diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index 70518860e..369a2d5ee 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -40,7 +40,7 @@ jobs: sudo dpkg --add-architecture i386 sudo apt update sudo apt install -y gcc-multilib libssl-dev:i386 pkg-config:i386 - - uses: dtolnay/rust-toolchain@1.87.0 + - uses: dtolnay/rust-toolchain@1.88.0 with: targets: ${{ matrix.bits == 32 && 'i686-unknown-linux-gnu' || '' }} - uses: Swatinem/rust-cache@v2 @@ -73,7 +73,7 @@ jobs: runs-on: ubuntu-latest steps: - uses: actions/checkout@v4 - - uses: dtolnay/rust-toolchain@1.87.0 + - uses: dtolnay/rust-toolchain@1.88.0 with: components: clippy, rustfmt - uses: Swatinem/rust-cache@v2 diff --git a/Cargo.lock b/Cargo.lock index dcd154367..c3fd68af5 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -181,9 +181,9 @@ checksum = "bef38d45163c2f1dde094a7dfd33ccf595c92905c8f8f4fdc18d06fb1037718a" [[package]] name = "bitflags" -version = "2.8.0" +version = "2.9.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8f68f53c83ab957f72c32642f3868eec03eb974d1fb82e453128456482613d36" +checksum = "1b8e56985ec62d17e9c1001dc89c88ecd7dc08e47eba5ec7c29c7b5eeecde967" dependencies = [ "serde", ] @@ -214,9 +214,9 @@ checksum = "64fa3c856b712db6612c019f14756e64e4bcea13337a6b33b696333a9eaa2d06" [[package]] name = "bytemuck" -version = "1.21.0" +version = "1.23.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ef657dfab802224e671f5818e9a4935f9b1957ed18e58292690cc39e7a4092a3" +checksum = "5c76a5792e44e4abe34d3abf15636779261d45a7450612059293d1d2cfc63422" dependencies = [ "bytemuck_derive", ] @@ -964,6 +964,69 @@ dependencies = [ "url", ] +[[package]] +name = "hayro" +version = "0.1.0" +source = "git+https://github.com/LaurenzV/hayro?rev=d651e18#d651e188a747c188db2e206733eaf17e8f622a5d" +dependencies = [ + "bytemuck", + "hayro-interpret", + "image", + "kurbo", + "rustc-hash", + "smallvec", +] + +[[package]] +name = "hayro-font" +version = "0.1.0" +source = "git+https://github.com/LaurenzV/hayro?rev=d651e18#d651e188a747c188db2e206733eaf17e8f622a5d" +dependencies = [ + "log", + "phf", +] + +[[package]] +name = "hayro-interpret" +version = "0.1.0" +source = "git+https://github.com/LaurenzV/hayro?rev=d651e18#d651e188a747c188db2e206733eaf17e8f622a5d" +dependencies = [ + "bitflags 2.9.1", + "hayro-font", + "hayro-syntax", + "kurbo", + "log", + "phf", + "qcms", + "skrifa", + "smallvec", + "yoke 0.8.0", +] + +[[package]] +name = "hayro-syntax" +version = "0.0.1" +source = "git+https://github.com/LaurenzV/hayro?rev=d651e18#d651e188a747c188db2e206733eaf17e8f622a5d" +dependencies = [ + "flate2", + "kurbo", + "log", + "rustc-hash", + "smallvec", + "zune-jpeg", +] + +[[package]] +name = "hayro-write" +version = "0.1.0" +source = "git+https://github.com/LaurenzV/hayro?rev=d651e18#d651e188a747c188db2e206733eaf17e8f622a5d" +dependencies = [ + "flate2", + "hayro-syntax", + "log", + "pdf-writer", +] + [[package]] name = "heck" version = "0.5.0" @@ -1206,9 +1269,9 @@ checksum = "cb56e1aa765b4b4f3aadfab769793b7087bb03a4ea4920644a6d238e2df5b9ed" [[package]] name = "image" -version = "0.25.5" +version = "0.25.6" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "cd6f44aed642f18953a158afeb30206f4d50da59fbc66ecb53c66488de73563b" +checksum = "db35664ce6b9810857a38a906215e75a9c879f0696556a39f59c62829710251a" dependencies = [ "bytemuck", "byteorder-lite", @@ -1271,7 +1334,7 @@ version = "0.11.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "f37dccff2791ab604f9babef0ba14fbe0be30bd368dc541e2b08d07c8aa908f3" dependencies = [ - "bitflags 2.8.0", + "bitflags 2.9.1", "inotify-sys", "libc", ] @@ -1367,7 +1430,7 @@ dependencies = [ [[package]] name = "krilla" version = "0.4.0" -source = "git+https://github.com/LaurenzV/krilla?rev=20c14fe#20c14fefee5002566b3d6668b338bbe2168784e7" +source = "git+https://github.com/LaurenzV/krilla/?rev=2da9d6c#2da9d6c6cb6a6baa6379592c72ae4eb7e16aa7b4" dependencies = [ "base64", "bumpalo", @@ -1376,6 +1439,7 @@ dependencies = [ "float-cmp 0.10.0", "fxhash", "gif", + "hayro-write", "image-webp", "imagesize", "once_cell", @@ -1385,6 +1449,7 @@ dependencies = [ "rustybuzz", "siphasher", "skrifa", + "smallvec", "subsetter", "tiny-skia-path", "xmp-writer", @@ -1395,7 +1460,7 @@ dependencies = [ [[package]] name = "krilla-svg" version = "0.1.0" -source = "git+https://github.com/LaurenzV/krilla?rev=20c14fe#20c14fefee5002566b3d6668b338bbe2168784e7" +source = "git+https://github.com/LaurenzV/krilla/?rev=2da9d6c#2da9d6c6cb6a6baa6379592c72ae4eb7e16aa7b4" dependencies = [ "flate2", "fontdb", @@ -1408,9 +1473,9 @@ dependencies = [ [[package]] name = "kurbo" -version = "0.11.1" +version = "0.11.