imprimatur 0.7.0 → 0.8.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9812ca055242755ab6117d2c5eba9bb1991e02a21572682f9f36c2a0481e2139
-  data.tar.gz: 8e1d32d440ec79ea732c62dd0a830c0b13f76655df7f9b85d60d539e0d429936
+  metadata.gz: 7c343d074a7ed5e2ba382910345401adaaa92269eb915c901db443d97ed6f68a
+  data.tar.gz: 00cfe2c403d09f50708469c6228813f8ab9361775b848ac23457389008e643ab
 SHA512:
-  metadata.gz: 3572b1a17819ae0fa85346788343f19ef9410d19f48b7257edb95dc00544ae1234991a5d0b50284ed3d10e21f34ca991f965aa7956fecb6a03244805870cea13
-  data.tar.gz: 9399ac644f99a699a12a2a8bdcefe73977bddcec75571e8053d984d829ddc33817fd40bb6c9c79fd15ce89f56fbe472035c276e3177430b9a35ab20023b66f16
+  metadata.gz: e003c94d4f867c4d957a7921a8ef50eb36f0393196535064bfab45cfead076f881a39dbbc7d2e73aa15487e57046a18ea3df1483f8b493f926a8d6e362e6ee7c
+  data.tar.gz: ac5657121fd5b657880fee907b3a4508087aff20a15ad057b1d43254796d767e908eec39952ac44a0aa65f814a55d36db8c9b56119effe647d8ea018bcb2e6b4

data/CHANGELOG.md

@@ -11,6 +11,18 @@
 - Allow a bounded exit grace when macOS cannot inspect an exiting worker;
   persistent monitor failures still terminate and reap a live worker.
 
+## [0.8.0](https://github.com/seuros/imprimatur/compare/v0.7.0...v0.8.0) (2026-10-27)
+
+
+### Features
+
+* **reader:** accept drop_repeated in the Ruby API ([4e537d6](https://github.com/seuros/imprimatur/commit/4e537d641bafd8d3cdb8b1c20d16f4cb766ad073))
+
+
+### Bug Fixes
+
+* **pdf:** read printed page numbers from headers too, not just footers ([c036434](https://github.com/seuros/imprimatur/commit/c0364341ee8fc36be2c97d25fd06da0aa76a4ad6))
+
 ## [0.7.0](https://github.com/seuros/imprimatur/compare/v0.6.1...v0.7.0) (2026-10-26)
 
 

data/Cargo.lock

@@ -513,7 +513,7 @@ dependencies = [
 
 [[package]]
 name = "imprimatur"
-version = "0.7.0"
+version = "0.8.0"
 dependencies = [
  "aes",
  "crc32fast",
@@ -536,7 +536,7 @@ dependencies = [
 
 [[package]]
 name = "imprimatur-cli"
-version = "0.7.0"
+version = "0.8.0"
 dependencies = [
  "imprimatur",
  "regex",
@@ -548,7 +548,7 @@ dependencies = [
 
 [[package]]
 name = "imprimatur_native"
-version = "0.7.0"
+version = "0.8.0"
 dependencies = [
  "imprimatur",
  "magnus",

data/Cargo.toml

@@ -15,7 +15,7 @@ crc32fast = "1.5.2"
 datamatrix = "0.3.3"
 font-test-data = "0.10.0"
 hayro = { version = "0.8.0", default-features = false, features = ["embed-fonts", "embed-cmaps"] }
-imprimatur = { path = "ext/imprimatur_native/core", version = "0.7.0" }
+imprimatur = { path = "ext/imprimatur_native/core", version = "0.8.0" }
 libproc = "0.14.11"
 magnus = { version = "0.9.2", default-features = false }
 md-5 = "0.11.0"

