Change Log

Notable changes in each release are documented below.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

  • Security - in case of vulnerabilities.
  • Changed - for changes in existing functionality.
  • Deprecated - for soon-to-be removed features.
  • Removed - for now removed features.
  • Fixed - for any bug fixes.

FHIR R5

7.1.0 - 20250211

Security

  • Removed unused h2 jar to resolve warnings regarding CVE-2022-45868
  • Updated library versions to mitigate CVEs:
    • CVE-2016-1000027
    • CVE-2023-5072
    • CVE-2023-5384
    • CVE-2023-6129
    • CVE-2023-6237
    • CVE-2023-33201
    • CVE-2023-34042
    • CVE-2023-34055
    • CVE-2023-41080
    • CVE-2023-42363
    • CVE-2023-42364
    • CVE-2023-42365
    • CVE-2023-42366
    • CVE-2023-42795
    • CVE-2023-44487
    • CVE-2023-45648
    • CVE-2023-46589
    • CVE-2023-52428
    • CVE-2024-0727
    • CVE-2024-0853
    • CVE-2024-1597
    • CVE-2024-2004
    • CVE-2024-2379
    • CVE-2024-2398
    • CVE-2024-2466
    • CVE-2024-4603
    • CVE-2024-4741
    • CVE-2024-5535
    • CVE-2024-6119
    • CVE-2024-6197
    • CVE-2024-6874
    • CVE-2024-7264
    • CVE-2024-8096
    • CVE-2024-9681
    • CVE-2024-11053
    • CVE-2024-12798
    • CVE-2024-20918
    • CVE-2024-20919
    • CVE-2024-20921
    • CVE-2024-20926
    • CVE-2024-20932
    • CVE-2024-20945
    • CVE-2024-20952
    • CVE-2024-21140
    • CVE-2024-21145
    • CVE-2024-21147
    • CVE-2024-21235
    • CVE-2024-21892
    • CVE-2024-22243
    • CVE-2024-22257
    • CVE-2024-22259
    • CVE-2024-22262
    • CVE-2024-23672
    • CVE-2024-24549
    • CVE-2024-25629
    • CVE-2024-25710
    • CVE-2024-26308
    • CVE-2024-29857
    • CVE-2024-30171
    • CVE-2024-34750
    • CVE-2024-38807
    • CVE-2024-38808
    • CVE-2024-38809
    • CVE-2024-38820
    • CVE-2024-38816
    • CVE-2024-38819
    • CVE-2024-38821
    • CVE-2024-38827
    • CVE-2024-38828
    • CVE-2024-45294
    • CVE-2024-47554
    • CVE-2024-50379
    • CVE-2024-51132
    • CVE-2024-52007
    • CVE-2024-55887
    • CVE-2024-56337

