media-type-version - extract the format version from a media type string

[Home | sourcehut | issues | PyPI | crates.io | Python API | ReadTheDocs]

Overview

The media-type-version library is designed to be used as the first step in parsing structured data, e.g. configuration files, serialized classes, etc. The caller extracts the media type string (e.g. a JSON "mediaType": "..." key) and passes it in for parsing. The caller then decides what to do with the extracted version information - is this version supported, what fields are expected to be there, should any extraneous fields produce errors, and so on.

The media type string is expected to be in a <prefix>.vX.Y<suffix> format, with a fixed prefix and suffix. The prefix will usually be a vendor-specific media type. The version part consists of two unsigned integer numbers. The suffix, if used, may correspond to the file format. A sample media type string identifying a TOML configuration file for a text-processing program could be vnd.ringlet.textproc.publync.config/publync.v0.2+toml

The library

The media-type-version library provides a single function, extract(), that parses a media type string, strips the specified prefix and suffix, and looks for a .vX.Y version string left. It then returns the (X, Y) version tuple.

Python example:

mtv_cfg: Final = media_type_version.Config(
    log=logging.Logger(...),
    prefix="vnd.acme/thing",
    suffix="+toml",
)

ver_major, ver_minor = media_type_version.extract(mtv_cfg, "vnd.acme/thing.v3.12+toml")

Rust example:

use media_type_version::{Config as MTVConfig, Error as MTVError, Version as MTVersion};

let cfg = MTVConfig::builder()
    .prefix("vnd.acme/thing")
    .suffix("+toml")
    .build()
    .map_err(MTVError::into_owned_error)?;
assert_eq!(
    media_type_version::extract(&cfg, "vnd.acme/thing.v3.12+toml").as_tuple(),
    (3, 12)
);

The mtv-extract tool

The media-type-version library also provides a command-line tool called mtv-extract that can be used to extract format versions from various sources.

The mtv-extract tool supports the following top-level command-line options:

• -q: quiet operation; only display warnings and error messages
• -v: verbose operation; display diagnostic output

The "features" subcommand

The features subcommand will display a single line of output starting with the prefix "Features: " and containing a space-separated list of "name=version" pairs. This output format is intended to be machine-readable, so that other programs may examine it using e.g. the feature-check library and command-line tool.

The "lines" subcommand

The lines subcommand will read a series of strings from the specified files, parse them as media-type strings with the specified prefix and suffix, and output a line consisting of two tab-separated numbers for each parsed string:

$ { echo vnd.acme/thing.v3.47; echo vnd.acme/thing.v42.616; } | mtv-extract -q lines -p vnd.acme/thing -- -
3       47
42      616
$

The lines subcommand supports the following command-line options:

• -p prefix (required): the prefix to strip from the media type string
• -s suffix: the optional suffix to strip from the media type string

Supported features

The features subcommand of the mtv-extract tool, as well as the FEATURES constant in both the Python and Rust implementations, may currently list the following features:

media-type-version

The version of the media-type-version library itself.

cmd-features

0.1

The command-line tool supports the features subcommand with the output format described above.

cmd-lines

0.1

The command-line tool supports the lines subcommand with the mandatory prefix option and the optional suffix one. It requires at least one file name to read from, and it supports - for reading from the standard input stream.

extract

0.1

The library supports the extract() function. It accepts a Config parameter containing the prefix and suffix strings, and a string parameter to parse. The string must contain a ".vX.Y" version specification between the prefix and the suffix.

Contact

The media-type-version library was written by Peter Pentchev. It is developed in a sourcehut repository. This documentation is hosted at Ringlet with a copy at ReadTheDocs.