data/README.md

@@ -32,7 +32,7 @@ end
 gem "imprimatur"
 ```
 
-Precompiled gems ship for Linux (x86_64 and aarch64, glibc 2.34+ and musl),
+Precompiled gems ship for Linux (x86_64 and aarch64, glibc 2.34+),
 macOS 13+ (arm64, x86_64) and FreeBSD 15.1 (amd64): nothing to compile.
 Anywhere else, the source gem compiles the extension on install, with the
 toolchain from [rustup.rs](https://rustup.rs) (1.99+).

data/ext/imprimatur_native/cli/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "imprimatur-cli"
-version = "0.7.0"
+version = "0.8.0"
 description = "Read PDF, XLSX, XLS, ODS and DOCX documents as Markdown or JSON, search them, extract their images."
 keywords = ["pdf", "xlsx", "docx", "markdown", "cli"]
 categories = ["command-line-utilities", "text-processing"]

data/ext/imprimatur_native/cli/src/cli/markdown.rs

@@ -16,7 +16,8 @@ use crate::app::{self, App, Out, Status};
 /// PDF: the text layer, placed glyph by glyph, with headings by font size,
 /// running headers and footers left out, and pages after the first marked
 /// `<!-- page N -->` (or `<!-- page N (label L) -->` with its printed page
-/// label, from `/PageLabels` or a footer that looks like one). Encrypted
+/// label, from `/PageLabels` or a header or footer line that looks like
+/// one). Encrypted
 /// PDFs open when their user password is empty. XLSX, XLS and ODS: each
 /// sheet as a table, values as shown, then its comments. DOCX: headings,
 /// lists, emphasis, links and tables.

data/ext/imprimatur_native/core/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "imprimatur"
-version = "0.7.0"
+version = "0.8.0"
 description = "XLSX, PDF and DOCX authoring, and PDF, XLSX, XLS, ODS and DOCX reading, with its own XML, zip and PDF code."
 keywords = ["xlsx", "excel", "pdf", "docx", "spreadsheet"]
 categories = ["encoding", "text-processing"]

data/ext/imprimatur_native/core/src/read/markdown.rs

@@ -44,7 +44,10 @@ pub enum Mode {
 }
 
 fn pdf(out: &mut String, pages: &[Page]) {
-    for page in pages.iter().filter(|page| !page.blocks.is_empty()) {
+    // Every page past the first gets its marker, blocks or none: an empty
+    // page (left bare by `--drop-repeated`, or genuinely blank) still needs
+    // one, so the markers stay a reliable page count and index.
+    for page in pages {
         if page.number > 1 {
             end_block(out);
             // Writing to a String cannot fail.

data/ext/imprimatur_native/core/src/read/pdf/layout.rs

@@ -293,27 +293,41 @@ pub fn pages(
             reason = "`blocks` keeps its buffer for the next page"
         )]
         let blocks = blocks.drain(..).collect();
-        let label = labels.map_or_else(
-            || footer_label(page, &boilerplate),
-            |labels| labels.get(index).cloned().flatten(),
-        );
+        let label = labels
+            .map_or_else(|| printed_label(page), |labels| labels.get(index).cloned().flatten());
         out.push(Page { number, blocks, images: page.images, label });
     }
     out
 }
 
-fn footer_label(page: &PageLines, boilerplate: &[u64]) -> Option<String> {
-    let mut candidates = page.lines.iter().filter(|line: &&Line| {
-        line.margin != 0
-            && line.y < page.height * EDGE
-            && boilerplate.binary_search(&line.margin).is_err()
-    });
-    let line = candidates.next()?;
-    if candidates.next().is_some() {
+/// The page's own printed label, from a single header or footer line that
+/// looks like one: a bare page number or roman numeral ("`-i-`", "`7`"), or
+/// the same after a short label word ("`Page 70`" -> "`70`"). Unlike the
+/// boilerplate this document's other margin lines are checked against, a
+/// running "`Page 70`", "`Page 71`", ... header is not noise here: each
+/// instance *is* that page's label, not a repeated line to drop.
+fn printed_label(page: &PageLines) -> Option<String> {
+    let mut candidates =
+        page.lines.iter().filter(|line: &&Line| line.margin != 0).filter_map(|line| {
+            let text = page.text(line).trim();
