Biotite

Changelog

Additions
- Added `sequence.ProteinSequence.get_molecular_weight()` method
- Added `application.sra` subpackage as interface to *NCBI SRA tools*
- `FastqDumpApp` is used for fetching FASTQ files from the *NCBI SRA*
- Added `iter_read()` static method to `sequence.io.fasta.FastaFile`
and `sequence.io.fasta.FastqFile`
- This method is used to parse header-sequence-pairs from FASTA/FASTQ
files without the necessity to keep the entire file in memory.
- `set_sequence` and `set_sequences` in `sequence.io.fasta`
and `sequence.io.fasta` support writing RNA sequences with the new
`as_rna` parameter

Fixes
- Fixed missing whitespace at the end of `_loop` category labels in
*PDBx/mmCIF* files (224)
- Fixed inconsistent handling of model IDs over different file formats
for structures where the first model ID is greater than `1` (227)
- Removed warning in `structure.density()`
- `sequence.io.fastq.get_sequence()` and `sequence.io.fastq.get_sequences()`
properly handle RNA and ambiguous sequences now
- Fixed `start` parameter in `structure.renumber_atom_ids` and
`structure.renumber_res_ids`
- Updated fetch URL for FASTA files in `database.rcsb.fetch()`

Changelog

Additions
- Improved example gallery
- Added minigalleries in the API reference to get tangible examples for the
respective function/class
- Added support for animated *Matplotlib* plots
- Using [Ammolite](https://ammolite.biotite-python.org/) for rendering
PyMOL images
- Added support for new RCSB search API
- New `database.rcsb.Query` classes, that reflect the entirety of the new
search API, including sequence, sequence motif and structure searches
- Multiple `database.rcsb.Query` objects can be combined/negated using the
operators `|`, `&` and `~`
- Added the `return_type`, `sort_by` and `range` parameter to
`database.rcsb.search()`
- Added `database.rcsb.count()` function to count the number of results a
`database.rcsb.Query` would yield in a less costly way than
`database.rcsb.search()`
- Increased indexing speed in `biotite.structure.BondList`
- Added attribute `sequence.Sequence.alphabet` property, that is equivalent to
`sequence.Sequence.get_alphabet()`
- Added convenience functions `fastq.get_sequence()`, `fastq.get_sequences()`,
`fastq.set_sequence()` and `fastq.set_sequences()`
- Drastically increased writing speed of `sequence.io.fasta.FastaFile`
- Increased mapping speed of `sequence.AlphabetMapper`
- Added `sequence.Alphabet.is_letter_alphabet()` method
- Added general sequence I/O convenience functions
`sequence.io.load_sequence()`, `sequence.io.load_sequences()`,
`sequence.io.save_sequence()` and `sequence.io.save_sequences()` that derive
the appropriate `File` class from the suffix of the file name.


Changes
- The `omit_chain` parameter has been removed from `database.rcsb.search()`
- The old `database.rcsb.Query` classes have been removed
- Removed `python setup.py test` and `python setup.py build_sphinx` commands,
please use `pytest` and `sphinx-build` directly instead
- Renamed `sequence.NucleotideSequence.alphabet` to
`sequence.NucleotideSequence.alphabet_unamb`
- `sequence.io.fastq.FastqFile` returns its entries only as `str` instead of
`sequence.NucleotideSequence` for consistency with
`sequence.io.fastq.FastaFile`
- The method `sequence.io.fastq.FastqFile.get_sequence()` is deprecated
- The method `sequence.io.fastq.FastqFile.get_seq_string()` returns the
sequence as a `str` instead of a `sequence.NucleotideSequence`


Fixes
- Fixed `expect_looped` parameter in
`structure.io.pdbx.PDBxFile.get_category()`
- Fixed error in `structure.io.pdbx.PDBxFile`, that was raised, if a *PDBx*
field and its single-line value are in separate lines
- Added check for boolean mask length, when a boolean mask is given as index
to `biotite.structure.BondList`
- Changed `chain_id` *dtype* from `'U3'` to `'U4'` (215)

