Files
mopac-pdf/internal/pdfinfo/pdfinfo.go
T
mrcharles 10e930e46d mopac-pdf v0: markdown to beautiful PDFs via digest-pinned typst
Go CLI that turns the markdown currency of this stack (Redmine notes,
Discourse posts, briefing output) into typeset PDFs: two embedded typst
templates (report with title page/TOC/headers, dense brief), front-matter
(title/subtitle/author/date/classification/template), GFM tables, and
bar charts rendered in pure Go (go-chart) from fenced chart data blocks.
Engine is a prebuilt typst 0.15.1 container pinned by digest and run
--network none; PDF bytes to stdout or -o. Exit codes 0/1/2.

All dev in docker (dev.sh/Makefile); unit tests + golden typst fixtures
plus a host-side smoke against the real engine container.

Generated with Crush

Assisted-by: Crush:glm-5.2
2026-08-29 05:44:24 -05:00

39 lines
1.2 KiB
Go

// Package pdfinfo extracts basic facts from PDF bytes with a tiny parser —
// no third-party PDF library, just enough to count pages in golden tests
// and smoke checks.
package pdfinfo
import (
"bytes"
"fmt"
"regexp"
"strconv"
)
// Pages returns the page count of a PDF by scanning for the /Type /Pages
// objects' /Count entries (the catalog root carries the total; we take the
// maximum to be safe with nested page trees).
func Pages(pdf []byte) (int, error) {
if len(pdf) == 0 || !bytes.HasPrefix(pdf, []byte("%PDF-")) {
return 0, fmt.Errorf("pdfinfo: not a PDF (missing %%%%PDF- header)")
}
re := regexp.MustCompile(`/Type\s*/Pages[^>]*?/Count\s+(\d+)`)
best := 0
for _, m := range re.FindAllStringSubmatch(string(pdf), -1) {
if n, err := strconv.Atoi(m[1]); err == nil && n > best {
best = n
}
}
if best == 0 {
// Compressed object streams can hide the count; require at least
// the header then fall back to /Type /Page occurrences (\b keeps
// /Pages from matching).
pageRe := regexp.MustCompile(`/Type\s*/Page\b`)
best = len(pageRe.FindAll(pdf, -1))
if best == 0 {
return 0, fmt.Errorf("pdfinfo: no page count found")
}
}
return best, nil
}