board-setup — printboard

Print the Enabler board's papers — the right ones, at the right size, the right number of copies, and the right version — in one command.

The papers live as slides in one Google Slides deck (the source of truth that colleagues edit). printboard exports that deck to a PDF on demand, finds each paper's slide by its title, and sends it to the printer at the size and count declared in manifest.json.

See CONTEXT.md for the vocabulary and the decisions behind this.

Usage

printboard                            # ACTIVATION: every variant of every paper, at its count
printboard generic                    # only the generic (blank) variants — ×5 each
printboard generic tech-working-conditions   # one paper, one variant — 1 copy (quick reprint)
printboard with-examples              # only the worked-example references — ×1 each
printboard generic takt --count 5     # a named paper, but this many copies
printboard --printer _4e_etage
printboard generic --to-pdf ~/Downloads/board.pdf   # export a PDF instead of printing

printboard --list-titles              # debug: show page → title for the live deck
printboard --dry-run                  # show the lp commands without printing
printboard printer                    # pick the default printer from a menu (A3/A4 shown)
printboard setup                      # one-time: authorise rclone (read-only Drive)
printboard doctor                     # check deps, auth, and that the deck exports

• No version → activation: prints every variant of every paper at its version count (with-examples ×1, template ×1, generic ×5). Add a version to filter (printboard generic) and/or a paper id to target one (printboard generic product-radar). Order doesn't matter — each token is recognised as a version or a paper.
• Naming a paper prints 1 copy. version_counts and per-variant counts are board-activation quantities, so a single-paper request (e.g. printboard dantotsu-2) is treated as a quick reprint and defaults to 1 — not the activation count. Use --count N when you do want more.
• Versions are with-examples, template, generic — read from the slide title.
• The printer (must support A3 + A4) is resolved in this order: --printer <name> (per run) → "printer" in ~/.config/printboard/config.json (persistent, printboard only) → print.printer in the manifest → the system default lp uses (set it with lpoptions -d <printer>). printboard doctor shows the one in effect.
• printboard printer is the easy way to set the persistent one: it lists every printer with its friendly name and whether it does A3+A4, and saves your pick to the config above. printboard printer <name> sets one directly; --list just shows them; --unset reverts to the system default.
• --to-pdf <path> writes the selected papers to one PDF instead of sending them to a printer — one page per variant, each sized to its A3/A4. No printer needed, and --count doesn't apply (a PDF is a master; set copies when you print it). Point it at a directory to auto-name the file.

The deck ↔ tool contract

When you rework the deck, follow this so the tool can find each paper:

1. Each printable paper is one slide whose title is its canonical name (e.g. Tech Working Conditions), as real text (not inside an image).
2. That title must be unique in the deck and appear as its own line — the tool matches a standalone title line, so spec/list slides are ignored.
3. A paper with several versions gets one slide per version, titled <Paper> — <version> — e.g. Tech Working Conditions — with examples and Tech Working Conditions — generic. The manifest maps each to its version. A single-variant paper can keep a bare title (e.g. Kanban).
4. Anything that isn't a paper (specs, materials, the "Printed paper needed" table) needs no change — it simply won't match.

printboard refuses to print when a title is missing, ambiguous (on several slides), or only a fuzzy substring match — and tells you what to fix. Use printboard generic --list-titles --from-pdf <export> to check the deck, or --force to override (uses the first matching page).

manifest.json

The source of truth for what Slides can't store — size, count, version-variant titles, and pillar per paper:

{
  "deck":  { "rclone_remote": "gdrive" },        // the private deck id lives in local config, not here
  "print": { "printer": null, "orientation": "landscape", "safe_margin_mm": 0,
             "version_counts": { "with-examples": 1, "template": 1, "generic": 5 },
             "lp_options": ["sides=one-sided"] },
  "papers": [
    { "id": "tech-working-conditions", "pillar": "Tech-enabled Network of Teams", "size": "A3",
      "variants": { "with-examples": "Tech Working Conditions — with examples",
                    "generic":       "Tech Working Conditions" } }
  ]
}

Counts come from version_counts; override one variant with an object: "generic": { "title": "…", "count": 3 }. A variant whose title is the bare paper name matches the unlabelled (blank) slide — the tool gives each contested page to the most specific title, so … - template wins its slide and the bare one falls to generic.

Each page is scaled to its exact A3/A4 size (vector, via ghostscript) before printing, so it fills the sheet on any printer regardless of the driver's fit-to-page behaviour. safe_margin_mm insets the content (0 = full bleed; keep content off the edges in the slide design, or set e.g. 6 to avoid the printer's unprintable border).

The deck id is not stored here — it's private, kept per-machine in ~/.config/printboard/config.json (written by printboard setup), so it never ships in the formula. Manifest lookup order: --manifest → $PRINTBOARD_MANIFEST → ~/.config/printboard/manifest.json → the copy bundled next to the script.

Setup

Requires poppler (pdftotext), rclone, and python3.

brew install poppler rclone
printboard setup --deck "<Google Slides URL or file id>"   # rclone OAuth + save the deck id locally
printboard doctor                                          # verify deps, auth, and that the deck exports

setup authorises rclone (browser, read-only — one time per user) and saves the deck id to ~/.config/printboard/config.json. The deck is org-restricted, so each user authorises as themselves; you do not need to choose a Shared Drive (the deck is fetched by id, so the remote's root is irrelevant). If your Workspace blocks third-party OAuth apps, create your own OAuth client in Google Cloud Console and supply its client_id/secret (re-run rclone config).

Then a no-print check of the live deck:

printboard generic --dry-run

Install (for non-developers)

One line, pasted into Terminal (⌘-Space → "Terminal"). It installs Homebrew if you don't have it, then printboard and everything it needs:

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/jcalixte/board-setup/main/install.sh)"

Then the two one-time steps it prints — a browser sign-in to the deck:

printboard setup --deck "<paste the Google Slides deck URL>"
printboard doctor        # checks deps, auth, and that the deck exports

Install (Homebrew)

brew install jcalixte/tap/printboard

The formula lives in printboard.rb and is mirrored to jcalixte/homebrew-tap → Formula/printboard.rb. Cutting a release is automated: commit and push the change, then run ./release.sh <version> (e.g. ./release.sh 1.4.2) — it tags the repo, hashes the tarball, and bumps url/sha256 in both formulas. See the script header for details.