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "89234b2cc610a7dd927ebde6b41dd1a5d4214cffaef4cf1fb2195d592f92518f" +checksum = "1077d333efea6170d9ccb96d3c3026f300ca0773da4938cc4c811daa6df68b0c" dependencies = [ "arrayvec", "smallvec", @@ -1462,7 +1527,7 @@ version = "0.1.3" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "c0ff37bd590ca25063e35af745c343cb7a0271906fb7b37e4813e8f79f00268d" dependencies = [ - "bitflags 2.8.0", + "bitflags 2.9.1", "libc", "redox_syscall", ] @@ -1628,7 +1693,7 @@ version = "8.0.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "2fee8403b3d66ac7b26aee6e40a897d85dc5ce26f44da36b8b73e987cc52e943" dependencies = [ - "bitflags 2.8.0", + "bitflags 2.9.1", "filetime", "fsevent-sys", "inotify", @@ -1710,7 +1775,7 @@ version = "0.10.72" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "fedfea7d58a1f73118430a55da6a286e7b044961736ce96a16a17068ea25e5da" dependencies = [ - "bitflags 2.8.0", + "bitflags 2.9.1", "cfg-if", "foreign-types", "libc", @@ -1847,7 +1912,7 @@ version = "0.13.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "3ea27c5015ab81753fc61e49f8cde74999346605ee148bb20008ef3d3150e0dc" dependencies = [ - "bitflags 2.8.0", + "bitflags 2.9.1", "itoa", "memchr", "ryu", @@ -2005,7 +2070,7 @@ version = "0.9.6" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "57206b407293d2bcd3af849ce869d52068623f19e1b5ff8e8778e3309439682b" dependencies = [ - "bitflags 2.8.0", + "bitflags 2.9.1", "getopts", "memchr", "unicase", @@ -2118,7 +2183,7 @@ version = "0.5.8" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "03a862b389f93e68874fbf580b9de08dd02facb9a788ebadaf4a3fd33cf58834" dependencies = [ - "bitflags 2.8.0", + "bitflags 2.9.1", ] [[package]] @@ -2221,7 +2286,7 @@ version = "0.38.44" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "fdb5bc1ae2baa591800df16c9ca78619bf65c0488b41b96ccec5d11220d8c154" dependencies = [ - "bitflags 2.8.0", + "bitflags 2.9.1", "errno", "libc", "linux-raw-sys", @@ -2240,7 +2305,7 @@ version = "0.20.1" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "fd3c7c96f8a08ee34eff8857b11b49b07d71d1c3f4e88f8a88d4c9e9f90b1702" dependencies = [ - "bitflags 2.8.0", + "bitflags 2.9.1", "bytemuck", "core_maths", "log", @@ -2288,7 +2353,7 @@ version = "2.11.1" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "897b2245f0b511c87893af39b033e5ca9cce68824c4d7e7630b5a1d339658d02" dependencies = [ - "bitflags 2.8.0", + "bitflags 2.9.1", "core-foundation", "core-foundation-sys", "libc", @@ -2451,9 +2516,9 @@ dependencies = [ [[package]] name = "smallvec" -version = "1.13.2" +version = "1.15.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "3c5e1a9a646d36c3599cd173a41282daf47c44583ad367b8e6837255952e5c67" +checksum = "67b1b7a3b5fe4f1376887184045fcf45c69e92af734b7aaddc05fb777b6fbd03" [[package]] name = "spin" @@ -2861,7 +2926,7 @@ dependencies = [ [[package]] name = "typst-assets" version = "0.13.1" -source = "git+https://github.com/typst/typst-assets?rev=edf0d64#edf0d648376e29738a05a933af9ea99bb81557b1" +source = "git+https://github.com/LaurenzV/typst-assets?rev=d89cc82#d89cc821f84a5667714491019c0b64087b2608bd" [[package]] name = "typst-cli" @@ -2911,7 +2976,7 @@ dependencies = [ [[package]] name = "typst-dev-assets" version = "0.13.1" -source = "git+https://github.com/typst/typst-dev-assets?rev=bfa947f#bfa947f3433d7d13a995168c40ae788a2ebfe648" +source = "git+https://github.com/LaurenzV/typst-dev-assets?rev=180c145#180c145cf810c8b2a7ed77355b351f3aed07d0c6" [[package]] name = "typst-docs" @@ -3063,7 +3128,7 @@ name = "typst-library" version = "0.13.1" dependencies = [ "az", - "bitflags 2.8.0", + "bitflags 2.9.1", "bumpalo", "chinese-number", "ciborium", @@ -3075,6 +3140,7 @@ dependencies = [ "fontdb", "glidesort", "hayagriva", + "hayro-syntax", "icu_properties", "icu_provider", "icu_provider_blob", @@ -3173,6 +3239,7 @@ version = "0.13.1" dependencies = [ "bytemuck", "comemo", + "hayro", "image", "pixglyph", "resvg", @@ -3191,6 +3258,7 @@ dependencies = [ "comemo", "ecow", "flate2", + "hayro", "image", "ttf-parser", "typst-library", @@ -3589,7 +3657,7 @@ version = "0.221.2" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "9845c470a2e10b61dd42c385839cdd6496363ed63b5c9e420b5488b77bd22083" dependencies = [ - "bitflags 2.8.0", + "bitflags 2.9.1", "indexmap 2.7.1", ] @@ -3724,7 +3792,7 @@ version = "0.33.