup flow does not.
*/
import { useCallback, useEffect, useLayoutEffect, useMemo, useRef, useState } from 'react'
import { useNavigate, useParams } from 'react-router-dom'
import { api } from '../../api'
import { AutoTextarea, useToast } from '../../components'
import { ConfirmDialog } from '../../Dialog'
import { classifyError } from '../../errors'
import { ExternalLinkDialog } from '../../ExternalLinkDialog'
import { ModelSetupNotice } from '../../ModelSetupNotice'
import { ModelStatusBadge } from '../../ModelStatusBadge'
import { blocksPlay, useModelStatus } from '../../modelStatus'
import { Composer } from './Composer'
import { FailureNotice } from './FailureNotice'
import { SidePanel, PANELS } from './SidePanel'
import { Transcript, displayText } from './Transcript'
const PLAYER_TYPES = ['do', 'say', 'story']
export default function Play() {
const { id } = useParams()
const navigate = useNavigate()
const toast = useToast()
const { status: modelStatus, refresh: refreshModelStatus } = useModelStatus()
const [adventure, setAdventure] = useState(null)
const [actions, setActions] = useState([])
const [input, setInput] = useState('')
const [direction, setDirection] = useState(false)
const [streaming, setStreaming] = useState(null)
const [reasoningStream, setReasoningStream] = useState(null)
const [busy, setBusy] = useState(false)
const [failure, setFailure] = useState(null) // classified, from errors.js
const [partial, setPartial] = useState(null) // prose that streamed then failed
const [editing, setEditing] = useState(null)
const [editWarning, setEditWarning] = useState(null) // pending edit awaiting confirmation
const [panel, setPanel] = useState(null)
const [stateKey, setStateKey] = useState(0)
const [inspectActionId, setInspectActionId] = useState(null)
const [preview, setPreview] = useState(null)
const [pinned, setPinned] = useState(null)
const [takesKey, setTakesKey] = useState(0)
const [total, setTotal] = useState(0)
const [hasMore, setHasMore] = useState(false)
const [loadingOlder, setLoadingOlder] = useState(false)
const [history, setHistory] = useState({ undo: false, redo: false })
const [externalLink, setExternalLink] = useState(null)
// Set when a retrieved passage in the Context panel is clicked, so the
// Knowledge panel opens on that source instead of on the whole list (§57).
const [focusSourceId, setFocusSourceId] = useState(null)
const storyEndRef = useRef(null)
const abortRef = useRef(null)
const pinnedRef = useRef(true)
const restoreScrollRef = useRef(null)
const loadingOlderRef = useRef(false)
const retryAfterRef = useRef(0)
// What was streaming when a failure arrived, read in the catch. State would
// still hold the pre-failure value there, so this is a ref.
const streamedRef = useRef('')
// The words the reader submitted for the turn in flight, so A05 can give them
// back. A ref because the SSE error path runs inside `handleEvent`, which must
// not re-created on every keystroke.
const keepInputRef = useRef(null)
const firstNarrationId = useMemo(
() => actions.find((a) => a.type === 'start' || a.type === 'ai')?.id ?? null,
[actions],
)
const previewCutoff = useMemo(
() => (preview ? actions.findIndex((a) => a.id === preview.actionId) : -1),
[preview, actions],
)
const waitingForFirstToken = streaming === '' && reasoningStream === null
const lastIsAi = actions.length > 0 && actions[actions.length - 1].type === 'ai'
useEffect(() => {
api.getAdventure(id)
.then((adv) => {
setAdventure(adv)
setActions(adv.actions)
setTotal(adv.action_count ?? adv.actions.length)
setHasMore(adv.actions.length < (adv.action_count ?? adv.actions.length))
setHistory({ undo: !!adv.can_undo, redo: !!adv.can_redo })
})
.catch(() => navigate('/'))
}, [id, navigate])
const noteHistory = useCallback((page) => {
setHistory({ undo: !!page.can_undo, redo: !!page.can_redo })
}, [])
const adoptWindow = useCallback((page) => {
pinnedRef.current = true
setPreview(null)
setPinned(null)
setActions(page.actions)
setTotal(page.total)
setHasMore(page.has_more)
noteHistory(page)
setStateKey((k) => k + 1)
}, [noteHistory])
const loadOlder = useCallback(async () => {
if (loadingOlderRef.current || !hasMore) return
if (Date.now() < retryAfterRef.current) return
const oldest = actions[0]
if (!oldest) return
loadingOlderRef.current = true
setLoadingOlder(true)
try {
const page = await api.getActions(id, { beforeId: oldest.id })
if (page.actions.length) {
pinnedRef.current = false
restoreScrollRef.current = {
height: document.documentElement.scrollHeight,
top: window.scrollY,
}
setActions((prev) => {
const known = new Set(prev.map((a) => a.id))
return [...page.actions.filter((a) => !known.has(a.id)), ...prev]
})
}
setTotal(page.total)
setHasMore(page.has_more)
noteHistory(page)
} catch {
// Leave hasMore alone so scrolling can try again, but hold off: parked
// near the top, momentum scrolling fires this many times a second.
