Release 3.1.2

• Patch fix to no longer require schema_version to be populated in the uns of an h5ad when validating using cellxgene-schema validate.

Release 3.1.1

• Add 'migrate' command for migrating 3.0.0 schema_version datasets to comply with schema_version 3.1.0

Release 3.1.0

• Defined suspension_type requirements for new assays, and changed requirements for some previously supported assays (see full list of validation requirements here)
  • Added MARS-seq
  • Added BD Rhapsody Whole Transcriptome Analysis
  • Added BD Rhapsody Targeted mRNA
  • Added inDrop
  • Added STRT-seq
  • Added TruDrop
  • Added GEXSCOPE technology
  • Added SPLiT-seq
  • Changed spatial transcriptomics by high-throughput sequencing [EFO:0030005] and its children to spatial transcriptomics [EFO:0008994] and its children
  • Updated Seq-Well [EFO:0008919] to Seq-Well [EFO:0008919] and its children
• Add labels option on cellxgene-schema validate now writes current schema version in h5ad uns as 'schema_version'
• cellxgene-schema validate now validates ontology terms against a new set of ontologies (see full list here)
  • Updated CL to the 2023-06-22 release
  • Updated EFO to the 2023-07-17 EFO 3.56.0 release
  • Updated MONDO to the 2023-07-03 release
  • Updated NCBITaxon to the 2023-06-20 release
  • Updated PATO to the 2023-05-18 release
  • Updated UBERON to the 2023-06-28 release

Release 3.0.3

Changed

• Deprecated cellxgene-schema convert in favor of cellxgene-schema migrate.
• Add cellxgene-schema migrate to migrate a dataset from the previous schema version to the current schema version.

Release 3.0.2

• A "view" is used when adding labels to a validated h5ad to improve performance and reduce memory requirements.
• Chunked reads are used when validating an h5ad. This reduces the memory required to validate. Chunked reads are not
  used when the X or raw.X value or in CSC format.

Release 3.0.1

Changed
The convert command used for 2.0.0 -> 3.0.0. schema conversions now compresses the output file.

Release 3.0.0

Added

• All new schema requirements and restrictions outlined in schema version 3.0.0 are strictly enforced by validate command. For failures, an error message is displayed and validation fails.
• Tests that mirror the requirements in schema version 3.0.0.
• convert CLI command, for automating conversion of cellxgene schema 2.x.x compliant h5ads to 3.0.0 compliant h5ads, where possible.
  • suspension_type and donor_id may require manual curation to be 3.x.x compliant, please see linked schema definition for details.
• remove-labels CLI command, for locally removing data portal added labels from datasets.

Changed

• Updated pinned versions of ontology and feature references (see ./cellxgene_schema/ontology_files/).
• Updated various dependency versions (see requirements.txt)
• Require Python >= 3.8 for installation

Removed

• Support for validating cellxgene schema versions 2.x.x.

2.1.2

What's Changed

• Add support to generate metrics about testing coverage in schema CLI and upload to Codecov by @maniarathi in #204
• feat: use gene_id as gene_name if gene_name is incorrect id (#208) by @seve in #209
• feat: filter PAR_Y genes by @seve in #207
• fix: add feature name and reference to raw.var by @seve in #210
• Release 2.1.2 by @seve in #214

Full Changelog: 2.1.1...2.1.2

Release 2.1.1

[2.1.1] - 2022-01-27

Changed

• Gene reference files now have an extra column for gene length.
• Gene reference files no longer have transcripts.
• Improved CLI responsiveness in certain situations

Release 2.1.0

[2.1.0] - 2021-11-16

Added

• A check for convertibility to Seurat format. Prints warnings if conversion is not possible.