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "3268f3d866458b787f390cf61f4bbb563b922d091359f9608842999eaee3943c" dependencies = [ - "bitflags 2.8.0", + "bitflags 2.9.1", ] [[package]] diff --git a/Cargo.toml b/Cargo.toml index 63ea32b94..ff2c2acaf 100644 --- a/Cargo.toml +++ b/Cargo.toml @@ -32,8 +32,8 @@ typst-svg = { path = "crates/typst-svg", version = "0.13.1" } typst-syntax = { path = "crates/typst-syntax", version = "0.13.1" } typst-timing = { path = "crates/typst-timing", version = "0.13.1" } typst-utils = { path = "crates/typst-utils", version = "0.13.1" } -typst-assets = { git = "https://github.com/typst/typst-assets", rev = "edf0d64" } -typst-dev-assets = { git = "https://github.com/typst/typst-dev-assets", rev = "bfa947f" } +typst-assets = { git = "https://github.com/LaurenzV/typst-assets", rev = "d89cc82" } +typst-dev-assets = { git = "https://github.com/LaurenzV/typst-dev-assets", rev = "180c145" } arrayvec = "0.7.4" az = "1.2" base64 = "0.22" @@ -61,6 +61,8 @@ fontdb = { version = "0.23", default-features = false } fs_extra = "1.3" glidesort = "0.1.2" hayagriva = "0.8.1" +hayro-syntax = { git = "https://github.com/LaurenzV/hayro", rev = "d651e18" } +hayro = { git = "https://github.com/LaurenzV/hayro", rev = "d651e18" } heck = "0.5" hypher = "0.1.4" icu_properties = { version = "1.4", features = ["serde"] } @@ -73,8 +75,8 @@ image = { version = "0.25.5", default-features = false, features = ["png", "jpeg indexmap = { version = "2", features = ["serde"] } infer = { version = "0.19.0", default-features = false } kamadak-exif = "0.6" -krilla = { git = "https://github.com/LaurenzV/krilla", rev = "20c14fe", default-features = false, features = ["raster-images", "comemo", "rayon"] } -krilla-svg = { git = "https://github.com/LaurenzV/krilla", rev = "20c14fe" } +krilla = { git = "https://github.com/LaurenzV/krilla/", rev = "2da9d6c", default-features = false, features = ["raster-images", "comemo", "rayon", "pdf"] } +krilla-svg = { git = "https://github.com/LaurenzV/krilla/", rev = "2da9d6c"} kurbo = "0.11" libfuzzer-sys = "0.4" lipsum = "0.9" diff --git a/crates/typst-cli/src/args.rs b/crates/typst-cli/src/args.rs index 7459be0f2..093735cbb 100644 --- a/crates/typst-cli/src/args.rs +++ b/crates/typst-cli/src/args.rs @@ -471,6 +471,7 @@ display_possible_values!(DiagnosticFormat); #[derive(Debug, Copy, Clone, Eq, PartialEq, ValueEnum)] pub enum Feature { Html, + PdfEmbedding, } display_possible_values!(Feature); diff --git a/crates/typst-cli/src/world.rs b/crates/typst-cli/src/world.rs index 8ad766b14..ab5216831 100644 --- a/crates/typst-cli/src/world.rs +++ b/crates/typst-cli/src/world.rs @@ -117,6 +117,7 @@ impl SystemWorld { .iter() .map(|&feature| match feature { Feature::Html => typst::Feature::Html, + Feature::PdfEmbedding => typst::Feature::PdfEmbedding, }) .collect(); diff --git a/crates/typst-library/Cargo.toml b/crates/typst-library/Cargo.toml index f4b219882..2297f7ca7 100644 --- a/crates/typst-library/Cargo.toml +++ b/crates/typst-library/Cargo.toml @@ -31,6 +31,7 @@ flate2 = { workspace = true } fontdb = { workspace = true } glidesort = { workspace = true } hayagriva = { workspace = true } +hayro-syntax = { workspace = true } icu_properties = { workspace = true } icu_provider = { workspace = true } icu_provider_blob = { workspace = true } diff --git a/crates/typst-library/src/lib.rs b/crates/typst-library/src/lib.rs index 025e997c6..edbc182fa 100644 --- a/crates/typst-library/src/lib.rs +++ b/crates/typst-library/src/lib.rs @@ -237,6 +237,7 @@ impl FromIterator for Features { #[non_exhaustive] pub enum Feature { Html, + PdfEmbedding, } /// A group of related standard library definitions. diff --git a/crates/typst-library/src/visualize/image/mod.rs b/crates/typst-library/src/visualize/image/mod.rs index f1fa6381b..26a486089 100644 --- a/crates/typst-library/src/visualize/image/mod.rs +++ b/crates/typst-library/src/visualize/image/mod.rs @@ -1,8 +1,10 @@ //! Image handling. +mod pdf; mod raster; mod svg; +pub use self::pdf::PdfImage; pub use self::raster::{ ExchangeFormat, PixelEncoding, PixelFormat, RasterFormat, RasterImage, }; @@ -13,10 +15,12 @@ use std::fmt::{self, Debug, Formatter}; use std::sync::Arc; use ecow::EcoString; +use hayro_syntax::LoadPdfError; +use typst_library::{Feature, World}; use typst_syntax::{Span, Spanned}; use typst_utils::LazyHash; -use crate::diag::{warning, At, LoadedWithin, SourceResult, StrResult}; +use crate::diag::{bail, warning, At, LoadedWithin, SourceResult, StrResult}; use crate::engine::Engine; use crate::foundations::{ cast, elem, func, scope, Bytes, Cast, Content, Derived, NativeElement, Packed, Smart, @@ -26,6 +30,7 @@ use crate::layout::{Length, Rel, Sizing}; use crate::loading::{DataSource, Load, LoadSource, Loaded, Readable}; use crate::model::Figurable; use crate::text::{families, LocalName}; +use crate::visualize::image::pdf::PdfDocument; /// A raster or vector graphic. /// @@ -126,6 +131,11 @@ pub struct ImageElem { /// A text describing the image. pub alt: Option, + /// The page number that should be embedded as an image. This attribute only has an effect + /// for PDF files. + #[default(1)] + pub page: usize, + /// How the image should adjust itself to a given area (the area is defined /// by the `width` and `height` fields). Note that `fit` doesn't visually /// change anything if the area's aspect ratio is the same as the image's @@ -261,6 +271,64 @@ impl Packed { ) .within(loaded)?, ), + ImageFormat::Vector(VectorFormat::Pdf) => { + if engine.world.library().features.is_enabled(Feature::PdfEmbedding) { + let document = match PdfDocument::new(loaded.data.clone()) { + Ok(doc) => doc, + Err(e) => match e { + LoadPdfError::Encryption => { + bail!