* lockdown bulk upload routes
* able to view lettings bulk upload errors
* add error count to bulk upload results
* coverage for bulk upload filename on results
* group bulk upload errors by row on results
* able to view bulk upload sales results
* scope lettings and sales bulk upload results
* fix linting
* call service correctly in test
* add bulk upload sales questions mapping
* appease linter
* bulk upload error shows correct question
- depending on log type it will show relevant question for the field
concerned
* use local disk for bulk upload for dev env
- this saves the need to connect to S3 to play with bulk upload in dev
environment
* improve namespacing of classes
* add job to process bulk uploads
* use local disk storage for dev file upload
* fix test
* use inline active job queue_adapter for dev
* use test active job queue adapter for test env
* remove rubocop violation
* delete bulk upload from disk after processing
* populate errors with cell, row + metadata
* update error message with something meaningful
* shim in sales validator
* able to parse sales bulk uploads
* change migration to add purchase_code to errors
* bulk upload error component renders purchaser code
- when a sales log
* populate purchaser_code for bulk upload errors
- when log type is sales
* remove superfluous private method
* Validate that the user belongs to either the managing or owning organisation
* do not reset created_by and remove user_organisation_chosen?
* Do not default managing organisation to owning organisation
* validate user belongs to organisation
* update tests to specify created_by
* clear create_by for support users
* refactor
* typo
* rebase lint
* Add `updated_by` to logs to track who was the last person to update a log
- Reset `created_by` automatically when a form is updated and the owner does not belong to the managing or owning organisation
* move reset_invalidated_dependent_fields!, update schema file and fix tests
Co-authored-by: James Rose <james@jbpr.net>
* Mark log as impacted by deactivation when location is deactivated
* Display affected logs in the table
* Route affected logs to tenancy start date question
* Update routes to get the tenancy start date page from form
* update next_page_redirect_path
* rename column
* Fix tests
* Add next_unresolved_page_id to pages
* Update unresulved when the log is corrected
* Mark logs as unresolved when scheme gets deactivated
* display correct content when there are no unresolved logs
* mark logs resolved after the user leaves check answers page
* Display link in success banner and reset banner when the link is clicked
* display inset hint text for unresolved log questions
* Display unresolved logs banner
* update banner message
* Persist the link in the banner
* update inset text
* Update success banner text
* Add unresolved and created_by scopes
* rename method
* add unresolved_log_redirect_page_id to form + typo and route
* Add UnresolvedLogHelper and extract flack notice message
* pluralize and return early
* remove flash[:notice] = nil
* to keep it consistent for sales log
* Extract unresolved path and fix a link
* extract resolve method and fix attribute nme
* update path
* typo
We import XML files that define logs from the previous CORE service. For an unknown reason, some of these files don't include the required namespace declarations to parse `user` values. Instead of changing the source, this change introduces a new method that forces the namespace declaration for user fields.
These fields are set using IDs from the previous CORE service. There was an incorrect assumption that these fields were integers so we were casting them as such. We have discovered that these fields are strings (e.g., `027`) so this change adjusts our types.
We were previously blocking the import of locations that belonged to schemes where the end date was before the current date. This meant that we were unable to import logs that are associated with schemes that _were_ valid, but have since expired.
We import XML files that define logs from the previous CORE service. For an unknown reason, some of these files don't include the required namespace declarations to parse `meta` values. Instead of changing the source, this change introduces a new method that forces the namespace declaration for meta fields.
Only import fields from the old CORE service that are logically valid. If a log is `saved` or `submitted-invalid` then it might have fields that are logically invalid at import. We can safely blank these out for the user to re-input.
* Remove an unused ethnic_other column
* remove sale_completion_date from lettings logs
* Add collection start year
* Remove sale_or_letting column
* Rename rent_type to rent_type_detail in the export
* Format dates
* refactor
* Fix test
* Set age and relat to nil when inconsistent
* Set homeless and reasonable preference due to homelessness to nil if inconsistent
* Age not answered case
* Refused is also valid
* Check relationship has been answered
While migrating users from the previous service to the new one we discovered that email addresses are not unique in the previous service. This means that one user in the new service might relate to multiple users in the previous service.
This change:
- Adds a new LegacyUser model that can belong to a User. Each LegacyUser model has one old_user_id that corresponds to the user ID in the legacy service.
- Updates the user import service so that we create this association for new users
- Creates a Rake script to backfill the association for existing users
Phil Lee
Kat
Jack
kosiakkatrina
James Rose
baarkerlounger
natdeanlewissoftwire
natdeanlewissoftwire