Rewrite of get_gambl_metadata and new helper functions

[rdmorin]

The main changes are:

• get_gambl_metadata replaces og_get_gambl_metadata functionality but allows backwards compatibility. This function does a variety of checks for consistency in the metadata/biopsy tables and reports issues it finds.
• check_gene_expression ensures RNA-seq is prioritized the way we want and always deduplicated
• get_gene_expression now relies on both the new functions and their new features

exome_priority_meta = get_gambl_metadata(capture_protocol_priority = "Exome")
2908 capture samples are missing a value for protocol. Assuming Exome.                                                                                 
55 biopsies are missing from the biopsy metadata. This should be fixed!                                                                                
affected cohorts:  DLBCL_Weng,DLBCL_Jain
27 biopsies with discrepancies in the pathology field. This should be fixed!
10 biopsies with discrepancies in the time_point field. This should be fixed!

> table(exome_priority_meta$protocol)

      Capture      chromium  ctDNA_Genome         Exome        genome        Genome         polyA         PolyA 
           34             3             5          3213            68          1854            11          1108 
Ribodepletion 
         1235 

not_exome_priority_meta = get_gambl_metadata(capture_protocol_priority = "Capture")
2908 capture samples are missing a value for protocol. Assuming Exome.                                                                                 
55 biopsies are missing from the biopsy metadata. This should be fixed!                                                                                
affected cohorts:  DLBCL_Weng,DLBCL_Jain
27 biopsies with discrepancies in the pathology field. This should be fixed!
10 biopsies with discrepancies in the time_point field. This should be fixed!

> table(not_exome_priority_meta$protocol)

      Capture      chromium  ctDNA_Genome         Exome        genome        Genome         polyA         PolyA 
           52             3             5          3195            68          1854            11          1108 
Ribodepletion 
         1235 

[rdmorin]

Would someone be able to test this again now that I've patched a few more bugs? It was a bit more involved to implement properly but it looks like the format (long/wide) is working as expected.