retryAfterRef.current = Date.now() + 3000
} finally {
loadingOlderRef.current = false
setLoadingOlder(false)
}
}, [actions, hasMore, id, noteHistory])
useLayoutEffect(() => {
const mark = restoreScrollRef.current
if (!mark) return
restoreScrollRef.current = null
const grown = document.documentElement.scrollHeight - mark.height
if (grown > 0) window.scrollTo({ top: mark.top + grown })
}, [actions])
useEffect(() => {
const onScroll = () => {
pinnedRef.current =
window.innerHeight + window.scrollY >= document.documentElement.scrollHeight - 120
if (window.scrollY < 400) loadOlder()
}
window.addEventListener('scroll', onScroll, { passive: true })
return () => window.removeEventListener('scroll', onScroll)
}, [loadOlder])
useEffect(() => {
if (pinnedRef.current) {
window.scrollTo({ top: document.documentElement.scrollHeight })
}
}, [actions, streaming, reasoningStream])
/**
* One failed turn, however the failure arrived.
*
* There are two paths and they must behave identically. A transport or HTTP
* failure throws, and `runTurn`'s catch sees it. A *generation* failure does
* not throw at all: the server reports it as an SSE `error` event and the
* request then completes normally, so the catch never runs.
*
* That difference was a real A05 defect. The failure notice said "what you
* typed is still in the box" while the box was empty, and the header went on
* claiming the model was connected — because restoring the input and
* re-testing the model both lived in the catch alone. Found by driving a real
* provider failure in a browser.
*/
const failTurn = useCallback((detail) => {
const classified = classifyError(detail)
setStreaming(null)
setReasoningStream(null)
setFailure(classified)
setPartial(streamedRef.current || null)
// A05: the player's words come back so the turn can simply be sent again.
if (keepInputRef.current !== null) {
setInput(keepInputRef.current)
keepInputRef.current = null
}
// A model problem may well have changed since the status was last checked —
// re-test, so the header stops claiming Connected while the notice says the
// opposite.
if (classified.kind === 'model') refreshModelStatus()
}, [refreshModelStatus])
const handleEvent = useCallback((event) => {
if (event.type === 'player') {
setActions((prev) => [...prev, event.action])
setTotal((n) => n + 1)
} else if (event.type === 'chunk') {
streamedRef.current += event.text
setStreaming((prev) => (prev ?? '') + event.text)
} else if (event.type === 'reasoning') {
setReasoningStream((prev) => (prev ?? '') + event.text)
} else if (event.type === 'done') {
setStreaming(null)
setReasoningStream(null)
setActions((prev) => [...prev, event.action])
setTotal((n) => n + 1)
setHistory({ undo: true, redo: false })
} else if (event.type === 'error') {
failTurn(event.detail)
}
}, [failTurn])
async function runTurn(run, { keepInput = null } = {}) {
const controller = new AbortController()
abortRef.current = controller
keepInputRef.current = keepInput
setBusy(true)
setFailure(null)
setPartial(null)
streamedRef.current = ''
setStreaming('')
pinnedRef.current = true
try {
await run(controller.signal)
} catch (err) {
if (err.name === 'AbortError') {
toast('Generation stopped.')
