Psm-utils

Added

- `Peptidoform`: Added support for `iter()` and `len()` methods
- `Peptidoform`: Added support for initialization from a `pyteomics.proforma.ProForma` object
- `PSM`: Add `precursor_mz_error` property
- `PSMList`: Added support for `append()` and `extend()` methods.
- `io`: Added new `io.ionbot` reader
- `io`: Added support for Proteome Discoverer MSF files
- `io.mzid`: Parse inverse reduced ion mobility from mzid files (e.g. from PEAKS)
- `io.mzid`: Add support for user to define custom score key
- `io.mzid`: Add `Proteome Discoverer Delta Score` to known scores (with spaces, no colons)
- `io.mzid`: Allow inconsistent presence of score in PSMs in a single mzid file

Changed

- `PSM`: Values of the `rescoring_features` dictionary are now coerced to floats
- io: Raise `PSMUtilsIOException` when passed filetype is not known
- `io`: Make io reader `read_file` method inheritable (code cleanup)
- `io.mzid`: Throw warning when no known score can be parsed from mzid file instead of error
- `io.mzid`: Move spectrum level parsing of rt and ion mobility to function
- `io.mzid`: Give `PeptideShaker PSM score` priority over other potential search engine scores (required for correct PeptideShaker mzid parsing)
- `io.percolator`: Add option to write `PSMScore` and `ChargeN` as features to PIN file. Default is now `False`.
- Formatting: Increase max line length to 99 (code formatting)

Fixed

- `PSMList`: Fix issue where `psm_list["protein_list"]` resulted in a Numpy error due to the inconsistent shape of the lists.
- `io.tsv`: Throw more descriptive `PSMUtilsIOException` when handeling tsv errors
- `io.msamanda`: Fix support for N/C-terminal modifications
- `io.Percolator.PercolatorTabWriter`: Allow rescoring features that are not in `feature_names` (`extrasaction` is now specified in `DictWriter`)
- Use raw strings for escape characters where needed
- Fix compatibility with sqlalchemy 2.0 (move of `declarative_base`)
- online: Remove useless == True
- docs: Set newer `build>os` configuration for readthedocs.org
- CI: Upgrade Github action versions

Fixed

- `PSMList`: Revert comparison operator change from v0.4.0 that results in broken `calculate_qvalues()` method (E711; Numpy array, not singleton)

Added

- Add + operator support for `PSMList`
- Add utility functions for m/z-mass conversion in new module `psm_utils.utils`
- `peptidoform`: Catch `ProFormaError` and reraise `PeptidoformException` with invalid peptidoform in message

Changed

- `io.msamanda`: Changed `REQUIRED_COLUMNS` to include new features from the MS Amanda output CSV file
- `io.peptide_record`Catch the `IndexError` when a modification has a position that is out of range for the peptide, and raise an `InvalidPeprecModificationError` instead.
- Rename optional dependency `doc` to `docs`
- Implement "raise from e" when applicable throughout package

Fixed

- Added missing `io.msamanda` API docs

Changed

- `io.sage`: Change `spectrum_fdr` to `spectrum_q` (crf. lazear/sage64).

Added

- Add reader for [Sage](https://github.com/lazear/sage) PSM files.
- `io.mzid`: Add reading/writing of PEP and q-values

Changed

- `psm`: The default values of `PSM.provenance_data`, `PSM.metadata` and `PSM.rescoring_features` are now `dict()` instead of `None`.
- `PSMList`: Also allow Numpy integers for indexing a single PSM
- `io.mzid.MzidReader`: Attempt to parse `retention time` or `scan start time` cvParams from both SpectrumIdentificationResult as SpectrumIdentificationItem levels. Note that according to the mzIdentML specification document (v1.1.1) neither cvParams are expected to be present at either level.
- `io.mzid.MzidReader`: Prefer `spectrum title` cvParam over `spectrumID` attribute for `PSM.spectrum_id` as these titles always match to the peak list files. In this case, `spectrumID` is saved in `metadata["mzid_spectrum_id"]`. Fall back to `spectrumID` if `spectrum title` is absent.
- `io.mzid.MzidWriter`: `PSM.retention_time` is now written as cvParam `retention time` instead of `scan start time`, and to the `SpectrumIdentificationItem` level instead of the `SpectrumIdentificationResult` level, as theoretically in psm_utils, multiple PSMs for the same spectrum can have different values for `retention_time`.
- `io.mzid.MzidWriter`: Write PSM score as cvParam `search engine specific score` instead of userParam `score`.
- `io.percolator.PercolatorTabWriter`: For PIN-style files: Use `SpecId` instead of `PSMId` and write `PSMScore` and `ChargeN` columns by default.
- Filter warnings from `psims.mzmlb` on import, as `mzmlb` is not used

Fixed

- `psm`: Fix missing qvalue and pep in docstring
- `peptidoform`: ProForma mass modifications are now correctly parsed within the `rename_modifications` function.
- `io.maxquant.MSMSReader`: Correctly parse empty `Proteins` column to `None`
- `io.percolator.PercolatorTabReader`: Correctly parse Percolator peptidoform notation if no leading or trailing amino acids are present (e.g. `.ACDK.` instead of `K.ACDK.E`).
- `io.percolator.PercolatorTabWriter`: ScanNr is now correctly written as an integer counting from the first PSM in the file.
- `io.percolator.PercolatorTabWriter`: If no protein information is present, write the peptidoform preceded by `PEP_` to the Proteins column.
- `io.idxml`: Read metadata as strings
- `io.mzid.MzidReader`: Set `PSM.retention_time` to `None` instead of `float('nan')` if missing from the PSM file.
- `io.mzid`: Fix reading of file if charge is missing
- `io.mzid`: Fix writing if protein_list is None
- `io.mzid`: Consider all `PeptideEvidence` entries for a `SpectrumIdentificationItem` to determine `is_decoy`
- `io.mzid`: Fix handling of mzIdentML files when `is_decoy` field is not present (fixes 30)
- `io.tsv`: Raise `PSMUtilsIOException` with clear error message when TSV `protein_list` cannot be read

Fixed

- Fix bug in `io._base_classes` (introduced in v0.2.2)
- Fix bug in TSVReader for reading TSV files with empty protein_list