Add ability to import PDF

[albertski]

Users can now import a PDF into a book directly from the table of contents toolbar. Each page of the PDF is imported as a separate textcpage, using the first line as the title.

Implementation:

• PdfImporter model wrapping the pdf-reader gem
• Books::ImportsController (new/create, editor-only)
• Import button in the create buttons toolbar
• Unit tests for PdfImporter and controller integration tests

Totally understand if this is something you don’t want to support.

[Copilot]

Pull request overview

Adds a PDF import workflow for books so editors can upload a PDF from the TOC toolbar and have each PDF page imported as a separate text Page/Leaf.

Changes:

• Introduces PdfImporter to parse PDFs (via pdf-reader) and create leaves/pages.
• Adds Books::ImportsController + route + books/imports/new form to upload a PDF (editor-only).
• Adds an “Import PDF” button to the existing create-buttons toolbar; adds unit/integration tests and PDF fixtures.

Reviewed changes

Copilot reviewed 8 out of 11 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
app/models/pdf_importer.rb	New importer that reads PDF pages and creates Leaf/Page records.
app/controllers/books/imports_controller.rb	New editor-only endpoints to render the form and run the import.
app/views/books/imports/new.html.erb	Upload form UI for selecting and importing a PDF.
app/views/books/_create_buttons.html.erb	Adds a new toolbar button to access the import flow.
config/routes.rb	Adds nested book import routes (new/create).
test/models/pdf_importer_test.rb	Unit tests for importer behavior (titles, body content, skipping blank pages, malformed PDFs).
test/controllers/books/imports_controller_test.rb	Integration tests for import endpoints + authorization.
test/fixtures/files/sample.pdf	PDF fixture with extractable text across 2 pages.
test/fixtures/files/blank.pdf	PDF fixture intended to have no extractable text.
Gemfile / Gemfile.lock	Adds pdf-reader and its transitive dependencies.

[!TIP]
If you aren't ready for review, convert to a draft PR.
Click "Convert to draft" or run gh pr ready --undo.
Click "Ready for review" or run gh pr ready to reengage.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The new "Import PDF" toolbar button is missing the disable-when-arranging / disable-when-deleting classes used by the other create buttons in this toolbar, so it will remain clickable during rearrange/delete modes while other actions are intentionally disabled. Consider reusing the same button classes/disable hooks for consistent behavior and styling.

[Copilot]

PdfImporter#import returns Leaf records (as also asserted in PdfImporterTest), but this action stores the result in a variable named pages. Renaming this to something like imported_leaves (and adjusting the notice string accordingly) would make the code clearer and reduce confusion between Page (leafable) vs Leaf (wrapper).

Suggested change

	pages = PdfImporter.new(@book, params[:pdf].tempfile).import
	if pages.any?
	redirect_to book_slug_url(@book), notice: "Successfully imported #{pages.size} #{"page".pluralize(pages.size)} from PDF."
	imported_leaves = PdfImporter.new(@book, params[:pdf].tempfile).import
	if imported_leaves.any?
	redirect_to book_slug_url(@book), notice: "Successfully imported #{imported_leaves.size} #{"leaf".pluralize(imported_leaves.size)} from PDF."

[Copilot]

The user-facing alert message includes the raw exception message from pdf-reader (Could not read PDF: #{e.message}). These messages can be noisy/inconsistent and may reveal implementation details; consider logging the exception details and showing a stable, user-friendly message instead (optionally with guidance like “Please verify the file is a valid, unencrypted PDF”).

Suggested change

	flash.now[:alert] = "Could not read PDF: #{e.message}"
	Rails.logger.warn("PDF import failed for book #{ @book&.id || 'unknown' }: #{e.class}: #{e.message}")
	flash.now[:alert] = "Could not read PDF. Please verify the file is a valid, unencrypted PDF."

[Copilot]

extract_title uses hard-coded title length limits (between?(3, 100)). To make the intent clearer and future changes safer, consider extracting these numbers into named constants (e.g., MIN_TITLE_LENGTH, MAX_TITLE_LENGTH) or documenting why those bounds were chosen.

[Copilot]

import creates multiple records (one per PDF page) but does not run inside a transaction. If an exception occurs midway (e.g., validation/database error), the book can end up partially imported. Wrapping the per-page creation loop in a single transaction (or otherwise ensuring all-or-nothing behavior) would keep imports consistent.