* Bump pandas 1.4 and pyarrow 6
* Use engine="pyarrow" for pd.read_csv()
* Refactoring
* Refactoring
* Refactoring
* Use bytes in pd.read_json()
* Fix test_contribution
* Fix pandas issue when 'arrays' are empty but 'names' contain values
* fix: ValueError: For argument "ascending" expected type bool, received type NoneType.
* Remove engine="pyarrow" and convert bytes to string
* make copy of selected df to fix regression
* Simplify pd.read_json() and pd.read_csv() for example data
Co-authored-by: Ville Brofeldt <ville.brofeldt@apple.com>
* pass datasource_type and datasource_id to form_data
* add datasource_type to delete command
* add datasource_type to delete command
* fix old keys implementation
* add more tests
* fix: set correct schema on config import
* Fix lint
* Fix test
* Fix tests
* Fix another test
* Fix another test
* Fix base test
* Add helper function
* Fix examples
* Fix test
* Fix test
* Fixing more tests
* feat(explore): new datasets have autocomplete filters enabled if UX_BETA is set
* Lint fix
* Set autocomplete to True in example datasets
* Use "IN" as a default operator in dnd filters
* Default operator only if UX_BETA enabled
* Add presto to the CI
Sample test data
Datetime conversion
Sample test data
Fix tests
* TODO to switch to timestamps
* Address feedback
* Update requirements
* Add TODOs
Co-authored-by: bogdan kyryliuk <bogdankyryliuk@dropbox.com>
* feat: [explore] don't save filters inherited from a dashboard
When navigating to explore from a dashboard context, the current
dashboard filter(s) are passed along to explore so that the context is
kept. So say you're filtering on "country=Romania", in your dashboard
and pivot to explore, that filter is still there and keep on exploring.
Now a common issue is that you'll want to make some tweak to your chart
that are unrelated to the filter, say toggling the legend off for
instance, and then save it. Now you back to your dashboard and even
though you started with an "all countries" dashboard, with a global
filter on country, now that one chart is stuck on "Romania". Typically
you notice this when filtering on something else, say "Italy" and then
that one chart now has two mutually exclusive filters, and show "No data".
Now, the fix is to flag the filter as "extra" (that's the not-so-good internal
name we use for these inherited filters) and make it clear that that
specific filter is special and won't be saved when saving the chart.
* fix build
* Typo fix in CONTRIBUTING.md
* Alter references to config.get('FOO') to use preferred config['FOO']
* Set missing configuration constants in superset/config.py
* Misc. CI fixes
* Add type annotation for FEATURE_FLATGS
* [WiP] improve load_examples
related to #7472, longer term we will generate the examples by exporting
them into tarball as in #7472. In the meantime, we need this subset of
the features:
* allowing specifying an alternate database connection for examples
* allowing a --only-metadata flag to `load_examples` to load only
dashboard and chart definitions, no actual data is loaded
* Improve logging
* Rename data->examples
* Load only if not exist
* By default do not load, add a force flag
* fix build
* set published to true