( + span, + "the PDF is encrypted or password-protected"; + hint: "such PDFs are currently not supported"; + hint: "preprocess the PDF to remove the encryption" + ); + } + LoadPdfError::Invalid => { + bail!( + span, + "the PDF could not be loaded"; + hint: "perhaps the PDF file is malformed" + ); + } + }, + }; + + let page_num = self.page.get(styles); + + if page_num == 0 { + bail!( + span, + "{page_num} is not a valid page number"; + hint: "page numbers for PDF start at 1" + ) + }; + + // The user provides the page number start from 1, but further down the pipeline, + // page numbers are 0-based. + let page_idx = page_num - 1; + let num_pages = document.len(); + + let Some(pdf_image) = PdfImage::new(document, page_idx) else { + let pages = if num_pages == 1 { "page" } else { "pages" }; + + bail!( + span, + "page {page_num} doesn't exist"; + hint: "the document only has {num_pages} {pages}" + ); + }; + + ImageKind::Pdf(pdf_image) + } else { + bail!( + span, + "embedding PDFs is currently an experimental, opt-in feature"; + hint: "enable the corresponding feature to try it out"; + hint: "convert your PDF to SVG instead" + ); + } + } }; Ok(Image::new(kind, self.alt.get_cloned(styles), self.scaling.get(styles))) @@ -286,6 +354,7 @@ impl Packed { "jpg" | "jpeg" => return Ok(ExchangeFormat::Jpg.into()), "gif" => return Ok(ExchangeFormat::Gif.into()), "svg" | "svgz" => return Ok(VectorFormat::Svg.into()), + "pdf" => return Ok(VectorFormat::Pdf.into()), "webp" => return Ok(ExchangeFormat::Webp.into()), _ => {} } @@ -373,6 +442,7 @@ impl Image { match &self.0.kind { ImageKind::Raster(raster) => raster.format().into(), ImageKind::Svg(_) => VectorFormat::Svg.into(), + ImageKind::Pdf(_) => VectorFormat::Pdf.into(), } } @@ -381,6 +451,7 @@ impl Image { match &self.0.kind { ImageKind::Raster(raster) => raster.width() as f64, ImageKind::Svg(svg) => svg.width(), + ImageKind::Pdf(pdf) => pdf.width() as f64, } } @@ -389,6 +460,7 @@ impl Image { match &self.0.kind { ImageKind::Raster(raster) => raster.height() as f64, ImageKind::Svg(svg) => svg.height(), + ImageKind::Pdf(pdf) => pdf.height() as f64, } } @@ -397,6 +469,7 @@ impl Image { match &self.0.kind { ImageKind::Raster(raster) => raster.dpi(), ImageKind::Svg(_) => Some(Image::USVG_DEFAULT_DPI), + ImageKind::Pdf(_) => Some(Image::DEFAULT_DPI), } } @@ -435,6 +508,8 @@ pub enum ImageKind { Raster(RasterImage), /// An SVG image. Svg(SvgImage), + /// A PDF image. + Pdf(PdfImage), } impl From for ImageKind { @@ -469,10 +544,20 @@ impl ImageFormat { return Some(Self::Vector(VectorFormat::Svg)); } + if is_pdf(data) { + return Some(Self::Vector(VectorFormat::Pdf)); + } + None } } +/// Checks whether the data looks like a PDF file. +fn is_pdf(data: &[u8]) -> bool { + let head = &data[..data.len().min(2048)]; + memchr::memmem::find(head, b"%PDF-").is_some() +} + /// Checks whether the data looks like an SVG or a compressed SVG. fn is_svg(data: &[u8]) -> bool { // Check for the gzip magic bytes. This check is perhaps a bit too @@ -493,6 +578,8 @@ fn is_svg(data: &[u8]) -> bool { pub enum VectorFormat { /// The vector graphics format of the web. Svg, + /// The PDF graphics format. + Pdf, } impl From for ImageFormat diff --git a/crates/typst-library/src/visualize/image/pdf.rs b/crates/typst-library/src/visualize/image/pdf.rs new file mode 100644 index 000000000..c4d6c4f18 --- /dev/null +++ b/crates/typst-library/src/visualize/image/pdf.rs @@ -0,0 +1,165 @@ +use std::hash::{Hash, Hasher}; +use std::sync::Arc; + +use hayro_syntax::page::Page; +use hayro_syntax::{LoadPdfError, Pdf}; + +use crate::foundations::Bytes; + +struct DocumentRepr { + pdf: Arc, + data: Bytes, + standard_fonts: Arc, +} + +impl Hash for DocumentRepr { + fn hash(&self, state: &mut H) { + self.data.hash(state); + } +} + +/// A PDF document. +#[derive(Clone, Hash)] +pub struct PdfDocument(Arc); + +impl PdfDocument { + /// Load a PDF document. + #[comemo::memoize] + #[typst_macros::time(name = "load pdf document")] + pub fn new(data: Bytes) -> Result { + let pdf = Arc::new(Pdf::new(Arc::new(data.clone()))?); + let standard_fonts = get_standard_fonts(); + + Ok(Self(Arc::new(DocumentRepr { data, pdf, standard_fonts }))) + } + + /// Return the number of pages in the PDF. + pub fn len(&self) -> usize { + self.0.pdf.pages().len() + } +} + +struct ImageRepr { + document: PdfDocument, + page_index: usize, + width: f32, + height: f32, +} + +impl Hash for ImageRepr { + fn hash(&self, state: &mut H) { + self.document.hash(state); + self.page_index.hash(state); + } +} + +/// A specific page of a PDF acting as an image. +#[derive(Clone, Hash)] +pub