Changelog

Additions
- Added `structure.filter_nucleotides()`
- `structure.io.pdbx.get_sequence()` is able to parse a
`sequence.NucleotideSequence` from a PDBx file in addition to
`sequence.ProteinSequence`
- Added `structure.base_pairs()` for determining base pairs in nucleic acid
structures
- Added `structure.get_residue_starts_for()`
- Added `structure.check_atom_id_continuity()`
- Added `structure.renumber_atom_ids()` and `structure.renumber_res_ids()`
to fix structures with discontinuous atom/residue IDs
- Added `get_model_count()` to `structure.io.pdb`, `structure.io.pdbx`,
`structure.io.mmtf` and `structure.io.gro` to obtain the total number
of models
- The `model` parameter in `get_structure()` in `structure.io.pdb`,
`structure.io.pdbx`, `structure.io.mmtf` and `structure.io.gro` supports
negative values to start indexing beginning from the last model
- Increased performance of residue and chain-related functions
(e.g. `structure.get_residue.starts()`)


Changes
- Revamped *altloc* ID handling (194)
- Instead of choosing each alternate location individually there are three
options:
- `'first'` choses always chooses the atoms with the first *altloc* ID
for each residue
- `'occupancy'` choses always chooses the atoms with the highest occupancy
for each residue
- `'all'` does not filter any *altloc* IDs and adds the `altloc_id`
annotation to the resulting `structure.AtomArray` or
`structure.AtomArrayStack`
- Renamed `structure.check_id_continuity()` into
`structure.check_res_id_continuity()`; `structure.check_id_continuity()`
is still available, but is deprecated


Fixes
- Fixed `structure.BondList` being iterable, yielding nonsense values
- Improved element guesses in `structure.io.pdb.PDBFile` when the
element column is missing (188)
- Fixed parsing of single models from `structure.io.mmtf.MMTFFile` (205)
- Fixed missing unit cell values in `structure.io.pdbx.get_structure()`
raising an error; the `box` attribute is set to `None` instead

Changelog

Additions
- More functionality for `structure.BondList`
- `__contains__()` method to test whether a bond exists
- `find_connected()` identifies systems of connected atoms (*aka* molecules)
- Added frame wise iteration of trajectory files for saving memory
- `structure.io.TrajectoryFile.read_iter()` yields coordinates, box and time for each frame
- `structure.io.TrajectoryFile.read_iter_structure()` yields an `structure.AtomArray` for each frame
- Added ability to read entire biological assemblies from mmCIF files
- `structure.io.pdbx.list_assemblies()` lists the available assemblies
- `structure.io.pdbx.get_assembly()` returns the given assembly as
`structure.AtomArray` or `structure.AtomArrayStack`
- Added the `expect_looped` parameter to
`structure.io.pdbx.PDBxFile.get_category`
- `structure.info.vdw_radius_single()` provides VdW radii also for more
uncommon elements
- Added `structure.get_residue_masks()`, which masks all residues to which the
given atoms belong
- Added `structure.repeat()` functions to repeat atoms multiple times in the
same model with different coordinates
- Added a bunch of new examples to the gallery


Changes
- `temp_file()` and `temp_dir()` is deprecated, use the Python standard library
module `tempfile` instead
- For all `File` classes, `read()` is now a class method,
e.g. ``pdbx_file = PDBxFile.read()``, the old instance method is deprecated
- `database.rcsb.fetch()` and `database.enrez.fetch()` overwrite an existing
file if it is empty


Fixes
- A newline character is appended to the end of file, when writing text files
- Fixed `structure.CellList` when using the `periodic` parameter in combination
with the `selection` parameter; before unallocated memory was potentially
accessed

Changelog

Fixes
- Fixed support for `msgpack` 1.0

Changelog

Additions
- Added `structure.from_template()` to create a `structure.AtomArrayStack` from an existing atom array (or stack) and coordinates
- Added `ignore` parameter to `sequence.io.genbank.get_annotation()` to ignore the given feature keys
- Added `sequence.graphics.plot_plasmid_map()` for visualizing `sequence.Annotation` objects as plasmid
- Added a bunch of new examples to the gallery
- Added support for Python 3.8 on Windows


Changes
- The output of the `score_matrix()` method of `sequence.align.SubstitutionMatrix` is not writable anymore, rendering a `SubstitutionMatrix` truly immutable
- Renamed `environment.yaml` to `environment.yml`
- A `sequence.Feature` must have at least one location


Fixes
- Fixed incorrect centroid calculation in `structure.superimpose()`, when providing a boolean mask
- Fixed installation of PyPI source distributions
- Fixed issues when reading text files with `\r\n` line breaks (line breaks with carriage return, typical for Windows)