+            page_number_label(text)
+        });
+    let label = candidates.next()?;
+    candidates.next().is_none().then_some(label)
+}
+
+/// `text` as a page label, if it looks like one, stripped of a leading
+/// label word such as "Page" when there is one.
+fn page_number_label(text: &str) -> Option<String> {
+    if looks_like_page_number(text) {
+        return Some(text.to_owned());
+    }
+    let word_end = text.find(|ch: char| !ch.is_ascii_alphabetic())?;
+    if word_end == 0 {
         return None;
     }
-    let text = page.text(line).trim();
-    looks_like_page_number(text).then(|| text.to_owned())
+    let rest = text[word_end..].trim_start_matches(|ch: char| !ch.is_alphanumeric());
+    looks_like_page_number(rest).then(|| rest.to_owned())
 }
 
 fn looks_like_page_number(text: &str) -> bool {

data/ext/imprimatur_native/ffi/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "imprimatur_native"
-version = "0.7.0"
+version = "0.8.0"
 description = "Ruby bindings for the imprimatur crate"
 publish = false
 edition.workspace = true

data/ext/imprimatur_native/ffi/src/lib.rs

@@ -323,13 +323,13 @@ fn init(ruby: &Ruby) -> Result<(), Error> {
     barcode::define(ruby, imprimatur)?;
 
     let native = imprimatur.define_module("Native")?;
-    native.define_singleton_method("markdown", function!(read::markdown_native, 4))?;
-    native.define_singleton_method("content", function!(read::content_native, 5))?;
-    native.define_singleton_method("images", function!(read::images_native, 4))?;
+    native.define_singleton_method("markdown", function!(read::markdown_native, 5))?;
+    native.define_singleton_method("content", function!(read::content_native, 6))?;
+    native.define_singleton_method("images", function!(read::images_native, 5))?;
     native.define_singleton_method("pdf_info", function!(pdf::info, 4))?;
     native.define_singleton_method("pdf_pages", function!(pdf::pages, 5))?;
     native.define_singleton_method("render_pdf_page", function!(pdf::render, 5))?;
-    native.define_singleton_method("sheets", function!(sheets::sheets_native, 5))?;
+    native.define_singleton_method("sheets", function!(sheets::sheets_native, 6))?;
     let zip_writer = native.define_class("ZipWriter", ruby.class_object())?;
     zip_writer.define_singleton_method("new", function!(archive::ZipWriter::new, 1))?;
     zip_writer.define_method("add_file", method!(archive::ZipWriter::add_file, 3))?;

data/ext/imprimatur_native/ffi/src/pdf.rs

@@ -7,7 +7,7 @@ use std::path::PathBuf;
 
 use crate::read::without_gvl;
 
-type Keywords = (Option<usize>, Option<u64>, Option<f64>);
+type Keywords = (Option<usize>, Option<u64>, Option<f64>, bool);
 
 pub fn info(
     ruby: &Ruby,

data/ext/imprimatur_native/ffi/src/read.rs

@@ -25,15 +25,19 @@ struct Read<T, F> {
 }
 
 /// Called by `Imprimatur.markdown`, which owns the keyword interface.
-/// `timeout` is in seconds; `nil` limits keep their defaults.
+/// `timeout` is in seconds; `nil` limits keep their defaults. `drop_repeated`
+/// drops PDF text repeated at the same spot on most pages, anywhere on the
+/// page (watermarks); off by default.
 pub fn markdown_native(
     ruby: &Ruby,
     document: RString,
     max_input: Option<usize>,
     max_inflated: Option<u64>,
     timeout: Option<f64>,
+    drop_repeated: bool,
 ) -> Result<RString, Error> {
-    let text = without_gvl(ruby, document, (max_input, max_inflated, timeout), markdown)?;
+    let text =
+        without_gvl(ruby, document, (max_input, max_inflated, timeout, drop_repeated), markdown)?;
     Ok(ruby.str_new(&text))
 }
 
@@ -45,9 +49,11 @@ pub fn content_native(
     max_input: Option<usize>,
     max_inflated: Option<u64>,
     timeout: Option<f64>,
+    drop_repeated: bool,
     styles: bool,
 ) -> Result<RString, Error> {
-    let read = without_gvl(ruby, document, (max_input, max_inflated, timeout), content)?;
+    let read =
+        without_gvl(ruby, document, (max_input, max_inflated, timeout, drop_repeated), content)?;
