go2com VS participle

MRI data is commonly saved in DICOM format, yes. But so are X-ray images, ultrasound and CT scans and a bunch of other imaging modalities. All of these may be single images or a series of images over time. Some (such as MRI and CT) also typically consist of multiple images for each point in time1, each at a slightly different position in the body to be able to visualize 3D data.

I know of tools like Kaitai that do essentially what I want, just curious about my options. I guess something like https://github.com/alecthomas/participle could work to parse the format, then I'd have to generate code based on the parsed data or something

This one was recommended in an older thread and looks like it's still actively developed: https://github.com/alecthomas/participle

I like participle: https://github.com/alecthomas/participle. It's pretty easy to get something working if it fits your use case.

Like the other commenters said, it's probably best to just split the string and go through it manually. If you need a real parser, check out: https://github.com/alecthomas/participle

I used https://github.com/alecthomas/participle to implement a custom DSL for some projects. I think there's also an SQL example in the repository. After the parsing is solved it's 'only' about doing something useful with the query but at least you've full control of all aspects, which features you'd like to support and so on.

I made a domain specific language using https://github.com/alecthomas/participle what I found interesting is the grammar defined in structure annotations. After the parser runs, you get a fully instantiated structure based upon the input. I then invoke structure methods to do the work described in the DSL.

Maybe instead of regex you could look into other options.

Shameless plug, but have you compared your Rust parser to https://github.com/alecthomas/participle?