Phase 1 Gaps

[dostiep]

1. Not all datasets needed for the study are represented in the USDM + BCs + DSSs **(API to develop)**

2. Not all dataset variables for the study are available via the USDM + BCs + DSSs **(API to develop)**

3. Attributes like Key Sequence are not available via the USDM + BCs + DSSs **(added placeholders)**

4. For required variables added via the Library, there is no metadata describing the subset of a codelist relevant for the study **(we've added placeholders instead of all values from the codelists)**

5. Attributes like Repeating are not available via the USDM + BCs + DSSs **(method from generator was created to address this)**

6. The dataset Structure is not available via the USDM + BCs + DSSs, but can be found via a Library API call **(done)**

7. No way to determine what Value Level Metadata is relevant to a study

8. We will need a best practice for how a sponsor can introduce their own BCs and DSSs into the define.xml generation process **(API to develop)**

9. Attributes like ReferenceData are not available via the USDM + BCs + DSSs, but an algorithm can be created to categorize them effectively **(done)**

10. When generating shell datasets for ADaM, we need a method to determine when a variable with a datatype of integer is representing a datetime. **(later)**

11. Mining CRFs is viewed as one source of additional study metadata, but many sponsors and vendors do not have ODM-based CRFs available **(later)**

12. Attributes like Code list datatype are not available via USDM + BCs + DSSs **(actually defaulted to text)**

13. Attributes like Mandatory are not available via the USDM + BCs + DSSs **(Fixed: Mandatory is derived from 'Req' variables from CDISC Library)**

14. Domains like RS have multiple codelists attached to a single variable - such as RSCAT's "ONCRSCAT", "CCCAT" and depending on study requirements, 
RSTESTCD will have multiple codelists "ONCRTSCD", "ECOG1TC" as well. USDM + BCs + DSSs don't provide all the metadata needed to support this. **(Need to test as it should be addressed with DSS. We need to add RS to a USDM JSON file )**

15. Some BCs have multiple DSSs and no way to determine which DSS should be used **(True if BC is referenced in the USDM file instead of a DSS)**

16. Dictionaries and Methods have not been implemented and are not provided by the USDM + BCs + DSSs

17. Some codelist terms have multiple decodes options, and no way to determine which to use **(We currently use the first value if multiple ones exist)**

18. Overall, many of the detailed gaps point to the fact that there be a missing model needed to drive the process. Something like a Data Management Plan model that carries operationally focused metadata needed for define.xml, ODM CRFs, etc. For example:
    a. Dictionaries and their versions
    b. Standard IG used and its version
    c. CT used and its version
    d. SDTM methods (maybe)