Fix - Adding filter of summarized field grouped by latitude and longitude on a map causes table visualization error (#31952)

* Fix #30058

* Add missing setting key

* Add a unit test attempt at #30058

* Mock store instead of using deprecated MetabaseSettings.set

* Remove LeafletGridHeatMap unit tests

* Add e2e test for #30058

* Fix typing

* Don't allow undefined "map-tile-server-url"

* Tag the test with repo and issue id

Replay is adding a significant overhead and is making every single run
fail. This is messing with our statistics and analytics.

While the Replay.io team is working on some optimizations that would
remove the overhead, we can significantly dial down the frequency of
these runs. We'll switch to running them only once per week, on Sunday.

For now...

Debian Bookworm has Java 17 as the default and Java 11 has lots of steps to install. Going back to Bullseye with Node 18

There was a discrepancy between TemplateTag id types as defined in metabase.lib.native (uuid) and metabase.mbql.schema (string). This PR makes string the uniform type for both. This results in errors when code using different specs for the same logical entity is used together. Here's an breaking example:

```
(let [card-id 328 ;; This is a valid card locally
      q   (str "SELECT * FROM {{#" card-id "}} LIMIT 2")
      tt  (lib-native/extract-template-tags q)
      res (qp/process-query
            {:database 1
             :type     :native
             :native   {:query         q
                        :template-tags tt}})]
  (is (some? res)))
```
This will break since the tt generated id is a uuid but process-query expects a string id.

Fixes #31252 

* require table primary key to enable basic actions

* add unit tests

* update model detail page test metadata

* remove disabled columns from isValid check

* Applying column indexes before filtering

* Adding test to GH 29140, updating test helpers in settings.cy.spec.js

* Adjusting indexes when sorting via table header

* e2e test adjustments

* Fix component name

* Show only 1 title line on small SmartScalars

* Style PreviousValueVariation and the separator
- Get rid of flexbox so that the default browser text-wrapping algorithm is used
- Be pixel-perfect about margin-left

* Bring back height calculation based on number of lines

* Format change

* Fix ellipsis in ScalarValue

* Make spacing consistent

* Style VariationIcon and VariationValue
- get rid of a utility class and inline CSS

* Add ellipsis to previous value, style tooltip

* Ensure consistent width of SmartScalar children

* Update JS to match CSS

* Refactor measureText to return the whole TextMetrics object

* Extract hardcoded padding out of ScalarValue

* Update measure text to return height explicitly

* Account for height in the findSize function

* Improve spacing

* Improve styling

* Fix build, bring back removed styling

* Make previous value responsive

* Revert #31724

* Fix typing
- Introduce TextWidthMeasurer type

* Remove redundant CSS and elements

* Rename formatChangeSmart to formatChangeAutoPrecision

* Fix ellipsis on previous value

* Improve 2 lines threshold value and introduce explanatory constant

* Format code

* Remove redundant code

* Update findSize tests

* Fix overflow in Scalars

* Add overflow test boilerplate

* Include more sizes in the test

* Format code

* Add tests for scalar
- add scalar-description to assertions

* Extract assertDescendantNotOverflowsContainer

* Remove redundant check

* Extract setupDashboardWithQuestionInCards

* Reorder cards from biggest to lowest horizontally

* Extract assertDescendantsNotOverflowDashcards

* Add test case for when sidebar is open

* Format code

* Ignore elements without dimensions in assertions

* Reorder cards from biggest to smallest vertically

* Add 1x cards in tests for scalars

* Add tests for various viewports

* Improve naming

* Format code

* Fix cursor: pointer being shown on container hover instead of icon hover

* Improve vertical icon alignment, make ScalarDescriptionPlaceholder padding consistent with ScalarDescriptionContainer

