Smart 1 Improvements (t_type consistency and optimization strategy) #93
Conversation
signedav
changed the title
…not a table created (like for smart1) only concern t_type option when it's relevant. concerning the selected strategy has to be implemented.
There was a problem hiding this comment.
Overall, looks good to me.
The complexity here is a bit out-of-scope for me at the moment, I should really have a look at some MB documentation on polymorphic relations, optimization strategies, and the like. I'd appreciate if you can point me to some recommended resources (up-to-date docs, presentations, etc).
| field_idx = ( | ||
| fields.lookupField("T_basket") | ||
| if not_pg | ||
| else fields.lookupField("t_basket") |
There was a problem hiding this comment.
It would be great to have a variable holding the correct basket field name regardless of the engine (e.g., fields.lookupField(BASKET_FIELD_NAME)).
I'd have to refresh a bit how the Model Baker code has evolved to remember if this could be added as a class member. Just creating a note here in case you can think of something evident.
There was a problem hiding this comment.
Thanks I will concern that.
| # check filter in widget | ||
| ews = t_basket_field.editorWidgetSetup() | ||
| map = ews.config() | ||
| dataset_table = "T_ILI2DB_DATASET" if not_pg else "t_ili2db_dataset" |
| if layer.layer.name() == "Strasse": | ||
| count += 1 | ||
| value_map = layer.layer.editFormConfig().widgetConfig( | ||
| "T_Type" if not_pg else "t_type" |
Yes, this first part was mainly about to provide the |
… and provide the baskets correctly
t_typeFieldGet possible values for
t_typevalue relation for gpkg and mssql.This is done over the
t_ili2db_column_proptable, where the values are provided for every system.With PG it used to parse the check constraint for that in
get_value_map_info, not sure if this becomes obsolete now, but I keep it for the moment.Where this functionality is implemented claeis/ili2db#241
Where it was implemented differently for pg opengisch/QgisModelBaker#389
Where it has been discussed opengisch/QgisModelBaker#134
And this here resolves opengisch/QgisModelBaker#797
Beautify naming
Since the value in
t_typeused to be an sql-table name describing the class (but neither an existing table (because it's not created with smart1) nor a proper ili-name), it's now displayed as the Ili-Element-name.Optimize Strategy
With Smart1 a strategy named GROUP makes no sense: There is HIDE or NONE, while HIDE hides all the irrelevant values for
t_typeand provides only the relevant baskets.