From ce6975e65a0e9f3b0d625735159583e9860e1fdb Mon Sep 17 00:00:00 2001
From: Tobias Schmitz <tobiasschmitz2001@gmail.com>
Date: Thu, 8 May 2025 10:46:28 +0200
Subject: [PATCH] feat: use the infer crate to determine if pdf embeds should
 be compressed

---
 Cargo.lock                    | 27 +++++++++++++++++++++++++++
 Cargo.toml                    |  1 +
 crates/typst-pdf/Cargo.toml   |  1 +
 crates/typst-pdf/src/embed.rs | 21 ++++++++++++++++++++-
 4 files changed, 49 insertions(+), 1 deletion(-)

diff --git a/Cargo.lock b/Cargo.lock
index ab2d2cc83..731ee422b 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -255,6 +255,17 @@ dependencies = [
  "shlex",
 ]
 
+[[package]]
+name = "cfb"
+version = "0.7.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d38f2da7a0a2c4ccf0065be06397cc26a81f4e528be095826eee9d4adbb8c60f"
+dependencies = [
+ "byteorder",
+ "fnv",
+ "uuid",
+]
+
 [[package]]
 name = "cfg-if"
 version = "1.0.0"
@@ -1259,6 +1270,15 @@ dependencies = [
  "serde",
 ]
 
+[[package]]
+name = "infer"
+version = "0.19.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a588916bfdfd92e71cacef98a63d9b1f0d74d6599980d11894290e7ddefffcf7"
+dependencies = [
+ "cfb",
+]
+
 [[package]]
 name = "inotify"
 version = "0.11.0"
@@ -3127,6 +3147,7 @@ dependencies = [
  "comemo",
  "ecow",
  "image",
+ "infer",
  "krilla",
  "krilla-svg",
  "serde",
@@ -3430,6 +3451,12 @@ version = "0.2.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "06abde3611657adf66d383f00b093d7faecc7fa57071cce2578660c9f1010821"
 
+[[package]]
+name = "uuid"
+version = "1.16.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "458f7a779bf54acc9f347480ac654f68407d3aab21269a6e3c9f922acd9e2da9"
+
 [[package]]
 name = "vcpkg"
 version = "0.2.15"
diff --git a/Cargo.toml b/Cargo.toml
index 12870b809..fb448b669 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -71,6 +71,7 @@ icu_segmenter = { version = "1.4", features = ["serde"] }
 if_chain = "1"
 image = { version = "0.25.5", default-features = false, features = ["png", "jpeg", "gif"] }
 indexmap = { version = "2", features = ["serde"] }
+infer = "0.19.0"
 kamadak-exif = "0.6"
 krilla = { version = "0.4.0", default-features = false, features = ["raster-images", "comemo", "rayon"] }
 krilla-svg = "0.1.0"
diff --git a/crates/typst-pdf/Cargo.toml b/crates/typst-pdf/Cargo.toml
index f6f08b5bc..5745d0530 100644
--- a/crates/typst-pdf/Cargo.toml
+++ b/crates/typst-pdf/Cargo.toml
@@ -23,6 +23,7 @@ bytemuck = { workspace = true }
 comemo = { workspace = true }
 ecow = { workspace = true }
 image = { workspace = true }
+infer = { workspace = true }
 krilla = { workspace = true }
 krilla-svg = { workspace = true }
 serde = { workspace = true }
diff --git a/crates/typst-pdf/src/embed.rs b/crates/typst-pdf/src/embed.rs
index 6ed65a2b6..b8d9e16b7 100644
--- a/crates/typst-pdf/src/embed.rs
+++ b/crates/typst-pdf/src/embed.rs
@@ -34,6 +34,7 @@ pub(crate) fn embed_files(
             },
         };
         let data: Arc<dyn AsRef<[u8]> + Send + Sync> = Arc::new(embed.data.clone());
+        let compress = should_compress(&embed.data);
 
         let file = EmbeddedFile {
             path,
@@ -41,7 +42,7 @@ pub(crate) fn embed_files(
             description,
             association_kind,
             data: data.into(),
-            compress: true,
+            compress,
             location: Some(span.into_raw().get()),
         };
 
@@ -52,3 +53,21 @@ pub(crate) fn embed_files(
 
     Ok(())
 }
+
+fn should_compress(data: &[u8]) -> bool {
+    let Some(ty) = infer::get(data) else {
+        return false;
+    };
+    match ty.matcher_type() {
+        infer::MatcherType::App => true,
+        infer::MatcherType::Archive => false,
+        infer::MatcherType::Audio => false,
+        infer::MatcherType::Book => true,
+        infer::MatcherType::Doc => true,
+        infer::MatcherType::Font => true,
+        infer::MatcherType::Image => false,
+        infer::MatcherType::Text => true,
+        infer::MatcherType::Video => false,
+        infer::MatcherType::Custom => true,
+    }
+}