Quickstart: vendor2mzml

vendor2mzml auto-detects the vendor format of its input and writes mzML (or a gzipped mzML, or a metadata summary).

Convert

# Thermo .raw
vendor2mzml convert sample.raw sample.mzML

# Bruker .d bundle directory
vendor2mzml convert sample.d sample.mzML --indexed

# Waters .raw bundle directory, gzipped output
vendor2mzml convert sample.raw sample.mzML.gz

--indexed writes the standards-compliant <indexList> plus SHA-1 hash, so downstream tools (mzML2HDF, ProteoWizard's msconvert, MzIdentML builders) can index-jump into the file.

Disk space

<spectrumList count> must hold the number of spectra actually written, which is known only after the last one. The writer therefore holds the encoded spectra until the run is read to the end: in memory up to 64 MiB, then in a temporary file in the output file's directory. The file is removed when the conversion ends, successful or not. While a large conversion runs, that directory needs room for the output plus the uncompressed spectra (for .mzML.gz output the spool is not gzipped).

Unreadable spectra

A spectrum or chromatogram the vendor reader cannot decode is left out of the output by default. convert then prints a warning on stderr and still exits 0, because the file was written:

warning: skipped 12 unreadable record(s) (12 spectra, 0 chromatograms); rerun with --strict to stop at the first one and see its error

--strict stops at the first unreadable record, prints its error, and exits 1. The partial output file is kept, and stderr says it is incomplete:

error: thermo reader: scan data record ending at offset 0x5a7dcb does not match a known layout
  sample.mzML is incomplete

vendor2mzml convert --strict sample.raw sample.mzML

info and validate always stop at the first unreadable spectrum and exit 1.

Centroid on the way out

vendor2mzml convert sample.raw sample.mzML --centroid
vendor2mzml convert sample.raw sample.mzML --centroid --centroid-min-intensity 500

--centroid centroids every profile-mode spectrum before writing (local-maxima peak picking; spectra already tagged centroid pass through unchanged). It is off by default - a plain convert never discards profile data. --centroid-min-intensity discards picked peaks below that height and is ignored unless --centroid is set. info takes the same two flags, for a peak-count summary post-centroiding without writing mzML.

Validate

vendor2mzml validate sample.raw
vendor2mzml validate sample.mzML
vendor2mzml validate --json sample.d

validate accepts any input convert accepts (Thermo .raw, Bruker .d, Waters .raw) plus pre-existing mzML files (.mzML, .mzML.gz, read via mzdata). It runs the conformance harness from openmassspec-core:

• monotonic spectrum indices,
• non-negative, non-decreasing retention times,
• equal-length m/z and intensity arrays,
• equal-length mobility arrays (if present),
• MS-level / polarity sanity,
• precursor presence on MSn spectra.

--json emits a single JSON object suitable for jq:

{"ok":true,"input":"sample.mzML","kind":"mzML","spectrum_count":3047,"elapsed_sec":0.214}

On failure, ok is false and error_kind names the conformance variant (for example, PeakArrayLengthMismatch).

Profile a run

vendor2mzml convert sample.raw sample.mzML --profile json

emits one JSON object on stderr after the conversion completes:

{"output":"sample.mzML","indexed":false,"spectra":3047,"spectra_skipped":0,"chromatograms":2,"chromatograms_skipped":0,"elapsed_sec":0.812}

spectra and chromatograms count what was written; the _skipped fields count unreadable records left out.

Use --profile text for a human-readable variant.

Inspect without converting

vendor2mzml info sample.d

vendor:        bruker
path:          sample.d
instrument:    timsTOF Pro (MS:1003005)
source format: Bruker TDF format (MS:1002817)
native id:     Bruker TDF nativeID format (MS:1002818)
software:      opentimstdf 1.0.4
spectra:       24425
  ms1:         1132
  ms2:         23293
  pos:         24425
rt range:      240.291 - 900.132 s (4.00 - 15.00 min)
decode time:   2.05 s

--json switches the output to a single JSON object suitable for piping into jq.

Exit codes

Code	Meaning
0	Success. For convert, also when unreadable records were skipped (see the stderr warning).
1	I/O error, unsupported format, an unreadable spectrum in info/validate, or convert --strict hit one.
2	validate: input format was not recognised (clap errors also).
3	validate: the conformance harness rejected the input.