* Add team filter to cohort query to speed up
* Fix querying
* Remove emails from CSV upload
* fix clickhouse email
* Add person email index
* back to concurrent
* Update demo data warning styles
Left-over from https://github.com/PostHog/posthog/pull/3103
* minor adjustment
Co-authored-by: Paolo D'Amico <paolodamico@users.noreply.github.com>
* Make PluginConfig.order mandatory
* upgrade plugin server to 0.6.10
* black
* Ensure no PluginConfig.order collission
* Clean up PluginConfig API
* Revert "Ensure no PluginConfig.order collission"
This reverts commit d8b2c907e4.
* Black
Co-authored-by: Marius Andra <marius.andra@gmail.com>
* Unify clickhouse and pg demo data creation
* Adapt yakkos demo data generation to normal demo
* Split demo.py into subfolder
* Add original, web data import
Note that elements are not added because they can't be added
automagically :(.
* Improve wording
* dont generate dashboards
* try get tests passing
* fix types
* Get cypress tests green
* fix form type
* split to enabled/disabled plugins and add empty and loading screens
* show ingestion order
* add "rearrange" button
* show alert when rearranging
* drag to sort plugins
* fixes
* move opt-out to "danger zone" in "advanced" tab
* refactor "custom" to "advanced" in the code
* rename to DangerZone
* start rearranging from dragging card
* show changed order
* remove extra button
* save order in the backend
* fix default config for newly installed plugins
* newly enabled plugins go to the end of the list
* mypy
* hide second "check again" button
* show code diff
* lowercase
* clarify string
* reload only if saved, add 403 status if no access, simplify dict
* set temporary order only on enabled plugins, avoiding rearranging "plugins to update" list
* simplify condition
* change empty component
* raise instead of 403
During tests it's making everything quite unreadable:
```
AXES: Cleaned up 0 expired access attempts from database that were older than 2021-01-26 07:30:23.776885+00:00
AXES: Successful login by {username: "user1@posthog.com", ip_address: "None", user_agent: "<unknown>", path_info: "<unknown>"}.
AXES: Cleaned up 0 expired access attempts from database that were older than 2021-01-26 07:30:23.786481+00:00
AXES: Successful logout by {username: "user1@posthog.com", ip_address: "None", user_agent: "<unknown>", path_info: "<unknown>"}.
```
* Migration for is_demo
* Populate is_demo when triggered
* Refactor: extract create_demo_team
* Also create ee sample data
* Create demo data if feature flag enabled (and in debug), redirect to root
* Show demo warning
* Solve rease issue
* Redirect to personalization
* Redirect to root
* Update antd to latest, which gives access to Alert action prop
* Update demowarning code
* update tests
* Update naming
* Update link
* Solve a regression case
This probably solves
https://sentry.io/organizations/posthog/issues/2180671668/?project=1899813&referrer=slack
but I get a slightly different error locally
* Quietly ignore (and report) errors in get_active_feature_flags
This method is used in mission-critical endpoints (/e, /decide), which
should avoid 500-ing due to application errors.
With capture_exception we still get a sentry error which should help us
debug
* refactor Signup location
* personalization route stub
* radio buttons base UI
* select radio options
* continue button
* save personalization state
* report personalization answers to PH
* all data & nicer UI
* mark personalization completed on DB
* cannot complete personalization twice
* rename components
* more fixes
* undo completed_personalization as per discussion, @macobo's PR coming
* fix ts
* Move adding feature flags to event to capture endpoint
* Remove unused imports
* Don't create empty event properties dict
* Don't ever create event properties if dict not present
* Test feature flags in test_capture
* Ensure that every event has a properties dict
* Only apply PLUGIN_SERVER_INGESTION_HANDOFF to plugin-whitelisted orgs
* Adjust tests for ensured event properies dict