setStreaming(null)
setReasoningStream(null)
} else {
// Whatever streamed is not story. It was never added to `actions` —
// only a `done` event does that — but the streaming pane has to be
// cleared or it reads as accepted text. `failTurn` does that too.
failTurn(err.message)
}
} finally {
abortRef.current = null
keepInputRef.current = null
setStreaming(null)
setReasoningStream(null)
setBusy(false)
}
}
function stopGeneration() {
abortRef.current?.abort()
}
function send(type) {
const text = input.trim()
const after_id = preview?.takeId
const run = (payload, keep) => runTurn(async (signal) => {
try {
await api.sendAction(id, payload, handleEvent, signal)
} finally {
if (after_id) await resync()
}
}, { keepInput: keep })
if (preview) setPinned({ ...preview, written: true })
setPreview(null)
if (type === 'continue') {
run({ type: 'continue', text: '', after_id }, null)
return
}
const payload = { type: text ? type : 'continue', text, after_id }
setInput('')
run(payload, text)
}
function retry() {
setPreview(null)
setActions((prev) =>
prev.length && prev[prev.length - 1].type === 'ai' ? prev.slice(0, -1) : prev)
runTurn(async (signal) => {
try {
await api.retry(id, handleEvent, signal)
} catch (err) {
// The optimistically removed action may still exist server-side, so
// resync rather than guess.
api.getAdventure(id).then((adv) => {
setActions(adv.actions)
setTotal(adv.action_count ?? adv.actions.length)
setHasMore(adv.actions.length < (adv.action_count ?? adv.actions.length))
}).catch(() => {})
throw err
}
})
}
async function moveHead(call) {
setFailure(null)
setPreview(null)
setPinned(null)
try {
const page = await call(id)
setActions(page.actions)
setTotal(page.total)
setHasMore(page.has_more)
noteHistory(page)
setStateKey((k) => k + 1)
} catch (err) {
setFailure(classifyError(err.message))
}
}
const undo = () => moveHead(api.undo)
const redo = () => moveHead(api.redo)
// A take-edit belongs to the preview that opened it.
useEffect(() => {
if (editing?.take && preview?.takeId !== editing.id) setEditing(null)
}, [editing, preview])
/**
* True when accepting this edit will start a different continuation.
*
* That is the case §16/§17 require an explanation for: story exists after the
* turn being changed, so the existing continuation is left behind (retained,
* not deleted). Editing the newest turn changes nothing downstream and needs
* no warning.
*/
const hasStoryAfter = useCallback((actionId) => {
const at = actions.findIndex((a) => a.id === actionId)
return at !== -1 && at < actions.length - 1
}, [actions])
/**
* Splits a stored player action into its marker and its words.
*
* The editor shows the words. The marker goes back on when the edit is saved,
* because it is what identifies a player turn in the replayed prompt — an
* edit that dropped it would quietly turn the reader's action into narration
* as far as the next turn's context is concerned.
*/
function splitMarker(target) {
const raw = target.text ?? ''
const shown = displayText({ type: target.kind }, raw)
return { marker: raw.slice(0, raw.length - shown.length), text: shown }
}
function beginEdit(target) {
// A take being read is a leaf by construction, so editing it never strands
// a continuation — no warning for that case.
const prepared = { ...target, ...splitMarker(target) }
if (target.take || !hasStoryAfter(target.id)) {
setEditing(prepared)
return
}
// Story exists after this turn, so the reader is about to change something
// the rest of the story was written from. Which operation that *is* differs
// by whose turn it was, and getting it wrong is D09's whole subject.
//
// A narrator turn goes through `PATCH`, because `_edit_narration` already
// forks it per `STORY-BRANCH-SEMANTICS.md` §§14-15: the corrected prose
// becomes a new continuation, downstream state is re-evaluated, and the old
// narration keeps its own future.
//
// A player turn must **not**. `PATCH` on a player row rewrites it in place
// and re-evaluates nothing — a correction, not a new continuation — and
// refuses outright when displaced history hangs off it. Either way D09 is
// unmet: no new continuation is created. Replaying the turn with the new
// words is what creates one, and that is `addTake` (SP9).