Changed

  • Preload will now update a Metadata Resource (one with a canonical), if it has a later Resource.date than the matching currently persisted one.
  • Fix reporting of used-supplement in ValueSet.expansion.parameter
  • Fix bug where Authorization header might not be included in certain HEAD requests for syndication content.
  • Fix version numbers in internal preload feed.
  • The ontoserver.feature.syndication.rejectInconsistentContentItemValues can be enabled to validate that a syndication entry's ContentItemIdentifier and ContentItemVersion match the resource's URL and version.
  • Report fhir-test-cases version in TerminologyCapabilities using Feature extension: http://hl7.org/fhir/uv/tx-tests/FeatureDefinition/test-version
  • preAdopt match.originMap in $translate output to include canonical for the ConceptMap.
  • Ontoserver log messages generated in the context of an API call now include a RequestId so they can be grouped/correlated in log analysis.
  • Migrated from old spring-security-oauth2 library.
  • Add support for check-system-version parameter to ValueSet/$expand.
  • Add check for used-codesystem when considering use of a pre-existing ValueSet expansion.
  • Added support for auto-detecting MMDDYYYY style versions (after 1300) when sorting for most recent
  • Added ontoserver.feature.compat.lookupTypedValue flag to force retention of old (broken) response for $lookup. If you need to the old valueString etc names in $lookup responses, then set ontoserver.feature.compat.lookupTypedValue=true
  • Uplift to support terminology fhir-tests: tests v1.6.0, runner v6.4.0
  • Removed dependency on old 3rd party Atom feed parser (Rome)
  • Added ValueSet filter operator EXISTS support for subproperties.
  • Add support for targetsystem parameter to $translate with unspecified ConceptMap.
  • Upgraded HAPI to 7.4.0 for latest FHIR Validation support: org.hl7.fhir.validation:6.3.11
  • Add support for target and targetsystem parameters to ConceptMap/$translate.
  • Support SCALE_TYP = "DOC" (for any case of “Doc”) in ValueSet definitions for legacy reasons: https://chat.fhir.org/#narrow/stream/179202-terminology/topic/LOINC.20properties.20in.20filters
  • Added validation of partition id and check digit for SNOMED Codes.
  • Ontoserver will now filter syndication feeds as a server as well as as a client
  • Add support for the “top” and “bottom” operators, !!> and !!<, from ECL 2.2
  • Add validation that itemIdentifier and itemVersion from a feed entry matches url and version in the FHIR Resource
  • Assign stable ids to Bundle / Package resources that do not have an id already provided
  • The ontoserver.search.summary configuration parameter can be used to control whether search requests default to summary or complete resources.
  • Cached internal syndication feed entries for significant performance improvement
  • Added support for the content-mode search parameter on CodeSystem
  • Added support for context, context-type and context-type-value search parameters for CodeSystem, ConceptMap, NamingSystem, StructureDefinition and ValueSet resources
  • Broadened support for filter operators on some of the FHIR specification's Defined Concept Properties
  • Add support for delete-history interactions (e.g., DELETE {Resource}/{id}/_history) from R6
  • Improved support for case-insensitive CodeSystems. This includes an index format change for FHIR CodeSystems (migration is performed automatically)
  • Tolerate CodeSystems that attempt to re-define the URIs for standard properties like parent and child.
  • Improved handling and validation of default CodeSystem properties like status
  • Enabled R5 support for the $convert operation
  • Added support for contained CodeSystems. For example, a ValueSet may contain a CodeSystem (or supplements) that are then used in operations on the ValueSet.
  • Added support for the useSupplement parameter, pre-adopted from R5.
  • Added support for ValueSet.expansion.property (via Extensions) pre-adopted from R5.
  • Added support for various SDC extensions:
  • MANY additional changes to support using Ontoserver with the FHIR Validator or IG Publisher as described here and to align with the tx tests in https://github.com/FHIR/fhir-test-cases
    • Improved consistency and detail in $validate-code messages
    • Additional machine-readable output for $validate-code
    • Eliminated redundant designations for use="display" translations
    • …
  • Added support for the Extension http://hl7.org/fhir/tools/StructureDefinion/valueset-expansion-param
  • Added spell-correction support to ValueSet/$expand?filter=...
  • Added the atom.syndication.republishUpstreamFeeds configuration parameter, which specifies the upstream feeds whose entries should be republished.
  • Added support for prometheus endpoint (disabled and not exposed by default) at /actuator/prometheus.
  • Exclude CodeSystems with content=example or content=not-present from use with $validate-code.
  • Added caching for TerminologyCapabilites to improve performance for large sets of CodeSystems.
  • Added support for SNOMED International's new Edition dependency element in a syndication feed.
  • Optional support ValueSet.identifier matching when resolving ValueSet uri in $expand.
  • Allow for inactive concepts to have an inactive FSN.

