0.4.0 (2015-09-09)¶

Changes since 0.3.7 (2015-06-23). See issues at milestone 0.4.0.

Special Attention¶

#227: x_to_r and r_to_x methods were renamed to x_to_n and n_to_x as part of support for non-coding transcripts.
#231: The UTA data provider will use a recently updated database by default (uta_20150827). Clients with custom configurations should use postgresql://anonymous:anonymous@uta.biocommons.org/uta/uta_20150827. (Note the change of hostname, username, and password as well; see Deprecations.)
#238: Most methods now raise HGVSDataNotAvailableError when expected data is not available. Previously, None was returned for some methods.
#244: Removed cache_transcripts argument from VariantMapper. This argument was deprecated in 0.3.0 and is now obsolete. Dataproviders are now expected to cache data.
#246: Remove hgvsX_to_hgvsY methods. These methods were deprecated in 0.3.0 and are now obsolete.
#247: Dup and Repeat “seq” instance variable renamed to “ref” for consistency.
EasyVariantMapper, Normalizer, and Validator now fetch sequence data at runtime, which may raise performance and privacy concerns. Users may wish to read Privacy Issues in the documentation.

UTA: UTA now uses anonymous:anonymous as the username:password. uta_public:uta_public will be obsolete shortly.

#248: Don’t raise validation exception when del sequence is empty [BROKEN: b6c07d329d36]

#44: Added variant normalization and use during mapping. Thanks to Meng Wang and Kevin Jacobs for contributions. (pull request #17)
#168: EasyVariantMapper supports replacing the reference sequence during mapping and enabled by default.
#227: Implement initial support for non-coding transcripts.
#230: Allow full IUPAC for NA and AA, with tests. Previously, the grammar admitted only ACGTU.
#233: Added get_similar_transcripts() to data UTA provider to expose UTA’s tx_similarity view.
#234, #241: Preferentially use transcript-protein accession associations from RefSeq when mapping c. to p. variants. Previously, when multiple protein accessions were associated with a single distinct sequence, the p. accession was arbitrary.
#236, #240: Added seqfetcher.SeqFetcher to fetch sequences from NCBI & Ensembl
#250: Implemented configuration module; hgvs.global_config is initialized once and available globally
#251: Parens now optional around p. edits; default is enabled per HGVS spec (hgvs.global_config.mapping.inferred_p_is_uncertain)
#255: variants normalized by EasyVariantMapper by default [BROKEN: 1b85d4deabc3]
#261: Replace reference default from config (hgvs.global_config.mapping.replace_reference)
UTA is now available as a docker image for local installation. See Local Installation of UTA (optional).

#213: Clarify warning message when validating intronic variants.
#254: Support inversion, conversion, and nadupn variants
Added misc/experimental/tx-seq-discrepancies to identify genomic locations of reference-transcript discrepancies
Added variant context method to evm (temporary location, but useful for debugging)
HGVSInternalError now subclasses HGVSError (not Exception) [BROKEN: ff6cd4dc51dc]
Lots of documentation updates.
Raise HGVSParserError (instead of ometa.runtime.ParseError) when parsing fails [BROKEN: efa93fe29d15]
The UTA data provider now checks for the requested schema on connection and provides more informative errors on failure.
hdp.data_version returns schema name for UTA since that that is the conventional use.
Use autocommit to prevent transaction overhead and locks [BROKEN: 65d69e41716e]

#263: Trying out a new tag-based changelog mechanism for 0.4.0 and 0.4 series.
All code will be mercilously reformmated with yapf occasionally using .style.yapf
Build and upload wheel packages (in addition to existing eggs and tarballs)
Docs significantly overhauled and moved to readthedocs.org with automatic webhook-based building
Enable users to set application_name when connecting [BROKEN: 835ac7771909]
_UTA_URL_KEY (dev use only) will switch URLs to any in hgvs/_data/defaults.ini
on import of hgvs, emit logging info line w/version [BROKEN: aa97f2c1cdc8]
sped up most tests by using setUpClass() rather than setUp() [BROKEN: a6d227f6a3e0]