Handle phenopackets with unparsable data

[ielis]

We need a strategy for handling phenopackets that do not include minimal required data for genotype-phenotype correlation analysis.

For instance, a phenopacket PMID_36446582_Novara2017_P2 does not include CNV coordinates, and ends up having no parsable variant.

Other fallible actions include unknown HPO term as of the used ontology.

Requirements

• errors are reported with phenopacket ID
• ideally all errors are reported, not just the first one
• phenopacket is not included in the analysis

Suggestion

• define a new exception, e.g. GenoPhenoParseException(BaseException) which can be thrown by PatientCreator. This enhances the parsing workflow and informs the user that the process is fallible. The exception has fields for sample ID and a list of specific issues.
• revise PhenopacketPatientCreator such that create_patient function keeps track of the issues, such as unparsable variant, invalid HPO term, ...
• if we encounter any issues, we pack them into the exception and raise it to indicate a failure

[lnrekerle]

This was actually part of my "update notebooks" PR that I just pushed. Forgot to add this task as one that would close with it. This is complete.