Fixed

  • Tightened validation of syndication feed entries; ensure that advertised item canonical matches actual resource canonical for Metadata Resources (CodeSystem, ConceptMap, StructureDefinition, and ValueSet).
  • Fixes to code identifying the appropriate display for a code in an expansion, and the set of associated designations.
  • Fix match.source value to be valueUri not valueString in $translate output.
  • Fix bug where Ontoserver could not verify the signature of JWT tokens while using secured syndication feeds.
  • Fixes to cache invalidation propagation when in scaled mode.
  • Fix bug with overly aggressive language matching for display validation.
  • Fix bugs with $validate-code, default / unspecified languages, and displays provided in the ValueSet.compose.
  • Fix bug handling empty search criteria for context-type-value.
  • Fix regression introduced in 6.20.1 with Resource and Resource_Blob table creation.
  • Fix bug when $find-matches is called without any values for property.
  • Fix bug with internal caching and CodeSystem supplement support.
  • Fix bug with internal caching and useSupplement parameter.
  • Fix bug with overly agressive language matching when validating display strings.
  • Fix bug where ValueSets including versioned CodeSystems would not match in a reference search.
  • Fix lazy initialisation of MappingR4.product / MappingR4.dependsOn bug.
  • Fix bug identifying latest binary index version in syndication feed.
  • Fix bug where $expand?designation=... for a “use” would erroneously filter out all languages.
  • More robust ZIP file handling for unicode filenames.
  • Ignore __MACOSX metadata directories in RF2 ZIP files.
  • Fix bug where Community write permissions (e.g., grouping/mycommunity.write) were being ignored.
  • Fix bug with handling of ValueSet filter criteria in the presence of SNOMED and LOINC supplements.
  • Fix bug with JWT token handling where the authorities field was empty and only scope was being used.
  • Fix bug in ValueSet filter regex handling wrt escaped chars: \t \n \r \f
  • Fix bug where content at /static was inaccessible with security enabled.
  • Fix bug where content at /fhir/.well-known could return 401 rather than 404.
  • Improve index cache invalidation following index deletion.
  • Fix bug where $closure would fail when no CodeSystem version was specified and > 1 version of the CodeSystem was available.
  • Fix bug where deduplication of Syndication feed entries did not correctly check contentItemIdentifier and contentItemVersion.
  • Fix bug where stale Lucene index readers were retained in cache.
  • Fix bug where an index that is OK gets marked in the database as not OK (e.g., due to a transient failure during an earlier repair task) but repair never updates the status back to OK.
  • Fixed bug where create/update operations on ConceptMap and other resources would fail due to primary key violations after an Ontoserver upgrade.
  • Fix bug where LOINC/SNOMED-CT CodeSystems indexed with older versions of Ontoserver would appear with caseSensitive=false.
  • Fix bug where display text from ValueSet compose would override a translation selected for the preferred language
  • Fix bug where older index versions could be preloaded over newer ones.
    • If you have been experiencing unexpectedly high disk usage, we recommend using the /api/vacuum endpoint to diagnose and/or remedy that.
  • Fix bug where specifying a fhirVersion other than 4.0 in the ContentType header would fail for an empty body.
  • Fix bug where Lucene indexes were not being closed during a cache flush.
  • Fix bug where failure to fetch a preload feed might not be flagged as a Preload Job failure.
  • Fix bug where caching of syndication feed entries always failed.
  • Only binary indexes, FHIR resources and FHIR packages can be preloaded (no RF2 or LOINC sources)
  • Fixed bug relating to ETags and caching of search results
  • Fix bug in multipart handling for $x-upload-external following Spring 3.0 migration
  • Fix bug in internal caching of feed entries when generating the syndication feed
  • Improved handling of LOINC Linguistic variants (translations) for server LOINC Parts
    • This results in a change in the LOINC index format version from 2.0.4 to 2.0.5
  • Improved logging for certain syndication errors