struct PdfImage(Arc); + +impl PdfImage { + /// Create a new PDF image. + /// + /// Returns `None` if the page index is not valid. + #[comemo::memoize] + pub fn new(document: PdfDocument, page: usize) -> Option { + let dimensions = document.0.pdf.pages().get(page)?.render_dimensions(); + + Some(Self(Arc::new(ImageRepr { + document, + page_index: page, + width: dimensions.0, + height: dimensions.1, + }))) + } + + /// Returns the PDF page of the image. + pub fn page(&self) -> &Page { + &self.pdf().pages()[self.0.page_index] + } + + /// Returns the underlying PDF document. + pub fn pdf(&self) -> &Arc { + &self.0.document.0.pdf + } + + /// Returns the width of the image. + pub fn width(&self) -> f32 { + self.0.width + } + + /// Returns the embedded standard fonts of the image. + pub fn standard_fonts(&self) -> &Arc { + &self.0.document.0.standard_fonts + } + + /// Returns the height of the image. + pub fn height(&self) -> f32 { + self.0.height + } + + /// Returns the page index of the image. + pub fn page_index(&self) -> usize { + self.0.page_index + } + + /// Returns the underlying Typst PDF document. + pub fn document(&self) -> &PdfDocument { + &self.0.document + } +} + +#[comemo::memoize] +fn get_standard_fonts() -> Arc { + let helvetica = VariantFont { + normal: Bytes::new(typst_assets::pdf::SANS), + bold: Bytes::new(typst_assets::pdf::SANS_BOLD), + italic: Bytes::new(typst_assets::pdf::SANS_ITALIC), + bold_italic: Bytes::new(typst_assets::pdf::SANS_BOLD_ITALIC), + }; + + let courier = VariantFont { + normal: Bytes::new(typst_assets::pdf::FIXED), + bold: Bytes::new(typst_assets::pdf::FIXED_BOLD), + italic: Bytes::new(typst_assets::pdf::FIXED_ITALIC), + bold_italic: Bytes::new(typst_assets::pdf::FIXED_BOLD_ITALIC), + }; + + let times = VariantFont { + normal: Bytes::new(typst_assets::pdf::SERIF), + bold: Bytes::new(typst_assets::pdf::SERIF_BOLD), + italic: Bytes::new(typst_assets::pdf::SERIF_ITALIC), + bold_italic: Bytes::new(typst_assets::pdf::SERIF_BOLD_ITALIC), + }; + + let symbol = Bytes::new(typst_assets::pdf::SYMBOL); + let zapf_dingbats = Bytes::new(typst_assets::pdf::DING_BATS); + + Arc::new(StandardFonts { helvetica, courier, times, symbol, zapf_dingbats }) +} + +/// A PDF font with multiple variants. +pub struct VariantFont { + /// The normal variant. + pub normal: Bytes, + /// The bold variant. + pub bold: Bytes, + /// The italic variant. + pub italic: Bytes, + /// The bold-italic variant. + pub bold_italic: Bytes, +} + +/// A structure holding the raw data of all PDF standard fonts. +pub struct StandardFonts { + /// The data for the `Helvetica` font family. + pub helvetica: VariantFont, + /// The data for the `Courier` font family. + pub courier: VariantFont, + /// The data for the `Times` font family. + pub times: VariantFont, + /// The data for the `Symbol` font family. + pub symbol: Bytes, + /// The data for the `Zapf Dingbats` font family. + pub zapf_dingbats: Bytes, +} diff --git a/crates/typst-pdf/src/convert.rs b/crates/typst-pdf/src/convert.rs index 9e2aa87b7..71a49394a 100644 --- a/crates/typst-pdf/src/convert.rs +++ b/crates/typst-pdf/src/convert.rs @@ -9,6 +9,7 @@ use krilla::embed::EmbedError; use krilla::error::KrillaError; use krilla::geom::PathBuilder; use krilla::page::{PageLabel, PageSettings}; +use krilla::pdf::PdfError; use krilla::surface::Surface; use krilla::{Document, SerializeSettings}; use krilla_svg::render_svg_glyph; @@ -363,6 +364,39 @@ fn finish( hint: "convert the image to 8 bit instead" ) } + KrillaError::Pdf(_, e, loc) => { + let span = to_span(loc); + match e { + // We already validated in `typst-library` that the page index is valid. + PdfError::InvalidPage(_) => unreachable!(), + PdfError::VersionMismatch(v) => { + let pdf_ver = v.as_str(); + let config_ver = configuration.version(); + let cur_ver = config_ver.as_str(); + + bail!(span, + "the version of the PDF file is too high"; + hint: "the current export target is {cur_ver}, while the PDF has version {pdf_ver}"; + hint: "raise the export target to {pdf_ver} or higher"; + hint: "preprocess the PDF to convert it to a lower version" + ); + } + } + } + KrillaError::DuplicateTagId(_, loc) => { + let span = to_span(loc); + bail!(span, + "duplicate tag id"; + hint: "this is a bug in typst, please report it" + ); + } + KrillaError::UnknownTagId(_, loc) => { + let span = to_span(loc); + bail!(span, + "unknown tag id"; + hint: "this is a bug in typst, please report it" + ); + } }, } } @@ -535,12 +569,12 @@ fn convert_error( } // The below errors cannot occur yet, only once Typst supports full PDF/A // and PDF/UA. But let's still add a message just to be on the safe side. - ValidationError::MissingAnnotationAltText => error!( + ValidationError::MissingAnnotationAltText(_) => error!