Reader

The entry point is RawFileReader. Opening a file parses the file header, audit tags, sample information, run header, scan event tree, scan index, error log, and generic data section. Scan peak data is read on demand from a separate std::fs::File handle.

use opentfraw::RawFileReader;

let raw = RawFileReader::open_path("sample.raw")?;

After open_path returns, the following fields are populated:

Field	Type	Description
`header`	`FileHeader`	Raw file magic/version header
`seq_row`	`SeqRow`	Sequence-table row (sample id, comment, vial, ...); surfaced to Python as `sample_info`
`raw_file_info`	`RawFileInfo`	Computer name, controller count, acquisition date, and other file-level metadata
`run_header`	`RunHeader`	Acquisition time range, first/last scan number
`scan_index`	`Vec<ScanIndexEntry>`	Per-scan byte offsets into the data stream
`scan_events`	`Vec<ScanEvent>`	Per-scan event metadata (filter, polarity, ...)
`scan_parameters_header`	`GenericDataHeader`	Column layout for `scan_parameters`
`scan_parameters`	`Vec<GenericRecord>`	Per-scan generic trailer values (`scan_parameters` in Python)
`error_log`	`Vec<ErrorEntry>`	Instrument error/status messages; surfaced to Python as `error_log()`
`inst_log_header`	`GenericDataHeader`	Column layout for `inst_log`
`inst_log`	`Vec<GenericRecord>`	Instrument status-log records; surfaced to Python as `status_log()`
`version`	`u32`	Raw file format version
`num_scans`	`u32`	Scan count from the run header
`data_addr`	`u64`	Data stream base address (for computing absolute scan offsets)
`flat_peaks`	`bool`	True if scan data uses flat-peak format (TSQ/SRM) instead of `PacketHeader`
`scan_format`	`ScanDataFormat`	Which decoder will be used for peaks
`device_family`	`DeviceFamily`	Heuristic instrument classification
`instrument_model`	`Option<&'static str>`	Reported instrument model from file metadata, if detected
`srm_q1_by_event`	`HashMap<u16, f64>`	SRM only: scan_event -> Q1 precursor m/z
`srm_q3_windows`	`HashMap<u16, Vec<(f32, f32)>>`	SRM only: scan_event -> Q3 isolation windows
`srm_ce_by_event`	`HashMap<u16, f64>`	SRM only: scan_event -> collision energy (eV); empty for v66/TSQ Altis files

For multi-controller files (MS + UV/PDA/Analog), controllers enumerates each controller's metadata as Vec<ControllerInfo> (index, run-header address, is_ms_controller, ControllerType, first/last scan, start/end time):

let mut file = std::fs::File::open("sample.raw")?;
let controllers = raw.controllers(&mut file)?;

Peaks are read with read_scan_peaks:

let mut file = std::fs::File::open("sample.raw")?;
let peaks = raw.read_scan_peaks(&mut file, 1)?;
for (mz, intensity) in peaks.mz.iter().zip(peaks.intensity.iter()) {
    println!("{mz:.4}\t{intensity:.0}");
}

read_scan_peaks dispatches on scan_format to one of the three decoders described in Scan data. Related read methods:

Method	Returns	Notes
`read_scan`	`ScanDataPacket`	Full packet (profile + centroids + FT labels); `PacketHeader` format only
`read_scan_labels`	`ScanDataPacket`	Centroids + FT labels, skipping the profile signal; `PacketHeader` format only
`read_scan_peaks`	`Vec<Peak>`	Centroided peaks, dispatching on `scan_format`; the recommended default
`read_peaks_only`	`Vec<Peak>`	Like `read_scan_peaks` but skips profile data on `PacketHeader` files for a 2-10x speedup
`read_scan_flat`	`Vec<Peak>`	Low-level: TSQ/SRM (v63) flat-peak format
`read_scan_srm_v66`	`Vec<Peak>`	Low-level: TSQ Quantiva/Altis (v66) SRM format

Error handling

Public functions return opentfraw::Result<T>. The error type is opentfraw::Error, which wraps the failure category (Io, Parse, UnsupportedVersion, BadMagic, ...) and a message.

Error handling​

Error handling