extract: Support Matroska input

Cargo.toml

@@ -9,7 +9,7 @@ license = "MIT"
 [dependencies]
 bitvec_helpers = { version = "3.1.6", default-features = false, features = ["bitstream-io"] }
 hdr10plus = { path = "./hdr10plus", features = ["hevc", "json"] }
-hevc_parser = { version = "0.6.4", features = ["hevc_io"] }
+hevc_parser = { version = "0.6.6", features = ["hevc_io"] }
 
 clap = { version = "4.5.29", features = ["derive", "wrap_help", "deprecated"] }
 serde_json = { version = "1.0.138", features = ["preserve_order"] }

README.md

@@ -21,11 +21,16 @@ Options that apply to the commands:
 
 ## Commands
 * ### **extract**
-    Extracts the HDR10+ metadata from HEVC SEI messages to a JSON file.  
+    Extracts the HDR10+ metadata from a HEVC file to a JSON file.  
     Also calculates the scene information for compatibility with Samsung tools.  
 
     If no output is specified, the file is only parsed partially to verify presence of metadata.
 
+    Input file:
+    - HEVC bitstream
+    - Matroska (experimental): MKV file containing a HEVC video track.
+
+
     **Flags**:
     * `--skip-reorder` Skip metadata reordering after extracting.
         - [Explanation on when to use `--skip-reorder`](README.md#wrong-metadata-order-workaround).
@@ -34,6 +39,9 @@ Options that apply to the commands:
     **Examples**:
     ```console
     hdr10plus_tool extract video.hevc -o metadata.json
+
+    # Experimental
+    hdr10plus_tool extract video.mkv -o metadata.json
     ```
     ```console
     ffmpeg -i input.mkv -map 0:v:0 -c copy -bsf:v hevc_mp4toannexb -f hevc - | hdr10plus_tool extract -o metadata.json -

src/commands/extract.rs

@@ -1,4 +1,4 @@
-use anyhow::{bail, Result};
+use anyhow::Result;
 use hevc_parser::io::IoFormat;
 
 use super::{input_from_either, CliOptions, ExtractArgs};
@@ -21,7 +21,7 @@ impl Extractor {
         let format = hevc_parser::io::format_from_path(&input)?;
 
         if format == IoFormat::Matroska {
-            bail!("Extractor: Matroska format unsupported");
+            println!("Extractor: Matroska input is experimental!");
         }
 
         if !options.verify && output.is_none() {

src/core/parser.rs

@@ -1,5 +1,5 @@
 use std::fs::File;
-use std::io::{stdout, BufRead, BufReader, BufWriter, Write};
+use std::io::{stdout, BufWriter, Write};
 use std::path::PathBuf;
 
 use anyhow::{bail, ensure, Result};
@@ -78,15 +78,13 @@ impl Parser {
         };
         let mut processor = HevcProcessor::new(format.clone(), processor_opts, chunk_size);
 
-        let stdin = std::io::stdin();
-        let mut reader = Box::new(stdin.lock()) as Box<dyn BufRead>;
-
-        if let IoFormat::Raw = format {
-            let file = File::open(&self.input)?;
-            reader = Box::new(BufReader::with_capacity(100_000, file));
-        }
+        let file_path = if let IoFormat::RawStdin = format {
+            None
+        } else {
+            Some(self.input.clone())
+        };
 
-        processor.process_io(&mut reader, self)
+        processor.process_file(self, file_path)
     }
 
     pub fn add_hdr10plus_sei(&mut self, nals: &[NALUnit], chunk: &[u8]) -> Result<()> {

tests/hevc/extract.rs

@@ -1989,3 +1989,35 @@ fn dts_injected() -> Result<()> {
 
     Ok(())
 }
+
+#[test]
+fn regular() -> Result<()> {
+    let input_file = Path::new("assets/hevc_tests/regular.hevc");
+    let temp = assert_fs::TempDir::new()?;
+
+    let output_json = temp.child("metadata.json");
+    let expected_json = Path::new("assets/hevc_tests/regular_metadata.json");
+
+    assert_cmd_output(input_file, output_json.as_ref(), true)?;
+    output_json
+        .assert(predicate::path::is_file())
+        .assert(predicate::path::eq_file(expected_json));
+
+    Ok(())
+}
+
+#[test]
+fn regular_mkv() -> Result<()> {
+    let input_file = Path::new("assets/hevc_tests/regular.mkv");
+    let temp = assert_fs::TempDir::new()?;
+
+    let output_json = temp.child("metadata.json");
+    let expected_json = Path::new("assets/hevc_tests/regular_metadata.json");
+
+    assert_cmd_output(input_file, output_json.as_ref(), true)?;
+    output_json
+        .assert(predicate::path::is_file())
+        .assert(predicate::path::eq_file(expected_json));
+
+    Ok(())
+}