( Span::detached(), "{prefix} missing annotation alt text"; hint: "please report this as a bug" ), - ValidationError::MissingAltText => error!( + ValidationError::MissingAltText(_) => error!( Span::detached(), "{prefix} missing alt text"; hint: "make sure your images and equations have alt text" @@ -576,6 +610,13 @@ fn convert_error( "{prefix} missing document date"; hint: "set the date of the document" ), + ValidationError::EmbeddedPDF(loc) => { + error!( + to_span(*loc), + "embedding PDFs is currently not supported in this export mode"; + hint: "try converting the PDF to SVG before embedding it" + ) + } } } diff --git a/crates/typst-pdf/src/image.rs b/crates/typst-pdf/src/image.rs index 93bdb1950..e23652a05 100644 --- a/crates/typst-pdf/src/image.rs +++ b/crates/typst-pdf/src/image.rs @@ -3,13 +3,14 @@ use std::sync::{Arc, OnceLock}; use image::{DynamicImage, EncodableLayout, GenericImageView, Rgba}; use krilla::image::{BitsPerComponent, CustomImage, ImageColorspace}; +use krilla::pdf::PdfDocument; use krilla::surface::Surface; use krilla_svg::{SurfaceExt, SvgSettings}; use typst_library::diag::{bail, SourceResult}; use typst_library::foundations::Smart; use typst_library::layout::{Abs, Angle, Ratio, Size, Transform}; use typst_library::visualize::{ - ExchangeFormat, Image, ImageKind, ImageScaling, RasterFormat, RasterImage, + ExchangeFormat, Image, ImageKind, ImageScaling, PdfImage, RasterFormat, RasterImage, }; use typst_syntax::Span; @@ -60,6 +61,9 @@ pub(crate) fn handle_image( SvgSettings { embed_text: true, ..Default::default() }, ); } + ImageKind::Pdf(pdf) => { + surface.draw_pdf_page(&convert_pdf(pdf), size.to_krilla(), pdf.page_index()) + } } if image.alt().is_some() { @@ -85,9 +89,9 @@ struct Repr { /// A wrapper around `RasterImage` so that we can implement `CustomImage`. #[derive(Clone)] -struct PdfImage(Arc); +struct PdfRasterImage(Arc); -impl PdfImage { +impl PdfRasterImage { pub fn new(raster: RasterImage) -> Self { Self(Arc::new(Repr { raster, @@ -97,7 +101,7 @@ impl PdfImage { } } -impl Hash for PdfImage { +impl Hash for PdfRasterImage { fn hash(&self, state: &mut H) { // `alpha_channel` and `actual_dynamic` are generated from the underlying `RasterImage`, // so this is enough. Since `raster` is prehashed, this is also very cheap. @@ -105,7 +109,7 @@ impl Hash for PdfImage { } } -impl CustomImage for PdfImage { +impl CustomImage for PdfRasterImage { fn color_channel(&self) -> &[u8] { self.0 .actual_dynamic @@ -196,10 +200,15 @@ fn convert_raster( interpolate, ) } else { - krilla::image::Image::from_custom(PdfImage::new(raster), interpolate) + krilla::image::Image::from_custom(PdfRasterImage::new(raster), interpolate) } } +#[comemo::memoize] +fn convert_pdf(pdf: &PdfImage) -> PdfDocument { + PdfDocument::new(pdf.pdf().clone()) +} + fn exif_transform(image: &RasterImage, size: Size) -> (Transform, Size) { let base = |hp: bool, vp: bool, mut base_ts: Transform, size: Size| { if hp { diff --git a/crates/typst-pdf/src/link.rs b/crates/typst-pdf/src/link.rs index 64cb8f0a2..2133be82c 100644 --- a/crates/typst-pdf/src/link.rs +++ b/crates/typst-pdf/src/link.rs @@ -49,7 +49,6 @@ pub(crate) fn handle_link( fc.push_annotation( LinkAnnotation::new( rect, - None, Target::Action(Action::Link(LinkAction::new(u.to_string()))), ) .into(), @@ -64,7 +63,6 @@ pub(crate) fn handle_link( fc.push_annotation( LinkAnnotation::new( rect, - None, Target::Destination(krilla::destination::Destination::Named( nd.clone(), )), @@ -83,7 +81,6 @@ pub(crate) fn handle_link( fc.push_annotation( LinkAnnotation::new( rect, - None, Target::Destination(krilla::destination::Destination::Xyz( XyzDestination::new(index, pos.point.to_krilla()), )), diff --git a/crates/typst-render/Cargo.toml b/crates/typst-render/Cargo.toml index 7d01d7e38..5032344ee 100644 --- a/crates/typst-render/Cargo.toml +++ b/crates/typst-render/Cargo.toml @@ -18,6 +18,7 @@ typst-macros = { workspace = true } typst-timing = { workspace = true } bytemuck = { workspace = true } comemo = { workspace = true } +hayro = { workspace = true } image = { workspace = true } pixglyph = { workspace = true } resvg = { workspace = true } diff --git a/crates/typst-render/src/image.rs b/crates/typst-render/src/image.rs index 7425bdd2f..7385794a8 100644 --- a/crates/typst-render/src/image.rs +++ b/crates/typst-render/src/image.rs @@ -1,11 +1,12 @@ -use std::sync::Arc; - +use hayro::{FontData, FontQuery, InterpreterSettings, RenderSettings, StandardFont}; use image::imageops::FilterType; use image::{GenericImageView, Rgba}; +use std::sync::Arc; use tiny_skia as sk; +use tiny_skia::IntSize; use typst_library::foundations::Smart; use typst_library::layout::Size; -use typst_library::visualize::{Image, ImageKind, ImageScaling}; +use typst_library::visualize::{Image, ImageKind, ImageScaling, PdfImage}; use crate::{AbsExt, State}; @@ -59,9 +60,9 @@ pub fn render_image( /// Prepare