     let options = json::Options { styles };
     Ok(ruby.str_new(&Json::Object(json::content_with(&read, options)).to_string()))
 }
@@ -60,9 +66,13 @@ pub fn images_native(
     max_input: Option<usize>,
     max_inflated: Option<u64>,
     timeout: Option<f64>,
+    drop_repeated: bool,
 ) -> Result<RArray, Error> {
-    let found =
-        without_gvl(ruby, document, (max_input, max_inflated, timeout), |bytes, limits| {
+    let found = without_gvl(
+        ruby,
+        document,
+        (max_input, max_inflated, timeout, drop_repeated),
+        |bytes, limits| {
             let mut found = Vec::new();
             images(bytes, limits, |image| {
                 let description = Json::Object(json::image(&image)).to_string();
@@ -70,7 +80,8 @@ pub fn images_native(
                 Ok(())
             })?;
             Ok(found)
-        })?;
+        },
+    )?;
     let pairs = ruby.ary_new_capa(found.len());
     for (data, description) in found {
         let pair = [
@@ -86,9 +97,14 @@ pub fn images_native(
 fn limits(
     ruby: &Ruby,
     cancel: &Arc<AtomicBool>,
-    (max_input, max_inflated, timeout): (Option<usize>, Option<u64>, Option<f64>),
+    (max_input, max_inflated, timeout, drop_repeated): (
+        Option<usize>,
+        Option<u64>,
+        Option<f64>,
+        bool,
+    ),
 ) -> Result<Limits, Error> {
-    let mut limits = Limits::default().cancel_on(Arc::clone(cancel));
+    let mut limits = Limits::default().cancel_on(Arc::clone(cancel)).drop_repeated(drop_repeated);
     if let Some(bytes) = max_input {
         limits = limits.max_input(bytes);
     }
@@ -107,7 +123,7 @@ fn limits(
 pub fn without_gvl<T, F>(
     ruby: &Ruby,
     document: RString,
-    keywords: (Option<usize>, Option<u64>, Option<f64>),
+    keywords: (Option<usize>, Option<u64>, Option<f64>, bool),
     job: F,
 ) -> Result<T, Error>
 where

data/ext/imprimatur_native/ffi/src/sheets.rs

@@ -23,9 +23,11 @@ pub fn sheets_native(
     max_input: Option<usize>,
     max_inflated: Option<u64>,
     timeout: Option<f64>,
+    drop_repeated: bool,
     styles: bool,
 ) -> Result<RArray, Error> {
-    let read = without_gvl(ruby, document, (max_input, max_inflated, timeout), content)?;
+    let read =
+        without_gvl(ruby, document, (max_input, max_inflated, timeout, drop_repeated), content)?;
     let Some(sheets) = read.sheets() else {
         let kind = if matches!(read, Content::Pdf(_)) { "a PDF" } else { "a Word document" };
         let error = imprimatur::Error::UnsupportedDocument(format!("{kind} is no spreadsheet"));

data/lib/imprimatur/reader.rb

@@ -41,8 +41,8 @@ module Imprimatur
     end
 
     # @api private
-    def self.limits(timeout: 60, max_bytes: nil, max_inflated: nil)
-      [max_bytes, max_inflated, timeout && Float(timeout)]
+    def self.limits(timeout: 60, max_bytes: nil, max_inflated: nil, drop_repeated: false)
+      [max_bytes, max_inflated, timeout && Float(timeout), drop_repeated ? true : false]
     end
 
     # A path or an IO is read no further than +max_bytes+ (one byte past it

data/lib/imprimatur/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Imprimatur
-  VERSION = '0.7.0'
+  VERSION = '0.8.0'
 end

data/lib/imprimatur.rb

@@ -72,6 +72,8 @@ module Imprimatur
     # @param timeout [Numeric, nil] seconds the read may take (60), +nil+ for none
     # @param max_bytes [Integer, nil] the largest document accepted (512 MiB)
     # @param max_inflated [Integer, nil] decompressed bytes allowed in all (1 GiB)
+    # @param drop_repeated [Boolean] PDF only: drop text repeated at the same
+    #   spot on most pages, anywhere on the page (watermarks); off by default
     # @return [String] Markdown, UTF-8
     # @raise [NoTextLayer] a PDF of scanned pages: run OCR instead
     # @raise [PasswordRequired] a PDF that needs a password to open

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: imprimatur
 version: !ruby/object:Gem::Version
-  version: 0.7.0
+  version: 0.8.0
 platform: ruby
 authors:
 - Abdelkader Boudih