* Add sleeps so tests will pass
* removed a space in a comment
* replaced some sleeps with waiting for loading indicator
* Removed context page object calls from filter_bar and doc_table services
* Add tmsLayers placeholder array to collections. Bind collections to wms-options scope
* Move tmsLayers attribute declaration to base visualization creation
* Add test to verify presence of road_map base layer option
* Move fieldsFetch logic into the vis editor
* Add annotations index pattern change detection
* Fix async update of state. Add functional test
* Add missing data archive
* Force fetch when component mount the first time
* Fix parameters naming
* Refactoring indexPatterns to fetch
* Replace xvfb with headless browser
* Update karma to run in headless chrome
Update Karma tests for an option to run chrome in headless.
* Fix lint issues
* Fix spaces
Closes#22840
In the functional tests we want a better way to include/exclude certain tests, especially as we move forward with #22359. This PR allows us to decorate suite objects with "tags", which won't clutter up the test names and can be used to filter out specific tests within a single test config. The functional test runner supports defining `--include-tag` and `--exclude-tag` CLI arguments, and multiple can be defined.
The method of filtering out tests for running against cloud has been updated to use this approach and I plan to do the same to #22359 once this is merged.
* Add legacy response handler for table vis.
The new legacy response handler introduced a regression in how nested
tables were handled within table vis. This adds a new table-specific
response handler to ensure splitting is preserved.
This is a short term solution and will be removed once we are able to
update table splitting to be consistent with other vis types.
* Ensure formatted dates are preserved in table titles.
* Update legacy table response handler based on feedback.
* Ensure AggConfigResult.rawData is preserved in legacy table response handler.
* Move legacy table response handler to core_plugins.
* Legacy table response handler - style cleanup.
* Remove unneeded aggConfigResult.rawData from legacy table response handler.
* Add basic unit tests for legacy table response handler.
* In table vis, exclude split columns when showing metrics at all levels.
* Add functional tests
* Fix missing check on empty response
* Fix test order and remove applying on each entered filter
* Change quotes on addNewFilterAggregation testsubject
* Rename test hasPieChartError method to expectPieChartError
* Refactor piechart zero-value slices data cleaning.
The previous implementation used to remove zero-value slices by mutating the vis data from the legend logic.
We moved the logic of "cleaning" the zero-value slices before rendering the piechart and/or the legend, so now piechart and legends are rendering themselves with the same data structure.
* Reverting _validatePieData method to the old one
Implements query bar portion of https://elastic.github.io/eui/#/layout/header. Filter bar will come in another PR.
Fixes#14086
Re-implements our query bar component in React using some EUI components. Existing typeahead and suggestion styles were copied over 1:1 for now after talking with Dave about it. In this PR I focused on reaching feature parity with the existing query bar. Some additional work would be needed before we could move this into EUI as a generic component that could be consumed by other plugins.
Still needs some new tests and I suspect some old tests will need to be updated, but other than that this PR is functionally complete and ready for reviews.
This PR adds two usage stats to our telemetry for KQL:
* How many times people click the opt in/out toggle in the query bar UI
* Which language Kibana admins have set as the global default in advanced settings
* fixing mangling of floating point numbers by console
* fixing tests
* fixing issue with large requests
* restoring old code for server side as it handles large responses better
* Use an instance of SavedObjectsSerializer for migrations and the repository
* Fixing spelling of serialization
* Making the serializer conditionally include and prepend id with ns
* Adding repository tests for the namespaces
* Implementing find
* Modifying the SOCs to pass the options with the namespace
* Centralizing omitting the namespace when using serializer.rawToSavedObject
* Passing the schema through to the SavedObjectRepositoryProvider
* Changing the schema to work with undefined ui exports schemas
* Adding schema tests
* Making the complimentary serialization test use the namespace
* Fixing uiExports
* Fixing some tests
* Fixing included fields for the find
* Fixing include field tests, they're checking length also...
* Updating Repository test after adding namespace to always included
fields
* Renaming UIExportsSavedObjectTypeSchema to SavedObjectsSchemaDefinition
* Completing rename... forgot to save usages
* Fixing issue with the serialization.isRawSavedObject and the trailing :
* extract reusable save component from DashboardSaveModal
* update discover search to use SavedObjectSaveModal
* create generic show_save_model that works for both discover and dashboard
* fix last bits of discover save
* remove old save functionallity
* migrate visualize save to EUI
* fix functional tests
* disable save button if title is empty
* mark title input as invalid when title is not provided
* fix funtional tests
* move find logic to SavedObjectFinder component since savedObjectClient is no longer coupled to angular
* implement flyout open saved searches
* remove old open stuff
* add jest test for OpenSearchPanel and simplify panel title
* fix functional tests
* fix _lab_mode functional test
There have been several failures in this test, seemingly caused by a lack of sorting in the results. It makes sense that since both migrations are run simultaneously that sometimes one would succeed and sometimes another would, so I've just sorted the results before checking.
cc: @chrisdavies
* Bring back most of the filtering tests, only bypass the known current issues.
* Clean up after the test suite run so it can be run multiple times in a row.
* goal and guage issues were because of a bad cache
