GitLab

Rework on how Bio::DB::HTS is used in HTSAdaptor

Add VCF4 format for validation

as spaces can appear within a field which means it is not safe to use space
as delimiter. Now I found a way to make tab delimiter work for me so the space
is no longer needed anyway.

Previoiusly because of the lack of VCF4 format definition, the user upload of VCF
type data (stored in text) could not be validated. This change is intended to add
this missing information so that those VCF records could be validated and parsed.

Also add space as a valid delimiter in addition to tab. I know this is violating
the VCF4.x spece. But I could not get tab delimited VCF record to work. I don't
know why.

Related to ENSEMBL-5050.

based on Mag's review comments at this pull request:
https://github.com/Ensembl/ensembl-io/pull/55

The Bio::DB::HTS object and other relatedHTS objects will be created
only once within a single HTSAdaptor. They will have only one copy and
be stored in $self->{_cache}->{_htsobj_handle}. Every use of HTS objects
should through this copy.

Also remove "close" calls to the HTS objects as these objects will
handle their resources properly when exiting. The DESTROY methods to
these objects has been added in
https://github.com/Ensembl/Bio-DB-HTS/pull/50
which will release the resources where needed.

This will provide a uniform interface to HTS objects, making it more
consistent. It prevents from creating multiple objects for the same
file and hence more resouces efficient too.

Also rename some variables to make them consistent throughout.

Related to ENSEMBL-5035.

* postreleasefix/89:
  File might not have a header

* postreleasefix/89:
  Configure Vcf module correctly

* postreleasefix/89:
  Put VCF indices into timestamped directories

* postreleasefix/89:
  Write tabix indices to their own directory

* postreleasefix/87:
  Check if $max is initialised

* postreleasefix/87:
  New method, copied in from old adaptor
  Extend methods to deal with whole chromosomes

Feature support upgrade for Genbank parser. Processing of coordinates delegated to calling code because of complexity and biological comprehension required.

* postreleasefix/86:
  Fix validation of optional fields

Improvements to munge_chr_id subroutine in HTSAdaptor