//
// The confirmation says "this will continue the story differently… what you
// have already read is kept". Routing a player edit through `PATCH` made
// that sentence false. Found by the browser run: the dialog appeared, the
// editor held the new text, and the edited action never reached the story.
const fork = target.kind !== 'ai'
setEditWarning({ ...prepared, mode: fork ? 'refork' : 'edit' })
}
function beginRetake(action, previewing) {
if (action.type === 'ai') {
if (hasStoryAfter(action.id)) {
setEditWarning({ id: action.id, text: '', kind: action.type, mode: 'retake' })
return
}
addTake(action.id, '')
} else {
const seed = previewing ? previewing.text : action.text
const prepared = splitMarker({ text: seed, kind: action.type })
if (hasStoryAfter(action.id)) {
setEditWarning({
id: action.id, ...prepared, kind: action.type, mode: 'refork',
})
return
}
setEditing({ id: action.id, ...prepared, fork: true, kind: action.type })
}
}
function acceptWarning() {
const pending = editWarning
setEditWarning(null)
if (!pending) return
if (pending.mode === 'retake') addTake(pending.id, '')
else if (pending.mode === 'refork') setEditing({ ...pending, fork: true })
else setEditing(pending)
}
async function saveEdit() {
const { id: actionId, fork, take, marker = '' } = editing
// The marker the action was stored with, restored around the edited words.
const text = marker + editing.text
setEditing(null)
if (fork) {
addTake(actionId, text)
return
}
try {
const updated = await api.updateAction(id, actionId, text)
// D10. Correcting narrator prose is not a row rewrite: the server answers
// with a different node, the old narration keeps its future on the line
// it was written on, and the state after it is re-evaluated. Only the
// server can say what the transcript is now.
if (updated.id !== actionId) {
await resync()
toast('Edited. The story state after it was recalculated.')
return
}
if (take) {
setPreview((p) => (p && p.takeId === actionId ? { ...p, text: updated.text } : p))
setTakesKey((k) => k + 1)
} else {
setActions((prev) => prev.map((a) => (a.id === actionId ? updated : a)))
}
setStateKey((k) => k + 1)
} catch (err) {
setFailure(classifyError(err.message))
}
}
function addTake(actionId, text) {
setPreview(null)
runTurn(async (signal) => {
try {
await api.addTake(id, actionId, text, handleEvent, signal)
} finally {
await resync()
}
})
}
async function resync() {
try {
const adv = await api.getAdventure(id)
setPinned(null)
setActions(adv.actions)
setTotal(adv.action_count ?? adv.actions.length)
setHasMore(adv.actions.length < (adv.action_count ?? adv.actions.length))
setHistory({ undo: !!adv.can_undo, redo: !!adv.can_redo })
setStateKey((k) => k + 1)
} catch { /* stale beats wrong */ }
}
function inspect(actionId) {
setInspectActionId(actionId)
setPanel('context')
}
const copy = useCallback((text) => {
navigator.clipboard?.writeText(text)
.then(() => toast('Copied.'))
.catch(() => toast('Could not copy.', 'error'))
}, [toast])
if (!adventure) return null
const editor = (
<div className="turn-edit">
<label className="sr-only" htmlFor="turn-editor">Edit this text</label>
<AutoTextarea
id="turn-editor"
autoFocus
value={editing?.text ?? ''}
onChange={(e) => setEditing({ ...editing, text: e.target.value })}
/>
<div className="edit-buttons">
<button type="button" className="primary" onClick={saveEdit}>
{editing?.fork ? 'Save and continue' : 'Save'}
</button>
<button type="button" onClick={() => setEditing(null)}>Cancel</button>
</div>
{editing && !editing.fork && editing.kind === 'ai' && (
<p className="edit-hint">
What the story believes after this point will be worked out again from
your version.
