Release version 1.5.0

New Features

• File aliases are now prefixed with the workflow ID to ensure alias uniqueness between call cached files used in different workflows

Bugfixes

• Corrected a previously incorrect analysis step run for TF ChIP-seq
• Fixed a bug that would cause the contents of small files uploaded from GCS to S3 to be duplicated

Release version 1.4.0

New Features

• cropped_read_length_tolerance is now posted for ChIP bams when crop_len is not 0
• mapped_run_type is now posted for all ChIP bams
• File uploads are now performed at the very end of the accessioning process
• Short options were added to the CLI for -m/--accession-metadata, -s/--server, and -p/--pipeline-type
• Most requests to the ENCODE portal are now cached, resulting in improved speed posting metadata objects
• The file upload mechanism was changed significantly. Files to upload are no longer downloaded to disk before upload to S3, and are instead streamed using an in-memory buffer. Furthermore, code can now reattempt file upload when the ENCODE portal indicates upload failed

Release version 1.3.1

Bugfixes

• Fixed a bug that would cause Analysis._search_down to fail when the task being searched had no output files

Release version 1.3.0

New Features

• Added a separate template for control ChIP-seq that does not post ChipAlignEnrichQualityMetrics
• Changed derived_from_inputs in templates to be a JSON boolean

Release version 1.2.2

Bugfixes

• Fixed a bug in accessioning bulk RNA QC that would cause attachments to be posted with a mime type the portal rejected as invalid

Release version 1.2.1

Bugfixes

Fixed a bug that could cause an error during generation of ChipAlignmentEnrichmentQualityMetrics when there were more than one set of parent fastqs of the file with the quality metric.

Release version 1.2.0

New Features

• Added Jsonnet-generated templates for accessioning full ChIP-seq experiments in one shot, use them by specifying --pipeline-type to be one of tf_chip, histone_chip, or mint_chip. Older peak call only templates are still present.
• cropped_read_length property now posted when accessioning bams. If the crop_length value read from the WDL metadata was 0, indicating no cropping was performed, then this value is not posted.

Release version 1.1.0

New Features

• Added a new template for accessioning peak-calling for MINT-ChIP experiments where multiple blacklists were used
• Accessioning code now posts an array of accessioned file IDs to the analyses property of the experiment on the portal

Bugfixes

• Fixed a bug introduced in 1.0.0 where reaccessioning would cause an error attempting to generate quality metrics

Release version 1.0.1

Bugfixes

• Fixed a bug in finding assembly for ChIP-seq introduced in versison 1.0.0.

Release version 1.0.0

Backwards Incompatible Changes

This release includes a number backwards incompatible changes due to refactoring of the internals. The command line usage remains the same, but the internal API has changed in a number of ways too numerous to list here.

New features

• Added a --dry-run command line flag that does not post anything and reports any potential file md5 conflicts
• Added bulk RNA accessioning code
• Added --no-log-file and --log-file-path command line options for more control over logging
• Added a mechanism to filter the derived_from to avoid diamond dependencies in workflow digraphs causing spurious derived_from relationships

Bugfixes

• Updated the histone ChIP-seq template to include preferred_default
• Updated encodeD to fix failing builds for integration tests