#@deckyard/abif

Applied Biosystems capillary sequencers write their runs as ABIF files: Sanger traces (.ab1), fragment-analysis runs (.fsa) and, from the 3500 series, HID runs (.hid). This deck reads them: the directory of tagged elements, every element type, and the usual contents - raw and analysed traces per dye, the dyes, base calls with qualities and peak locations, and the sample and run metadata.

• Reading a file: opening, the directory, values and the typed readers.
• The run: traces, dyes, base calls and metadata.
• The three file types: .ab1, .fsa and .hid, what each holds and how it differs.
• The format: the layout on disk, the element types, the common tags, and what 3500 HID files encrypt.
• Cheatsheet: every export.

#Conventions

Snippets assume this import:

import "@deckyard/abif/" as abif;

def f as abif.File init abif.open("sample.ab1");
io.printf("%s: %d bases\n", abif.info($f).sample, len(abif.calls($f).bases));

Input that breaks the format is an error, never repaired: parse checks the header and every directory entry against the file before anything is read, and a string tag that is malformed or not UTF-8 is refused. Every failure is Error{kind: "abif"}, except what fs.readBytes throws in open:

try {
    def f as abif.File init abif.open($path);
} catch (e) {
    io.printf("%s\n", $e.message);
}

A File holds the whole file, so a 1 MB .hid is 1 MB in memory. The deck has no capabilities.