* Account for title lines count in Scalar

* Refactor tests to allow better test case descriptons

* Use useful viewports

* Find a testable way to fix ellipsis

* Adjust test data

* Add assertion messages

* Fix tooltip not shown on value when ellipsis is applied

* Add test for tooltips in 1x2 card

* Add tests for 2x2 & 5x3 cards
- fix CSS issues breaking tests

* Add tests for 2x2 card

* Add tests for more card sizes

* Assert ellipsis is applied

* Fix assertions

* Update unit tests

* Rename TITLE_2_LINES_THRESHOLD to TITLE_2_LINES_HEIGHT_THRESHOLD

* Fix typo

* Make mock data reflect real data a bit more

* Fix cursor: pointer being applied on the container hover instead of icon

* Don't assume the "space" function will return values in pixels

* Move Scalar to its own directory

* Extract constants.ts and utils.ts from Scalar

* Move SmartScalar to its own directory

* Extract constants.ts and utils.ts from SmartScalar

* Extract styles computation out of SmartScalar

* Extract getValueWidth & getValueHeight to utils

* Update value height computation

* Update typing

* Make capitalization work as intended

* Fix inversed calculation

* Add missing gridSize prop

* Remove capitalization logic and improve breakpoints
- I believe the original intention was to capitalize only the first character, but every word is capitalized and it looks weird

* Extract createCardsRow

* Break down the tests into more suites because cypress can't consistently handle loading 20 dashboard card simultaneously

* Improve names

* Remove mistakenly added size 1 smart cards from tests

* Sort props to see differences

* Fix vertical center alignment in Scalar
- this change also affects SmartScalar, see discussion: https://metaboat.slack.com/archives/C057T1QTB3L/p1688020307409629?thread_ts=1687997714.132239&cid=C057T1QTB3L

* Replace within() + contains() with findByText()

* Don't export getTitleHeight
- move it to below where it's used

* Remove unused gridSize prop

* Use consistent code style

* Simplify separators logic

* Remove concatRecursively

* Fix typo

* Optimize test suite

* Fix typo

* Add unit tests for Scalar utils: getValueHeight & getValueWidth

* Add unit tests for SmartScalar utils

* Fix e2e tests
- update card width in test caes
- wrap cases requiring cy.viewport() in it to prevent test issues

* Migrate notebook aggregations to MLv2 — Operators (1) (#31527)

* Add aggregation types

* Add aggregation TypeScript wrappers

* Move `AggregateStep` to its own directory

* Port basic features to MLv2

* Deprecate `AggregationPopover`

* Fix MLv1's aggregation clause validation

* Fix import

* `requiresField` → `requiresColumn`

* Add basic tests for `AggregationPicker`

* Add `findAggregationOperator` test utility

* Add basic tests for `AggregateStep`

* Make stage index explicit

* Highlight selections, add picker back button

* Use long display name for aggregation clauses

* Remove `React` imports

* Enable temporal bucketing

* Use aggregation JS wrappers

* Fix `Icon` import

* Patch unit test according to BE changes

* Fix E2E test

* Simplify `Aggregation's` `dimension` method

* Fix formatting

* Migrate notebook aggregations to MLv2 — Metrics (2) (#31528)

* Add basic legacy metrics support

* Don't show archived metrics

* Highlight selected metric

* Extract `AggregationPopover` in notebook step

* Fix `onClose` prop

* Fix icon

* Add missing metric types

* Use MLv2 to manage metrics

* Remove legacy query logic

* Add `Aggregatable` type

* Migrate notebook aggregations to MLv2 — Inline expressions (3) (#31529)

* Revert "Remove legacy query logic"

This reverts commit 722804d3a6684762701351c658625b5439c75c2b.

* Use HTML labels in `ExpressionWidget`

* Support inline expressions for aggregations

* Fix import

* Fix picker navigation

* Fix type error

* Remove `ts-expect-error`

* Migrate notebook aggregations to MLv2 — Disable MLv1 validation (4) (#31530)

Running E2E tests using ReplayIO still adds a significant overhead, which
leads to a lot of failed runs. There's no point in re-running those tests
for at least two reasons:
1. they will fail again and we're just wasting CI infrastructure
2. our slack-failure-alert-bot will spam our Slack with the reports of
the failed runs that we don't (currently) care about

[ci skip]

* Add full-app embedding tests

* Clean up unnecessary `within` call

* Clarify code comments

* Improve clarity on test utils

 * don't navigate away from the current model if the user searches for an item in that model

* Add joinedThing and pickerInfo

Fixes #31920.

fix object detail record nav for non-integer keys\ndont allow navigation past the end of the object detail data with the keyboard (#31912)

* Extra Metric tests; include `:description` in the Metric `display-info`

* Add `available-metrics` function & TS wrappers

* Support passing a Metric to `lib/aggregate`

* Include `:selected` information in Metric display info

* Fix indentation

* Revert changes to arithmetic type-of calculation

* Test fixes 

* Address PR feedback

* Address PR feedback

* Update docstring