</p>
)}
</div>
)
return (
<div className={`play ${panel ? 'with-panel' : ''}`}>
<header className="play-header">
<div className="play-title">
<button
type="button"
className="linklike back"
onClick={() => navigate('/')}
aria-label="Back to your campaigns"
>
← Campaigns
</button>
<h1>{adventure.title}</h1>
</div>
<div className="play-header-right">
<ModelStatusBadge compact />
<nav className="panel-tabs" aria-label="Campaign panels">
{PANELS.map((p) => (
<button
key={p.id}
type="button"
className={panel === p.id ? 'active' : ''}
aria-pressed={panel === p.id}
onClick={() => {
if (p.id === 'context') setInspectActionId(null)
setPanel(panel === p.id ? null : p.id)
}}
>
{p.label}
</button>
))}
</nav>
</div>
</header>
<div className="play-main">
<div className="play-story">
<Transcript
actions={actions}
advId={id}
busy={busy}
streaming={streaming}
reasoningStream={reasoningStream}
waitingForFirstToken={waitingForFirstToken}
firstNarrationId={firstNarrationId}
previewCutoff={previewCutoff}
preview={preview}
pinned={pinned}
editing={editing}
takesKey={takesKey}
total={total}
hasMore={hasMore}
loadingOlder={loadingOlder}
onLoadOlder={loadOlder}
onPreview={setPreview}
onSwitchedBranch={adoptWindow}
onTakeError={(message) => setFailure(classifyError(message))}
onEdit={beginEdit}
onRetake={beginRetake}
onInspect={inspect}
onCopy={copy}
onLink={setExternalLink}
editor={editor}
endRef={storyEndRef}
/>
{/* §8: a reader whose model is down finds Send disabled. Telling
them why, and how to fix it, belongs here rather than only on
Settings — a disabled button with a red badge is a puzzle. The
browser run found this by loading the story screen with the
endpoint dead and finding no guidance on it at all. */}
{blocksPlay(modelStatus) && (
<div className="story-blocked">
<ModelSetupNotice />
</div>
)}
<FailureNotice
failure={failure}
partial={partial}
onRetry={() => { setFailure(null); setPartial(null); retry() }}
onDismiss={() => { setFailure(null); setPartial(null) }}
/>
<Composer
input={input}
setInput={setInput}
direction={direction}
setDirection={setDirection}
busy={busy}
canUndo={history.undo}
canRedo={history.redo}
canRetry={lastIsAi}
onSend={send}
onContinue={() => send('continue')}
onRetry={retry}
onUndo={undo}
onRedo={redo}
onSavePoint={() => setPanel('savepoints')}
onStop={stopGeneration}
/>
</div>
{panel && (
<SidePanel
panel={panel}
advId={id}
adventure={adventure}
setAdventure={setAdventure}
refreshKey={`${actions.length}:${stateKey}`}
inspectActionId={inspectActionId}
onClearInspect={() => setInspectActionId(null)}
onClose={() => setPanel(null)}
onRestored={adoptWindow}
onCorrected={() => setStateKey((k) => k + 1)}
onError={(message) => setFailure(classifyError(message))}
focusSourceId={focusSourceId}
onOpenSource={(sourceId) => {
setFocusSourceId(sourceId)
setPanel('knowledge')
}}
moments={total}
/>
)}
</div>
{editWarning && (
<ConfirmDialog
title={
editWarning.kind === 'ai'
? 'Change this part of the story?'
: 'Change what you did here?'
}
confirmLabel="Continue differently"
onCancel={() => setEditWarning(null)}
onConfirm={acceptWarning}
>
{/* §16's wording, in the product's own words. The thing a reader
needs to know is not that a branch is created — it is that the
story they have already read does not disappear. */}
<p>
Changing this will continue the story differently from here.
</p>
<p>
Everything you have already read after this point is kept. It stops
being the story you are in, and you can come back to it with Undo or
from a Save Point.
</p>
{editWarning.kind === 'ai' && (
<p>
What the story believes — where people are, what they know, what
they carry — will be worked out again from your version.
</p>
)}
</ConfirmDialog>
)}
{externalLink && (
<ExternalLinkDialog
href={externalLink}
onClose={() => setExternalLink(null)}
/>
)}
</div>
)
}
export { PLAYER_TYPES }