Releases: EBIBioSamples/biosamples-v4
v4.0.3
4.0.3
This is a bugfix release that addresses the following issues:
- Forward legacy group URLs /biosamples/groups/SAMEGxxxx to /biosamples/samples/SAMEGxxxxx
- Missing or malformed update and release date on legacy XML group submission will default to current datetime. It is not recommended that users intentionally rely on this.
- Index legacy XML group submissions, which was not happening due to an unexpected consequence of the interaction of components.
- Redirect /biosamples/sample and /biosamples/group URLs in case of typo
v4.0.2
4.0.2
This is a bugfix release that addresses the following issues:
- Fix javascript on SampleTab submission and accession
- Handle load-balanced accessioning
- Fix for storage of relationships source on new samples
v4.0.1
4.0.1
This is a bugfix release that addresses the following issues:
- Fix submission of new unaccessioned samples with relationships by inserting an assigned accession into the source of any relationships that are missing it.
- Fix curation pipeline of numeric organism iri to "http://purl.obolibrary.org/obo/NCBITaxon_+taxId" when it should be "http://purl.obolibrary.org/obo/NCBITaxon_"+taxId e.g. http://purl.obolibrary.org/obo/NCBITaxon_9606
- Allow CORS requests for legacy XML APIs.
- Updated homepage project sample links to use a filter search rather than a text search.
v4.0.0
4.0.0
Version 4.0.0 represents a re-architecture and re-engineering of the BioSamples software stack. It is now based on the Java Spring-Boot framework, utilising MongoDB for storage and Solr for indexing and search. Version 4.0.0 follows modern web standards and conventions, while remaining backwards compatible. This will also give us a strong and stable foundation to build more features and improvements from, more reliably and more rapidly.
Highlights include:
Submissions and updates will be available immediately via accession, and will be available via search within a few minutes or less. There is also improved handling of submissions and updates, with fewer errors and better feedback about any problems.
Integration with EBI AAP for login management and access to pre-publication samples, including use of ELIXIR AAI single sign-on accounts.
Separation of submitted sample information from curation of that information, including the ability for 3rd party (re-)curation of samples. Please contact us if you would be interested in more information and/or to supply curation information.
Improved handling of non-alphanumeric characters in attribute types e.g. "geographic location (country and/or sea)". Previous versions stripped these characters e.g. "geographic location country and or sea".
Improved faceting allowing selection of multiple values within same facet, fixed re-use and re-distribution of search URLs. This will be expanded in future with additional facet types where appropriate.
Support and recommend the use of content negotiation to accessing multiple formats at the same URIs. In addition to the content (HTML vs XML vs JSON) this also supports compression and caching through standard mechanisms.
Java client using Spring, and a Spring-Boot starter module for easy use. This is used by BioSamples internally and other teams at EMBL-EBI, so is high performance and battle tested.
Containerisation using Docker and Docker-Compose, which makes it easier to run a local version for client development or for local storage of sample information.
Data content
Ontology terms Numeric tax IDs (e.g. 9606) and short ontology terms (e.g. PATO:0000384) are being replaced with full IRIs (e.g. http://purl.obolibrary.org/obo/NCBITaxon_9606 and http://purl.obolibrary.org/obo/PATO_0000384 ) in many places and eventually everywhere.
Groups will continue to exist for backwards compatibility purposes. However, we are investigating future development to reduce or remove many of these in favour of alternatives such as filtering samples by external link, or delegating grouping of samples to other EMBL-EBI archives such as BioStudies.
JSON /biosamples
This is the preferred API for use, and uses the same URIs as the HTML pages, and utilising content negotiation to provide a JSON response. This is designed as a hypermedia as the engine of application state (HATEOS) API and therefore we recommend users do not use specific URLs but rather follow relationships between API endpoints, much like a user would use links between HTML pages. It is similar to the /biosamples/api JSON format, with a few critical differences to improve on various issues users have raised:
added release in full ISO 8601 format including time. The backwards-compatible releaseDate exists but should be considered deprecated and will be removed in a future release.
added update in full ISO 8601 format including time. The backwards-compatible updateDate exists but should be considered deprecated and will be removed in a future release.
removed description as a separate field, is now available as a characteristic.
remove relations rel link; equivalent information is now embedded in sample in relationships and externalReferences lists.
remove sample rel link; with relations now embedded, this link serves no purpose.
added curationLinks rel link.
ordering may be different.
fields are not displayed if empty or null.
characteristic names accurately reflect what was submitted and may now be multiple words and may include non alphanumeric characters (e.g brackets, greek letters, etc). In the /biosamples/api responses characteristic names were always camelCased and with non-alphanumeric characters removed.
external references directly embedded in the samples and the groups.
XML /biosamples/xml
We are maintaining this for backwards compatibility. Later in 2018 we will be consulting about future development of this API, with a view to removing it. This is in the context of the improved JSON /biosamples API using content negotiation and several long-standing issues with limitations arising from the XML schema in use. We strongly recommend users switch to the new JSON /biosamples API but nevertheless will be maintaining /biosamples/xml for backwards compatibility, although please note the following differences:
XML element TermSourceREF element Name and element URI are removed.
XML element Property attributes characteristic and comment always false.
elements and attributes may be in different order.
allows only one IRI on attributes, so in rare cases of multiple IRIs will not be complete.
Query parameter query has now a default value of * if none is provided.
Query parameter sort is ignored for the search, due to undefined behaviour and lack of usage.
JSON /biosamples/api
This API should be considered deprecated and we will aim to remove it by 2019. Any users of this should move to using the /biosamples URIs to retrieve JSON representations with an improved schema via content negotiation. Further announcements will be made in future for specific updates and deadlines.
ordering may be different from previous versions, and is not guaranteed for future versions.
fields are not displayed if empty or null.
/api/externallinksrelations/{id}/sample and /api/externallinksrelations/{id}/group are removed due to lack of usage.
fixed externalReferences and publications to be nested objects and not JSON strings.
Acknowledgements
This release has been made possible with the support of our funders:
EMBL-EBI Core Funds
EC -ELIXIR-EXCELERATE
WT- HIPSCI
IMI - EBiSC
ELIXIR – Meta Data Implementation Study
WT-GA4GH
v4.0.0-RC1
increment version to 4.0.0-RC1