Reader

opentimstdf::Reader is the entry point. Opening a bundle parses the GlobalMetadata table, validates the schema version, and prepares SQLite statements for repeated frame access.

Opening a bundle

use opentimstdf::Reader;
let reader = Reader::open("my_bundle.d")?;

Reader::open takes any path that resolves to a .d/ directory and contains both analysis.tdf (SQLite) and analysis.tdf_bin (block stream).

Methods

Method	Returns	Notes
open(path)	Result<Reader>	Open and validate a bundle.
bundle_dir()	&Path	The path the reader was opened on.
metadata()	Result<Metadata>	All GlobalMetadata rows decoded into a struct.
compression_type()	u32	1 (LZF) or 2 (zstd).
calibration()	Result<Calibration>	TOF/m/z and scan/(1/K0) calibrators for the most-referenced MzCalibration row (see Calibration).
bundle_calibration()	Result<BundleCalibration>	Calibration per MzCalibration row; for_frame(&frame) picks a frame's row.
mz_calibration_model()	Result<MzCalibrationModel>	Tables or RangeFallback.
mz_calibration_status()	Result<MzCalibrationStatus>	Model plus the reason when the range fallback is used.
range_fallback_calibration()	Result<Calibration>	The range-fallback model, built even when the tables are usable (for comparison).
frame(id)	Result<Frame>	A single frame's index row.
frames()	Result<Vec<Frame>>	All frames in ascending id order.
decode_peaks(&frame)	Result<Vec<Peak>>	Decode a frame's payload to (scan, tof, intensity) peaks. See Peaks and codecs.
dia_windows_for_frame(id)	Result<Option<DiaFrameWindows>>	diaPASEF isolation windows; None for non-DIA bundles.
pasef_msms_info_for_frame(id)	Result<Vec<PasefMsMsInfo>>	DDA PASEF per-frame info.
prm_msms_info_for_frame(id)	Result<Vec<PrmMsMsInfo>>	prm-PASEF per-frame entries.
prm_target(target_id)	Result<Option<PrmTarget>>	prm-PASEF target metadata.
precursor(precursor_id)	Result<Option<Precursor>>	A row from the Precursors table.

See the full type signatures on docs.rs.

Types and errors

The structs these methods return (Frame, Peak, Metadata, DiaWindow, PasefMsMsInfo, Precursor, ...) are #[non_exhaustive]. Read their fields directly; new fields can be added in a minor release, so code outside the crate cannot build them with struct literals (Peak has Peak::new).

opentimstdf::Error is also #[non_exhaustive], so a match on it needs a _ arm. A frame whose payload cannot be decoded returns Error::CorruptFrame, and a missing or out-of-range GlobalMetadata value needed by the calibration returns Error::InvalidMetadata. SQLite failures return Error::Sqlite, whose Display gives the SQLite message.

mzML export

let mut out = std::io::BufWriter::new(std::fs::File::create("run.mzML")?);
let summary = opentimstdf::mzml::write_indexed_mzml("my_bundle.d", &mut out)?;
println!("{} spectra", summary.spectra_written);

write_mzml and write_indexed_mzml stop at the first frame or chromatogram that cannot be decoded and return its error. To write everything else and count what was left out, build a TdfSource and use the openmassspec-core writer with RecordErrorPolicy::Skip:

use openmassspec_core::{RecordErrorPolicy, WriteOptions};

let mut src = opentimstdf::mzml::TdfSource::open("my_bundle.d")?;
let opts = WriteOptions::new().with_record_error_policy(RecordErrorPolicy::Skip);
let summary = openmassspec_core::write_mzml_with_options(&mut src, &mut out, &opts)?;
if summary.skipped() > 0 {
    eprintln!("{} records could not be decoded", summary.skipped());
}

TdfSource::iter_spectra yields one Err item per frame that fails and then continues with the next frame.