Currently Mica depends on Opal for getting the data dictionaries and summary statistics. The proposal is to make use the Magma abstraction to access ValueTables from other Datasource than the RestDatasource. For instance, the ExcelDatasource could be used to retrieve data dictionaries from a file stored in Mica.
The work would then consist of generalizing the StudyTable attached to a Dataset so that it is not limited to the "project" and "table" properties which are Opal specific. A URN could be used instead, to identify the underlying ValueTable. Its location would be resolved in the context of the Study/Dataset: for instance when a specific Opal server URL is defined for the Study, the table's location can be resolved by combining this Opal server URL and the StudyTable's URN.
This also opens the possibility to have a "study-table-source" plugin type, to support future alternate sources of value tables.
These changes would also affect the variable summary and contingency retrieval, as the statistics may be differently supported by the underlying Datasource: for instance, when using the ExcelDatasource, the summary statistics could be extracted statically from the variable attributes and the contingency feature would not be available.
Currently Mica depends on Opal for getting the data dictionaries and summary statistics. The proposal is to make use the Magma abstraction to access
ValueTable
s from otherDatasource
than theRestDatasource
. For instance, theExcelDatasource
could be used to retrieve data dictionaries from a file stored in Mica.The work would then consist of generalizing the
StudyTable
attached to aDataset
so that it is not limited to the "project" and "table" properties which are Opal specific. A URN could be used instead, to identify the underlyingValueTable
. Its location would be resolved in the context of the Study/Dataset: for instance when a specific Opal server URL is defined for the Study, the table's location can be resolved by combining this Opal server URL and the StudyTable's URN.This also opens the possibility to have a "study-table-source" plugin type, to support future alternate sources of value tables.
These changes would also affect the variable summary and contingency retrieval, as the statistics may be differently supported by the underlying Datasource: for instance, when using the ExcelDatasource, the summary statistics could be extracted statically from the variable attributes and the contingency feature would not be available.
Affected components: