Built with zmz

Built with zmz

zmz exists because I kept writing the same mzML layer for every tool I built. These are the tools that now share it. Each one uses a different slice of the library, which is a decent tour of what it can do.

  

mzValidate

mzValidate validates mzML files: XML syntax, mzML structure, binary integrity, index metadata, vocabulary semantics, and references, mostly in one streaming pass and without building a tree. A validator has to see the document exactly as written, so it works below the mzML model:

• Events for every element, attribute, and text run with its source byte offset, which is where each finding's location comes from;
• model.encodingTerm, so it classifies array encodings exactly as zmz's readers do;
• base64, zlib, and sha1 to check binary payloads and the indexedmzML checksum without decoding into a model.

mzarc

mzarc is a research codec that compresses mass-spectrometry spectra with domain-specific transforms. It needs the numbers fast and nothing else held in memory:

• StreamReader with an array selector, so it decodes only the arrays it archives;
• DecodedArrays, reused for the whole file, so a run of any size compresses in flat memory;
• the record helpers and model.Terms for the metadata each archived spectrum keeps.

Add yours

Built something on zmz? Open an issue with a link and a line about what it does, and I'll add it here.