* Merge alfresco-transform-model project into alfresco-transform-core - WITHOUT the git history
* Simplify Maven dependencies and build project
* Unify release process of combined t-model and t-core
The t-config returned by the T-Router & T-Engines should include directAccessUrl as an optional transform option when the T-engine supports it.
[trigger release] 2.5.7-A4
The bulk of the changes in this PR are to do with adding a coreVersion element to the transform element in the T-Engine config. For more detail see the class header of CoreVersionDecorator.
* Support the use of coreVersion so that it is possible to upgrade pods in any order
* Moved the majority of the RequestParamMap static finals to alfresco-transform-model and added a new one: "includeCoreVersion" parameter.
* Repeat Bump dependency.tika.version from 2.1.0 to 2.2.1
Original PR https://github.com/Alfresco/alfresco-transform-core/pull/506 was merged to master where it failed. There had been no build of the PR before the merge, which is why this branch has been created.
* Use non deprecated TikaCoreProperties.SUBJECT with tika 2.2.1.
The deprecated OfficeOpenXMLCore.SUBJECT value worked in 2.2.0 but not 2.2.1
* With the upgrade of Tika from 2.2.0 to 2.2.1, the deprecated OfficeOpenXMLCore.SUBJECT metadata value became being null and the replacement TikaCoreProperties.SUBJECT became a multi value in a few of our test cases. For backward compatibility with very old versions of Alfresco, we have historically been added a number of extra values including "subject" and "description" back into the raw metadata, before mapping them onto Alfresco properties. These values existed in the original version of Tika used by Alfresco, so it is possible there are custom mappings out there that using them.
To complicate matters a little, out standard mappings for some types put the raw "subject" value into cm:description property. What makes it interesting is that the extra "description" value is not used but has the value originally in our expected metadata extarct data. That is why the quick_*_json files have been modified.
This reverts commit 9776577a45.
Was not possible to perform the release of 2.5.5-A1 with this upgrade of Tika.
Possibly related to it forcing a change in the following files, which were then deleted in the build:
D alfresco-transform-core-aio/alfresco-transform-core-aio-boot/src/license/THIRD-PARTY.properties
D alfresco-transform-core-aio/alfresco-transform-core-aio/src/license/THIRD-PARTY.properties
D alfresco-transform-tika/alfresco-transform-tika-boot/src/license/THIRD-PARTY.properties
D alfresco-transform-tika/alfresco-transform-tika/src/license/THIRD-PARTY.properties
Revert ATS-969 Tika upgrade 1.x -> 2.x (#493)
As the build is deleting the following, resulting in the release job failure
D alfresco-transform-core-aio/alfresco-transform-core-aio-boot/src/license/THIRD-PARTY.properties
D alfresco-transform-core-aio/alfresco-transform-core-aio/src/license/THIRD-PARTY.properties
D alfresco-transform-tika/alfresco-transform-tika-boot/src/license/THIRD-PARTY.properties
D alfresco-transform-tika/alfresco-transform-tika/src/license/THIRD-PARTY.properties
[trigger release] 2.5.5-A1
By default T-Engines now provide the more readable TransformerDebug DEBUG messages, rather than the original detailed request and reply messages, which are still available as TRACE.
* T-Engines config need to be combined in the same predefined order in the all-in-one transformer,
T-Router and the Content Repository with individual T-Engines.
* T-Engines config need to be combined in the same predefined order in the all-in-one transformer,
T-Router and the Content Repository with individual T-Engines.
* Introduce more flexible reading of UTF-16 data, where there may be a BOM, but the
spec says there should not be one, or the BOM is clearly wrong when looking at the
following characters. The https://en.wikipedia.org/wiki/UTF-16 write up is nice and clear.
* Includes identical correction in data setup in AIOTransformRegistryTest and
MicsControllerTest for a problem found in TextToPdfContentTransformerTest.
* Includes upgrade to latest pdfbox: 2.0.22
Bug found while reviewing documents on how to create a custom metadata extractor. The original refactor had left the repo doing the mapping. It should have been passing the fully qualified repo properties to the T-Engine to do the mapping.
Linked to:
Alfresco/alfresco-community-repo#227Alfresco/acs-packaging#1826
ATS-829: Release T-Core (T-Engines) 2.3.6 [trigger release]
Linked to REPO-5219 Allow AGS AMP to specify metadata extract mapping
Added an extractMapping transform option to all metadata extractors to override the default one.
