From 763ae22f4fa14b926cff219bc5080c6dd0a45845 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Hubert=20Figui=C3=A8re?= <hub@figuiere.net>
Date: Mon, 24 Apr 2023 23:52:47 -0400
Subject: [PATCH] Read the XMP packet

- and expose it like exif
---
 examples/decode.rs |  1 +
 src/decoder.rs     | 13 +++++++++++++
 src/parser.rs      | 45 +++++++++++++++++++++++++++++++--------------
 tests/lib.rs       | 14 ++++++++++++++
 4 files changed, 59 insertions(+), 14 deletions(-)

diff --git a/examples/decode.rs b/examples/decode.rs
index 13934fe0..eee4e27d 100644
--- a/examples/decode.rs
+++ b/examples/decode.rs
@@ -23,6 +23,7 @@ fn main() {
 
     eprintln!("{:?}", info);
     eprintln!("Exif: {}", decoder.exif_data().is_some());
+    eprintln!("XMP: {}", decoder.xmp_data().is_some());
     eprintln!("ICC: {}", decoder.icc_profile().is_some());
 
     let output_file = File::create(output_path).unwrap();
diff --git a/src/decoder.rs b/src/decoder.rs
index 795ad1e4..fcf1812f 100644
--- a/src/decoder.rs
+++ b/src/decoder.rs
@@ -118,6 +118,8 @@ pub struct Decoder<R> {
     icc_markers: Vec<IccChunk>,
 
     exif_data: Option<Vec<u8>>,
+    xmp_data: Option<Vec<u8>>,
+    psir_data: Option<Vec<u8>>,
 
     // Used for progressive JPEGs.
     coefficients: Vec<Vec<i16>>,
@@ -144,6 +146,8 @@ impl<R: Read> Decoder<R> {
             is_mjpeg: false,
             icc_markers: Vec::new(),
             exif_data: None,
+            xmp_data: None,
+            psir_data: None,
             coefficients: Vec::new(),
             coefficients_finished: [0; MAX_COMPONENTS],
             decoding_buffer_size_limit: usize::MAX,
@@ -197,6 +201,13 @@ impl<R: Read> Decoder<R> {
         self.exif_data.as_deref()
     }
 
+    /// Returns the raw XMP packet if there is any.
+    ///
+    /// The returned value will be `None` until a call to `decode` has returned `Ok`.
+    pub fn xmp_data(&self) -> Option<&[u8]> {
+        self.xmp_data.as_deref()
+    }
+
     /// Returns the embeded icc profile if the image contains one.
     pub fn icc_profile(&self) -> Option<Vec<u8>> {
         let mut marker_present: [Option<&IccChunk>; 256] = [None; 256];
@@ -542,6 +553,8 @@ impl<R: Read> Decoder<R> {
                             AppData::Avi1 => self.is_mjpeg = true,
                             AppData::Icc(icc) => self.icc_markers.push(icc),
                             AppData::Exif(data) => self.exif_data = Some(data),
+                            AppData::Xmp(data) => self.xmp_data = Some(data),
+                            AppData::Psir(data) => self.psir_data = Some(data),
                         }
                     }
                 }
diff --git a/src/parser.rs b/src/parser.rs
index 72ba00dc..03383e52 100644
--- a/src/parser.rs
+++ b/src/parser.rs
@@ -95,6 +95,8 @@ pub enum AppData {
     Avi1,
     Icc(IccChunk),
     Exif(Vec<u8>),
+    Xmp(Vec<u8>),
+    Psir(Vec<u8>),
 }
 
 // http://www.sno.phy.queensu.ca/~phil/exiftool/TagNames/JPEG.html#Adobe
@@ -621,21 +623,20 @@ pub fn parse_app<R: Read>(reader: &mut R, marker: Marker) -> Result<Option<AppDa
                 }
             }
         }
-        // Exif Data
         APP(1) => {
-            if length >= 6 {
-                let mut buffer = [0u8; 6];
-                reader.read_exact(&mut buffer)?;
-                bytes_read = buffer.len();
-
-                // https://web.archive.org/web/20190624045241if_/http://www.cipa.jp:80/std/documents/e/DC-008-Translation-2019-E.pdf
-                // 4.5.4 Basic Structure of JPEG Compressed Data
-                if buffer == *b"Exif\x00\x00" {
-                    let mut data = vec![0; length - bytes_read];
-                    reader.read_exact(&mut data)?;
-                    bytes_read += data.len();
-                    result = Some(AppData::Exif(data));
-                }
+            let mut buffer = vec![0u8; length];
+            reader.read_exact(&mut buffer)?;
+            bytes_read = buffer.len();
+
+            // https://web.archive.org/web/20190624045241if_/http://www.cipa.jp:80/std/documents/e/DC-008-Translation-2019-E.pdf
+            // 4.5.4 Basic Structure of JPEG Compressed Data
+            if length >= 6 && buffer[0..6] == *b"Exif\x00\x00" {
+                result = Some(AppData::Exif(buffer[6..].to_vec()));
+            }
+            // XMP packet
+            // https://github.com/adobe/XMP-Toolkit-SDK/blob/main/docs/XMPSpecificationPart3.pdf
+            else if length >= 29 && buffer[0..29] == *b"http://ns.adobe.com/xap/1.0/\0" {
+                result = Some(AppData::Xmp(buffer[29..].to_vec()));
             }
         }
         APP(2) => {
@@ -658,6 +659,22 @@ pub fn parse_app<R: Read>(reader: &mut R, marker: Marker) -> Result<Option<AppDa
                 }
             }
         }
+        APP(13) => {
+            if length >= 14 {
+                let mut buffer = [0u8; 14];
+                reader.read_exact(&mut buffer)?;
+                bytes_read = buffer.len();
+
+                // PSIR (Photoshop)
+                // https://github.com/adobe/XMP-Toolkit-SDK/blob/main/docs/XMPSpecificationPart3.pdf
+                if buffer[0..14] == *b"Photoshop 3.0\0" {
+                    let mut data = vec![0; length - bytes_read];
+                    reader.read_exact(&mut data)?;
+                    bytes_read += data.len();
+                    result = Some(AppData::Psir(data));
+                }
+            }
+        }
         APP(14) => {
             if length >= 12 {
                 let mut buffer = [0u8; 12];
diff --git a/tests/lib.rs b/tests/lib.rs
index 3ee186be..037da227 100644
--- a/tests/lib.rs
+++ b/tests/lib.rs
@@ -154,3 +154,17 @@ fn read_exif_data() {
     // exif data start as a TIFF header
     assert_eq!(&exif_data[0..8], b"\x49\x49\x2A\x00\x08\x00\x00\x00");
 }
+
+#[test]
+fn read_xmp_data() {
+    let path = Path::new("tests")
+        .join("reftest")
+        .join("images")
+        .join("ycck.jpg");
+
+    let mut decoder = jpeg::Decoder::new(File::open(&path).unwrap());
+    decoder.decode().unwrap();
+
+    let xmp_data = decoder.xmp_data().unwrap();
+    assert_eq!(&xmp_data[0..9], b"<?xpacket");
+}