#webm #matroska #parser #serialization #encoding #source-file #media-file

webm-iterable

This crate extends the ebml-iterable library to provide an iterator over webm data. It provides a Matroska specification implementing the required traits to read webm files.

17 releases

new 0.6.3 Nov 12, 2024
0.6.2 Apr 4, 2024
0.6.1 Feb 27, 2024
0.6.0 Oct 20, 2023
0.3.0 Jul 15, 2021

#31 in Video

Download history 112/week @ 2024-07-29 62/week @ 2024-08-05 9/week @ 2024-08-26 37/week @ 2024-09-02 14/week @ 2024-09-09 6/week @ 2024-09-16 19/week @ 2024-09-23 176/week @ 2024-09-30 141/week @ 2024-10-07 282/week @ 2024-10-14 1796/week @ 2024-10-21 2073/week @ 2024-10-28 1431/week @ 2024-11-04 1884/week @ 2024-11-11

7,227 downloads per month
Used in 4 crates (via dash-mpd)

MIT license

74KB
902 lines

This crate was built to ease parsing files encoded in a Matroska container, such as WebMs or MKVs.

[dependencies]
webm-iterable = "0.6.3"

Usage

The WebmIterator type is an alias for ebml-iterable's TagIterator using MatroskaSpec as the generic type, and implements Rust's standard Iterator trait. This struct can be created with the new function on any source that implements the standard Read trait. The iterator outputs MatroskaSpec variants containing the tag data.

Note: The with_capacity method can be used to construct a WebmIterator with a specified default buffer size. This is only useful as a microoptimization to memory management if you know the maximum tag size of the file you're reading.

The data in the tag can then be modified as desired (encryption, compression, etc.) and reencoded using the WebmWriter type. WebmWriter simply wraps ebml-iterable's TagWriter. This struct can be created with the new function on any source that implements the standard Write trait.

See the ebml-iterable docs for more information on iterating over ebml data if needed.

Matroska-specific types

This crate provides three additional structs for special matroska data tags:

Block

pub struct Block {
    pub track: u64,
    pub timestamp: i16,

    pub invisible: bool,
    pub lacing: BlockLacing,
}

impl Block {
    pub fn read_frame_data(&self) -> Result<Vec<Frame>, WebmCoercionError>
    pub fn raw_frame_data(&self) -> &[u8]
}

These properties are specific to the Block element as defined by Matroska. The Block struct implements TryFrom<&MatroskaSpec> and Into<MatroskaSpec> to simplify coercion to and from regular variants.

SimpleBlock

pub struct SimpleBlock {
    pub track: u64,
    pub timestamp: i16,

    pub invisible: bool,
    pub lacing: Option<BlockLacing>,
    pub discardable: bool,
    pub keyframe: bool,
}

impl SimpleBlock {
    pub fn read_frame_data(&self) -> Result<Vec<Frame>, WebmCoercionError>
    pub fn raw_frame_data(&self) -> &[u8]
}

These properties are specific to the SimpleBlock element as defined by Matroska. The SimpleBlock struct also implements TryFrom<&MatroskaSpec> and Into<MatroskaSpec> to simplify coercion to and from regular variants.

Examples

This example reads a media file into memory and decodes it.

use std::fs::File;
use webm_iterable::WebmIterator;

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut src = File::open("media/test.webm").unwrap();
    let tag_iterator = WebmIterator::new(&mut src, &[]);

    for tag in tag_iterator {
        println!("[{:?}]", tag?);
    }

    Ok(())
}

This example does the same thing, but keeps track of the number of times each tag appears in the file.

use std::fs::File;
use std::collections::HashMap;
use webm_iterable::WebmIterator;
use webm_iterable::matroska_spec::EbmlTag;

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut src = File::open("media/test.webm").unwrap();
    let tag_iterator = WebmIterator::new(&mut src, &[]);
    let mut tag_counts = HashMap::new();

    for tag in tag_iterator {
        let count = tag_counts.entry(tag?.get_id()).or_insert(0);
        *count += 1;
    }
    
    println!("{:?}", tag_counts);
    Ok(())
}

This example grabs the audio from a webm and stores the result in a new file. The logic in this example is rather advanced - an explanation follows the code.

use std::fs::File;
use std::convert::TryInto;

use webm_iterable::{
    WebmIterator, 
    WebmWriter,
    matroska_spec::{MatroskaSpec, Master, Block, EbmlSpecification, EbmlTag},
};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    // 1
    let mut src = File::open("media/audiosample.webm").unwrap();
    let tag_iterator = WebmIterator::new(&mut src, &[MatroskaSpec::TrackEntry(Master::Start)]);
    let mut dest = File::create("media/audioout.webm").unwrap();
    let mut tag_writer = WebmWriter::new(&mut dest);
    let mut stripped_tracks = Vec::new();

    // 2
    for tag in tag_iterator {
        let tag = tag?;
        match tag {
            // 3
            MatroskaSpec::TrackEntry(master) => {
                let children = master.get_children();
                let is_audio_track = |tag: &MatroskaSpec| {
                    if let MatroskaSpec::TrackType(val) = tag {
                        return *val != 2;
                    } else {
                        false
                    }
                };

                if children.iter().any(is_audio_track) {
                    let track_number_variant = children.iter().find(|c| matches!(c, MatroskaSpec::TrackNumber(_))).expect("should have a k number child");
                    let track_number = track_number_variant.as_unsigned_int().expect("TrackNumber is an unsigned int variant");
                    stripped_tracks.push(*track_number);
                } else {
                    tag_writer.write(&MatroskaSpec::TrackEntry(Master::Full(children)))?;
                }
            },
            // 4
            MatroskaSpec::Block(ref data) => {
                let block: Block = data.try_into()?;
                if !stripped_tracks.iter().any(|t| *t == block.track) {
                    tag_writer.write(&tag)?;
                }
            },
            MatroskaSpec::SimpleBlock(ref data) => {
                let block: Block = data.try_into()?;
                if !stripped_tracks.iter().any(|t| *t == block.track) {
                    tag_writer.write(&tag)?;
                }
            },
            // 5
            _ => {
                tag_writer.write(&tag)?;
            }
        }
    }
    
    Ok(())
}

In the above example, we (1) build our iterator and writer based on local file paths and declare useful local variables, (2) iterate over the tags in the webm file, (3) identify any tracks that are not audio and store their numbers in the stripped_tracks variable; if they are audio, we write the "TrackEntry" out, (4) only write block data for tracks that are audio, and (5) write all other tags to the output destination.

Notes

  • Notice the second parameter passed into the WebmIterator::new() function. This parameter tells the decoder which "master" tags should be read as Master::Full variants rather than the standard Master::Start and Master::End variants. This greatly simplifies our iteration loop logic as we don't have to maintain an internal buffer for the "TrackEntry" tags that we are interested in processing.

State of this project

Parsing and writing complete files should both work. Streaming (using tags of unknown size) should now also supported as of version 0.4.0. If something is broken, please create an issue.

Any additional feature requests can also be submitted as an issue.

Author

Austin Blake

Dependencies

~2.5MB
~54K SLoC