Shooting Schedule & Day Out of Days#2
Open
newjordan wants to merge 3 commits into
Open
Conversation
ScriptBreak's arc is breakdown → schedule → shoot, and the app already points at scheduling (the shot-list CSV is "for StripBoard, or your 1st AD"; the scene CSV is "the skeleton of a schedule") but never delivered it. This adds the missing step, as an additive export in the single index.html — no build step, no dependencies, no network. - Suggested stripboard: scenes grouped into synthetic shoot days by master location -> lighting regime (DAY / dusk-dawn "MAGIC" / NIGHT) -> INT vs EXT -> scene number, then split under a page budget (default 5 pg/day, +/-1 pg tolerance; one location per day). Print/PDF view + CSV. - Day Out of Days: the standard cast x shoot-day grid with SW / W / WF / SWF / H status codes and per-cast Work / Hold / Total(span) counts. Print/PDF + CSV. - New "Shooting Schedule & Day Out of Days" card in Export (with a pages/day control), both CSVs added to the "Everything" zip bundle, and a Help entry. Honest by design: cast presence is inferred from dialogue cues only, so silent/background cast are not detected and the DOOD under-reports who is needed; page counts may be estimated; and it does not model cast/location availability, company moves, or turnaround. Every output carries a prominent "draft — verify with your 1st AD" caveat. Conventions follow standard AD practice (Movie Magic-style grouping; standard DOOD SWF codes). Reuses exportScenes()/scopeLabel() (honors the scope bar), the openBreakdown- Sheets() print pattern, csvCell()/download(), and makeZip(). The scheduling helpers are duplicated verbatim in the MCP server (next commit) to preserve app<->MCP output parity. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01A2NtcuBy6SQmKaBQtgvkog
…f_days Add two read-only tools to the headless MCP server so agents (Claude Code, Codex, Hermes, …) get the same shooting schedule and Day Out of Days as the desktop app: - get_schedule — the suggested stripboard: ordered shoot days with location, I/E, day/night, scene numbers, page eighths, and cast IDs, plus a byte- identical `csv`. - get_day_out_of_days — the cast DOOD grid (SW/W/WF/SWF/H codes, Work/Hold/ Total counts) with a `legend`, plus a byte-identical `csv`. Both accept `pagesPerDay` and the existing scope selectors, and return a `caveat` field spelling out that cast presence is dialogue-inferred (so the DOOD under-reports) and the schedule is a draft. The scheduling helpers are copied verbatim from index.html; rendering both sides for the same project and diffing confirms the schedule CSV is byte-identical between app and MCP. Docs: bump the tool count to eleven and document the two tools + caveat in the root README and mcp/README. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01A2NtcuBy6SQmKaBQtgvkog
Follow-ups from an adversarial audit of the schedule/DOOD feature: - CSV formula injection: csvCell now prefixes any cell beginning with = + - @ (or tab/CR) with an apostrophe, so a location or character name like =HYPERLINK(...) / =1+1 / @sum(...) can't execute when the CSV is opened in Excel/Sheets/LibreOffice. Applied identically in the app and the MCP (which also makes the two csvCell copies textually identical), and it protects the pre-existing shotlist/scenes CSVs too. - DOOD/schedule CSVs now carry a leading "DRAFT — …verify with your 1st AD" note row, so the caveat travels with the file once it's detached from the app (the payroll-relevant surface). Identical in app and MCP → CSV stays byte-for-byte across both. - MCP get_schedule / get_day_out_of_days coerce pagesPerDay with Number() and fall back to 5 for non-positive/NaN input (a string previously made the day budget NaN and silently disabled day-splitting). - Docs: scope the MCP "byte-for-byte" claim to all/scene/page scopes and state plainly that scope:"filter" reproduces the structured toggles only (not the app's live free-text search box) and that the generation-date header line is locale/day-dependent — the previous "any filter, byte-for-byte" was an overclaim. Verified: app and MCP schedule/DOOD CSVs remain byte-identical (incl. the note row and escaped injection payloads); pagesPerDay:"abc" -> 5; all app <script> blocks and the MCP pass node --check; 11 tools intact. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01A2NtcuBy6SQmKaBQtgvkog
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
1 participant
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.
Shooting Schedule & Day Out of Days
ScriptBreak's arc is breakdown → schedule → shoot, and the app already points at scheduling — the shot-list CSV is sold as "for StripBoard, or your 1st AD," the scene CSV as "the skeleton of a schedule" — but it stops at the breakdown and never delivers the schedule itself. This PR adds that missing step: a suggested shooting schedule (draft stripboard) and the standard cast Day Out of Days, the paperwork a 1st AD / line producer needs to take a breakdown to set.
It stays true to every project constraint: additive JS in the single
src/index.html— no build step, no dependencies, no network — and it's mirrored read-only in the MCP server so agents get it too.What's in it
Desktop app (
src/index.html) — a new Export → "Shooting Schedule & Day Out of Days" card with a pages/day control and four outputs:MCP server (
mcp/scriptbreak-mcp.mjs) — two new read-only tools,get_scheduleandget_day_out_of_days, acceptingpagesPerDay+ the existing scope selectors and returning structured JSON plus a byte-identicalcsvand acaveatfield. (9 → 11 tools.)Honest by design
Cast presence in ScriptBreak comes from dialogue cues only, so a performer on set with no line in a scene isn't in
sc.characters. A DOOD is entirely a cast-presence document, so this would under-report who's needed. Rather than hide that, every output — the print views, the zip README, the export-card copy, and the MCPcaveatfield — says so plainly: this is a draft starting point for your 1st AD, not a locked schedule; cast presence is dialogue-inferred, page counts may be estimated, and it doesn't model cast/location availability, company moves, or turnaround. Conventions follow standard AD practice (Movie-Magic-style grouping; standard DOOD SWF codes; all in-span non-work days shown as paid Hold, the conservative default).Verification
<script>blocks inindex.htmlpassnode --check; the MCP passesnode --checkand a full stdio round-trip (11 tools; both new tools exercised).=/+/-/@) and a leading "DRAFT — verify with your 1st AD" note row so the caveat travels with a detached file — and tightens the MCP's parity wording.🤖 Generated with Claude Code
https://claude.ai/code/session_01A2NtcuBy6SQmKaBQtgvkog