Better error handling when one of the pileup fails

[teng-gao]

In multi-sample mode of pileup_and_phase.R, pileup produces empty VCF in one of the samples (because the BAM is truncated), but script was not interrupted. Phasing did not happen and gives below error. When run on single-sample mode the script correctly stops.

character(0)
Creating VCFs
Error in `mutate()`:
ℹ In argument: `snp_index = 1:n()`.
Caused by error:
! `snp_index` must be size 0 or 1, not 2.
Backtrace:
   ▆
 1. ├─numbat:::genotype(...)
 2. │ ├─... %>% arrange(CHROM, POS)
 3. │ └─base::lapply(...)
 4. │  └─numbat (local) FUN(X[[i]], ...)
 5. │   └─numbat:::get_snps(vcf)
 6. │    └─... %>% filter(!is.na(AR))
 7. ├─dplyr::arrange(., CHROM, POS)
 8. ├─dplyr::mutate(., AR = AD/DP)
 9. ├─dplyr::summarise(...)
 10. ├─dplyr::group_by(., CHROM, POS, REF, ALT, snp_id)
 11. ├─dplyr::bind_rows(.)
 12. │ └─rlang::list2(...)
 13. ├─dplyr::filter(., !is.na(AR))
 14. ├─dplyr::mutate(., AR = AD/DP)
 15. ├─dplyr::mutate(., snp_index = 1:n())
 16. ├─dplyr:::mutate.data.frame(., snp_index = 1:n())
 17. │ └─dplyr:::mutate_cols(.data, dplyr_quosures(...), by)
 18. │  ├─base::withCallingHandlers(...)
 19. │  └─dplyr:::mutate_col(dots[[i]], data, mask, new_columns)
 20. │   └─mask$eval_all_mutate(quo)
 21. │    └─dplyr (local) eval()
 22. ├─dplyr:::dplyr_internal_error(...)
 23. │ └─rlang::abort(class = c(class, "dplyr:::internal_error"), dplyr_error_data = data)
 24. │  └─rlang:::signal_abort(cnd, .file)
 25. │   └─base::signalCondition(cnd)
 26. └─dplyr (local) `<fn>`(`<dpl:::__>`)
 27.  └─rlang::abort(message, class = error_class, parent = parent, call = error_call)
Execution halted