  • Fix bug in syndication feed generation when atom.syndication.publish.index.enabled=true
  • Fixed $validate-code bug where inferSystem was used with case-insensitive CodeSystems
  • Fixed bugs with actuator endpoints including health when operating in secure mode
  • Removed bogus warnings about an “Unsupported authentication class”
  • Avoid considering CodeSystems with content = example / missing for operations like $lookup, $subsumes, and validation
  • Work around possible case of missing id in persisted JSON blobs
  • Fix bug affecting identifier:missing queries
  • Improved some validation messages for post-coordinated expressions
  • Fixed bug where failures to fetch a Syndication feed were cached indefinitely
  • Fixed performance issues with SEARCH and the reference parameter. e.g., [base]/ValueSet?reference=http://loinc.org
  • Fixed bug where certain incorrect SNOMED CT versions were allowed in a ValueSet definition
  • Fixed bug where ValueSet-specified display text would be ignored when SNOMED CT Editions were used in a ValueSet definition
  • Improved handling of an invalid displayLangauge parameter
  • Fix bug with regex and case-insensitive code systems
  • Fix bugs affecting _id search parameters
  • Fix bug where SNOMED designations with use display could have incorrect language
  • Fixed bug with pre-existing LOINC indexes and case-sensitivity
  • Fixed resolution of implicit ValueSets for CodeSystems provided via the tx-resource parameter.
  • Fixed bug processing CodeSystems without an id provided via the tx-resource parameter.
  • Fixed bug handling a valid coding in CodeableConcept that is not also in the ValueSet.
  • Fixed regression with $expand, the property parameter, and implicit SNOMED reference sets.
  • Add support for caseInsensitive CodeSystem in $validate-code
  • Fixed healthcheck.sh -l semantics to only look at startup preload, and add -L to look for any preload (eg including redoPreload)
  • Fixed bug where failing HEAD requests on syndication feeds would not trigger authorisation token renewal
  • Fixed bug where SNOMED FSN's were mis-characterised during validation
  • Fixed ValueSet resolution bug where tx-resource supplied ValueSets could conflict with persisted ValueSets.
  • Work around Infinispan change that affects Search results when running in a scaled deployment mode.
  • Improved consistency in CodeSystem version resolution within a ValueSet.compose when performing an $expand or $validate-code.
  • Fixed bug with ECL cardinality constraints on role-grouped attributes.
  • Fixed bug in PATCH interactions that used delete with where conditions
  • Improved consistency in CodeSystem version resolution within a ValueSet.compose when performing an $expand or $validate-code.
  • Fixed bug with ECL cardinality constraints on role-grouped attributes.
  • Fixed bug with ECL cardinality constraints of the form [0..n] not matching concepts with no (non-IS A) relationships.
  • Fixed bug resolving appropriate CodeSystem version for the LOINC all-codes implicit ValueSet.
  • Fixed bug storing compressed JSON form of very large resources.
  • Fixed bug where certain hash collisions could cause incorrect $expand, $validate-code, or $validate behaviour. This was extremely rare except in cases where two ValueSet compose clauses differed in only a single concept from a very large CodeSystem.
  • SNOMED Binary Index format changed to 2.0.10; Fixed bug with handling of ECL wildcard and regular expression term matches containing whitespace.
  • Fixed bug where ValueSet.expansion.offset was not being set when paging.
  • Cleaned up unused default Spring Boot user management.

7.0.2 - 20231215

Security

  • Updated library versions to mitigate CVEs:
    • CVE-2023-6378

Fixed

  • Fixed bug storing compressed JSON form of very large resources.
  • Fix bug with version conversion from R4 to R5.

7.0.1 - 20230922

Fixed

  • Fixed bug in run.sh launch script

7.0.0 - 20230918

Security

  • Updated library versions to mitigate CVEs:
    • CVE-2023-38286

Changed

  • Updated to FHIR R5 as supported FHIR Version
  • Binary index versions have all changed to 3.0.0 to support a new version of Lucene

Fixed

  • Fixed bug where ValueSet.expansion.offset was not being set when paging.

FHIR R4

FHIR STU3

FHIR DSTU2.1