a texture for an image at a scaled size. #[comemo::memoize] fn build_texture(image: &Image, w: u32, h: u32) -> Option> { - let mut texture = sk::Pixmap::new(w, h)?; - match image.kind() { + let texture = match image.kind() { ImageKind::Raster(raster) => { + let mut texture = sk::Pixmap::new(w, h)?; let w = texture.width(); let h = texture.height(); @@ -85,15 +86,70 @@ fn build_texture(image: &Image, w: u32, h: u32) -> Option> { let Rgba([r, g, b, a]) = src; *dest = sk::ColorU8::from_rgba(r, g, b, a).premultiply(); } + + texture } ImageKind::Svg(svg) => { + let mut texture = sk::Pixmap::new(w, h)?; let tree = svg.tree(); let ts = tiny_skia::Transform::from_scale( w as f32 / tree.size().width(), h as f32 / tree.size().height(), ); resvg::render(tree, ts, &mut texture.as_mut()); + + texture } - } + ImageKind::Pdf(pdf) => build_pdf_texture(pdf, w, h)?, + }; + Some(Arc::new(texture)) } + +// Keep this in sync with `typst-svg`! +fn build_pdf_texture(pdf: &PdfImage, w: u32, h: u32) -> Option { + let sf = pdf.standard_fonts().clone(); + + let select_standard_font = move |font: StandardFont| -> Option<(FontData, u32)> { + let bytes = match font { + StandardFont::Helvetica => sf.helvetica.normal.clone(), + StandardFont::HelveticaBold => sf.helvetica.bold.clone(), + StandardFont::HelveticaOblique => sf.helvetica.italic.clone(), + StandardFont::HelveticaBoldOblique => sf.helvetica.bold_italic.clone(), + StandardFont::Courier => sf.courier.normal.clone(), + StandardFont::CourierBold => sf.courier.bold.clone(), + StandardFont::CourierOblique => sf.courier.italic.clone(), + StandardFont::CourierBoldOblique => sf.courier.bold_italic.clone(), + StandardFont::TimesRoman => sf.times.normal.clone(), + StandardFont::TimesBold => sf.times.bold.clone(), + StandardFont::TimesItalic => sf.times.italic.clone(), + StandardFont::TimesBoldItalic => sf.times.bold_italic.clone(), + StandardFont::ZapfDingBats => sf.zapf_dingbats.clone(), + StandardFont::Symbol => sf.symbol.clone(), + }; + + let font_data: Arc + Send + Sync> = Arc::new(bytes.clone()); + + Some((font_data, 0)) + }; + + let interpreter_settings = InterpreterSettings { + font_resolver: Arc::new(move |query| match query { + FontQuery::Standard(s) => select_standard_font(*s), + FontQuery::Fallback(f) => select_standard_font(f.pick_standard_font()), + }), + warning_sink: Arc::new(|_| {}), + }; + let page = pdf.page(); + + let render_settings = RenderSettings { + x_scale: w as f32 / pdf.width(), + y_scale: h as f32 / pdf.height(), + width: Some(w as u16), + height: Some(h as u16), + }; + + let hayro_pix = hayro::render(page, &interpreter_settings, &render_settings); + + sk::Pixmap::from_vec(hayro_pix.take_u8(), IntSize::from_wh(w, h)?) +} diff --git a/crates/typst-svg/Cargo.toml b/crates/typst-svg/Cargo.toml index 5416621e5..e1984d54e 100644 --- a/crates/typst-svg/Cargo.toml +++ b/crates/typst-svg/Cargo.toml @@ -21,6 +21,7 @@ base64 = { workspace = true } comemo = { workspace = true } ecow = { workspace = true } flate2 = { workspace = true } +hayro = { workspace = true } image = { workspace = true } ttf-parser = { workspace = true } xmlparser = { workspace = true } diff --git a/crates/typst-svg/src/image.rs b/crates/typst-svg/src/image.rs index fd4aecd4f..40a982462 100644 --- a/crates/typst-svg/src/image.rs +++ b/crates/typst-svg/src/image.rs @@ -1,10 +1,14 @@ +use std::borrow::Cow; +use std::sync::Arc; + use base64::Engine; use ecow::{eco_format, EcoString}; +use hayro::{FontData, FontQuery, InterpreterSettings, RenderSettings, StandardFont}; use image::{codecs::png::PngEncoder, ImageEncoder}; use typst_library::foundations::Smart; use typst_library::layout::{Abs, Axes}; use typst_library::visualize::{ - ExchangeFormat, Image, ImageKind, ImageScaling, RasterFormat, + ExchangeFormat, Image, ImageKind, ImageScaling, PdfImage, RasterFormat, }; use crate::SVGRenderer; @@ -44,7 +48,7 @@ pub fn convert_image_scaling(scaling: Smart) -> Option<&'static st #[comemo::memoize] pub fn convert_image_to_base64_url(image: &Image) -> EcoString { let mut buf; - let (format, data): (&str, &[u8]) = match image.kind() { + let (format, data): (&str, Cow<[u8]>) = match image.kind() { ImageKind::Raster(raster) => match raster.format() { RasterFormat::Exchange(format) => ( match format { @@ -53,7 +57,7 @@ pub fn convert_image_to_base64_url(image: &Image) -> EcoString { ExchangeFormat::Gif => "gif", ExchangeFormat::Webp => "webp", }, - raster.data(), + Cow::Borrowed(raster.data()), ), RasterFormat::Pixel(_) => ("png", { buf = vec![]; @@ -62,10 +66,29 @@ pub fn convert_image_to_base64_url(image: &Image) -> EcoString { encoder.set_icc_profile(icc_profile.to_vec()).ok(); } raster.dynamic().write_with_encoder(encoder).unwrap(); - buf.as_slice() + Cow::Borrowed(buf.as_slice()) }), }, - ImageKind::Svg(svg) => ("svg+xml", svg.data()), + ImageKind::Svg(svg) => ("svg+xml", Cow::Borrowed(svg.data())), + ImageKind::Pdf(pdf) => { + // To