3rd party libraries to get a green build.
* Upgrade cxf-rt-transports-http and woodstox-core to avoid issues
* Upgrade to org.springframework.boot:spring-boot-starter-parent:2.3.5.RELEASE to avoid problem in org.springframework:spring-web
* Upgrade to activemq 5.15.13 to avoid problem in activemq-broker 5.15.12
* Metadata extract code added to T-Engines
* Required a refactor of duplicate code to avoid 3x more duplication:
- try catches used to return return exit codes
- calls to java libraries or commands to external processes
- building of transform options in controllers, adaptors
* integration tests based on current extracts performed in the repo
* included extract code for libreoffice, and embed code even though not used out of the box any more. There may well be custom extracts using them that move to T-Engines
* removal of unused imports
* minor autoOrient / allowEnlargement bug fixes that were not included in Paddington on the T-Engine side.
* REPO-5188 SourceEncoding should not be used to select transforms as it is always provided to T-Engines
* Fix test. Now that SourceEncoding is in the options but is not used to select the transformer, we need to add another
option to force the use of the textToPdf transform rather than libreoffice.
Set priorities on 5 Tika transform routes so they are not used if LibreOffice is available. This currently has no impact on the Transform Service (priorities are not used for routing), but does on repository LocalTransforms as the wrong T-Engine is selected.
* REPO-4639: Split tika engine_config.json into separate transformers.
* WIP: REPO-4639 Content conversion failed using Tika
The Tika T-Engine "transform" option does not exist when called via the Transform Service or Local transforms, which resulted in no transforms taking place. However this value is really not be needed as the T-Engine should be able to read its own engine_config.xml to work out which sub transform to use. Transforms only worked via Legacy transforms, which used a T-Engine.
This code is based on tried and tested ACS repository code. It has been further simplified.
TODO:
- replace the ConfigFileFinder class just added with something that uses Spring to read the JSON. i.e. simplify it.
- replace the CombinedConfig class just added with something that does not need the InLineTransformer. i.e. simplify it.
- create tests based on the repo tests
- remove the source and target mimetype checks in Tika as a check against engine_config.xml is cleaner.
- repeat the process for the Misc T-Engine as it has similar code checking source and target mimetypes.
- remove the transform option passed by the legacy transforms.
* Removed CombindConfig and ConfigFileFnder classes.
* Extracted AbstractTransformRegistry so that it may be used in the ACS repository too.
TODO AbstractTransformRegistry and AbstractTransformRegistry need to be moved to the alfresco-transform-model pakage
* tidy up only
* REPO-4639: Add priority to duplicate transforms.
* REPO-4639: Refactor TikaTransformationIT to use the new Tika /transform specifications
Changes AbstractTransformerControllerTest as the engine_config is now loaded in TransformRegistryImpl instead of AbstractTransformerController
* Rename to TransformServiceRegistry, so we don't have to change the repo code.
* Added the baseUrl parameter to the register method and fixed the missed rename in the last commit.
* Javadoc change only
* Moved common classes (with repo) AbbstractTransformRegistry and TransformServiceRegistry to alfresco-transform-model
* Replace (simplify) all the isTransformable calls with a check against the JSON.
- Tests now only pass targetEncoding to the 'string' transformer.
* Fix failing tests.
* Revert port change
* REPO-4639 : Add priorities to misc engine_config
* REPO-4639 : Add priorities to pdf-renderer and imagemagick engine_config
* Remove test that is @Ignored
* Pick up alfresco-transformer-model 1.0.2.7-REPO-4639-1
* REPO-4639 : Add priorities to libreoffice engine_config
* REPO-4639 : Add priorities to tika engine_config
* REPO-4639 : Remove all priorities with value equal to 50 (default) from engine_config
* Switch over to using TransformServiceRegistry in org.alfresco.transform.client.registry
Reintroduce the noExtensionSourceFilenameTest having removed @Ignore.
* New whitesource issue on commons-compress 1.18. Upgrading to 1.19.
* Removed the text/javascript -> text/plain test as this is not supported
* Modifications as a result of changes to method names in alfresco-transform-model
* Pick up alfresco-transform-model 1.0.2.7-ATS545-2
* Remove unused imports
alandavis
