feat(dashboard): wire reading-pace forecasts into a Time to finish section (EXP-04)

[ChelseaKR]

What & why

app.forecast shipped as a pure, tested module (EXP-04) but nothing rendered it — the feature was invisible to users. This wires it into the dashboard: build_view computes a per-book BookForecast for every currently-reading title, and the renderer shows an accessible "Time to finish" section.

• Remaining pages come from the book's reading stat (total_pages − pages_read); a book with no stat yields the module's honest Forecast.unknown() rather than a guess.
• The section is a <table> mirroring the existing sections: a range, never a single number, with the window basis disclosed ("from your last N reading days"), and unestimable books shown as such ("not enough recent reading to estimate"). No color-only meaning.
• BookForecast lives in app.view; app.render imports it under TYPE_CHECKING (annotations are already lazy via from __future__ import annotations), so no import cycle is introduced.

Definition of Done

• Tests added: an estimable book renders an hours range with its basis and passes the a11y structural check (check_html); a book with no page stat renders the honest "not enough recent reading to estimate" text; the demo render includes the "Time to finish" heading.
• Docs updated: CHANGELOG.md (Unreleased → Added) and the EXP-04 note in docs/ideation/03-expansions.md (dashboard wiring now done).
• No secrets, tokens, or real library data in the diff — tests use constructed fixtures.
• Four hard guardrails intact: this is read-only view/render plumbing over already-ingested data; no reading data leaves the instance; no catalog access touched; no author identity inferred.
• Commit signed off and Conventional Commits.

Verification note (please confirm on CI): this repo targets Python 3.14, which this build sandbox can't provide (interpreter downloads are blocked here), so I could not run the full make verify against 3.14 locally. What I did verify: the full suite (389 tests, including the two new forecast tests) passes when run on Python 3.12 with a temporary 3.12-compat shim that I did not commit — the shipped diff contains only the EXP-04 changes; and ruff check + ruff format --check are clean under the repo's target-version = "py314" config on the changed files. The 3.14 mypy/pytest gate runs in CI; I'll watch it and fix any fallout.

If this touches a workflow or a dependency

N/A — no workflow, uses:, or dependency change; uv.lock untouched.

---

Generated by Claude Code