Closed pasqLisena closed 8 years ago
You made a typo in listing the possibilities in Group 3: you have twice F25 and F31 with different suffix names.
I agree that you should instantiate the classes and properties only when you have the information in the source data, and not systematically.
You made a typo in listing the possibilities in Group 3: you have twice F25 and F31 with different suffix names.
The typo was in the following note about the refactoring. I edited it now.
I don't understand. How can we have: F31_Performance AND F31_AutrePerformance; F25_PerformancePlan AND F25_AutrePerformancePlan? Same Fxx but different concept?
It should be refactored, there is an issue still open #5 but not yet resolved.
Closed with last commit.
Here are the classes that we currently instantiate in marc2rdf:
About the 3rd group, a refactoring is in progress in #5 .
Currently
For each record, the application makes one instances for each class (F22, F14, F15, F40, etc.). Here is the relevant part of the code.
My proposal
For the Group 1 this is ok: the record actually describe them.
For Groups 2 and 3, the source xml does not contains the information all the time. It could give us the info about the 1st execution, the 1st execution in France and the 1st publication, but not all the times. My proposal is to make them "optional" and include them in the output only if we have the info.
For Group 4, for sure we will add an instance, but for notices d'oeuvre we could have multiple instances - see #17.