make sure the image isn't pixelated, we always scale up so the lowest + // dimension has at least 1000 pixels. However, we only scale up as much so that the + // largest dimension doesn't exceed 3000 pixels. + const MIN_RES: f32 = 1000.0; + const MAX_RES: f32 = 3000.0; + + let base_width = pdf.width(); + let w_scale = (MIN_RES / base_width).max(MAX_RES / base_width); + let base_height = pdf.height(); + let h_scale = (MIN_RES / base_height).min(MAX_RES / base_height); + + let total_scale = w_scale.min(h_scale); + + let width = (base_width * total_scale).ceil() as u32; + let height = (base_height * total_scale).ceil() as u32; + + ("png", Cow::Owned(pdf_to_png(pdf, width, height))) + } }; let mut url = eco_format!("data:image/{format};base64,"); @@ -73,3 +96,52 @@ pub fn convert_image_to_base64_url(image: &Image) -> EcoString { url.push_str(&data); url } + +// Keep this in sync with `typst-png`! +#[comemo::memoize] +fn pdf_to_png(pdf: &PdfImage, w: u32, h: u32) -> Vec { + let sf = pdf.standard_fonts().clone(); + + let select_standard_font = move |font: StandardFont| -> Option<(FontData, u32)> { + let bytes = match font { + StandardFont::Helvetica => sf.helvetica.normal.clone(), + StandardFont::HelveticaBold => sf.helvetica.bold.clone(), + StandardFont::HelveticaOblique => sf.helvetica.italic.clone(), + StandardFont::HelveticaBoldOblique => sf.helvetica.bold_italic.clone(), + StandardFont::Courier => sf.courier.normal.clone(), + StandardFont::CourierBold => sf.courier.bold.clone(), + StandardFont::CourierOblique => sf.courier.italic.clone(), + StandardFont::CourierBoldOblique => sf.courier.bold_italic.clone(), + StandardFont::TimesRoman => sf.times.normal.clone(), + StandardFont::TimesBold => sf.times.bold.clone(), + StandardFont::TimesItalic => sf.times.italic.clone(), + StandardFont::TimesBoldItalic => sf.times.bold_italic.clone(), + StandardFont::ZapfDingBats => sf.zapf_dingbats.clone(), + StandardFont::Symbol => sf.symbol.clone(), + }; + + let font_data: Arc + Send + Sync> = Arc::new(bytes.clone()); + + Some((font_data, 0)) + }; + + let interpreter_settings = InterpreterSettings { + font_resolver: Arc::new(move |query| match query { + FontQuery::Standard(s) => select_standard_font(*s), + FontQuery::Fallback(f) => select_standard_font(f.pick_standard_font()), + }), + warning_sink: Arc::new(|_| {}), + }; + let page = pdf.page(); + + let render_settings = RenderSettings { + x_scale: w as f32 / pdf.width(), + y_scale: h as f32 / pdf.height(), + width: Some(w as u16), + height: Some(h as u16), + }; + + let hayro_pix = hayro::render(page, &interpreter_settings, &render_settings); + + hayro_pix.take_png() +} diff --git a/tests/ref/image-pdf-multiple-pages.png b/tests/ref/image-pdf-multiple-pages.png new file mode 100644 index 000000000..a6be4a313 Binary files /dev/null and b/tests/ref/image-pdf-multiple-pages.png differ diff --git a/tests/ref/image-pdf.png b/tests/ref/image-pdf.png new file mode 100644 index 000000000..6619dcee6 Binary files /dev/null and b/tests/ref/image-pdf.png differ diff --git a/tests/src/world.rs b/tests/src/world.rs index 4b6cf5a34..0c37b7bed 100644 --- a/tests/src/world.rs +++ b/tests/src/world.rs @@ -198,7 +198,7 @@ fn library() -> Library { // exactly 100pt wide. Page height is unbounded and font size is 10pt so // that it multiplies to nice round numbers. let mut lib = Library::builder() - .with_features([Feature::Html].into_iter().collect()) + .with_features([Feature::Html, Feature::PdfEmbedding].into_iter().collect()) .build(); // Hook up helpers into the global scope. diff --git a/tests/suite/visualize/image.typ b/tests/suite/visualize/image.typ index 36ec06cb1..cb192d4bb 100644 --- a/tests/suite/visualize/image.typ +++ b/tests/suite/visualize/image.typ @@ -258,7 +258,7 @@ A #box(image("/assets/images/tiger.jpg", height: 1cm, width: 80%)) B --- image-png-but-pixmap-format --- #image( read("/assets/images/tiger.jpg", encoding: none), - // Error: 11-18 expected "png", "jpg", "gif", "webp", dictionary, "svg", or auto + // Error: 11-18 expected "png", "jpg", "gif", "webp", dictionary, "svg", "pdf", or auto format: "rgba8", ) @@ -289,3 +289,16 @@ A #box(image("/assets/images/tiger.jpg", height: 1cm, width: 80%)) B ..rotations.map(v => raw(str(v), lang: "typc")), ..rotations.map(rotated) ) + +--- image-pdf --- +#image("/assets/images/matplotlib.pdf") + +--- image-pdf-invalid-page --- +// Error: 2-49 page 2 doesn't exist +// Hint: 2-49 the document only has 1 page +#image("/assets/images/matplotlib.pdf", page: 2) + +--- image-pdf-multiple-pages --- +#image("/assets/images/diagrams.pdf", page: 1) +#image("/assets/images/diagrams.pdf", page: 3) +#image("/assets/images/diagrams.pdf", page: 2) \ No newline at end of file