Psm-utils

Added

- ✨ `io.alphadia`: Read support for AlphaDIA `precursors.tsv` (103 by rodvrees)
- ✨ `io.fragpipe`: Read support for FragPipe `psm.tsv` (103 by rodvrees)
- ✨ `io.diann`: Read support for DIA-NN TSV (103 by rodvrees)

Changed

- 💥 `psm_list`: When returning a PSM property across the full PSMList (e.g. `psm_list["peptidoform"]`), `np.fromiter` is now used instead of `np.array`. This fixes an issue where if all peptidoforms have the same length, a 3D array of parsed sequences (amino acids and modifications) was be returned instead of an array of `Peptidoform` object. However, this does mean that all resulting arrays will have the `object` dtype instead of the previously coerced dtypes. This might lead to issues downstream. (102)
- ➖ `io.idxml`: Make pyOpenMS an optional dependency, working around https://github.com/OpenMS/OpenMS/issues/7600 for now. For `idxml` support, install psm_utils with the `idxml` extra dependencies. (#107 by paretje)

Fixed

- 🐛 `io.pepxml`: Fix modification location and mass parsing. Position had an off-by-one error and the reported mass was the sum of the residue and modification instead of the modification alone. (fixes 100, 104)

Fixed

- `io`: Fix Sage filename pattern for automatic file type inference
- `io.flashlfq`: Fix writing PSMs without protein accession
- `io.flashlfq`: Fix column names `Peptide Monoisotopic Mass` and `Protein Accession`.
- `io.idxml`: Fix parsing if spectra file name not present [92](https://github.com/compomics/psm_utils/issues/92)

Added

- `Peptidoform`: Add `modified_sequence` property to return the modified sequence in ProForma format, but without charge state.
- `io`: Add support for reading and writing FlashLFQ generic TSV files.

Fixed

- `io.percolator`: Fix and improve ScanNr inferring and writing
- `io.percolator`: Infer style from file extension if not provided (enables dynamic style determination in, for instance, `convert` function).

Added

- Peptidoform: Allow comparison between a peptidoform and a peptidoform string; allow direct indexing with square brackets, which indexes or slices parsed_sequence (in 89)

Fixed

- TSV: Avoid flooding logs when reading a different file format by raising exception when three consecutive rows could not be parsed (in 88)

Fixed

- `io.xtandem`: Fix parsing PSMs and complete protein names in XTandem (by julianu in 83)
- `io.tsv`: Fix warning formatting when parsing TSV (by paretje in 85)
- `io`: Fix support for mzIdentML and pepXML files from Comet (by paretje in 87)