Allow to find a runfolder in 'analysis' or 'outgoing'
in the presence of a duplicate runfolder in 'incoming'.

The duplicates in 'incoming' are sometimes created by
instruments well after the run was mirrored and the
runfolder moved to 'analysis'.

Recently, the default LIMS driver type in samplesheet
generation was changed to 'ml_warehouse'. Following
this change, the code for auto-generation of MiSeq
default samplesheets is now simplified to exclude
an explicit creation of LIMS objects.

Ability to disregard duplicate runfolders.

Simplified samplesheet auto-generation code.

* Samplesheet generation tests - move to mlwh driver.

Previously the tests uses XML feeds as a source of LIMS
data. All production code is using mlwh database as a
source of LIMS data. Amend tests to use mlwh data and
the default ml_warehouse lims driver.

Amended how an undefined sample public name is dealt with,
i.e. converted an error to a warning and assigned a sensible
fall-back value.

Some samplesheets are generated without names of study owners,
followers, etc. This is due to the fact that some studies
do not have fixtures for user data.

The database fixtures were generated from the current data,
this explains multiple differences between the old and new
content of the generated samplesheets.

In some tests newer production batches were used instead of
hand-crafted data.

Deleted tests for old scenarios that do not exist in production
any longer.

Deleted tests that are duplicated by similar tests with more
contemporary test data.

* Deleted now redundant XML test fixtures

Use correct name instrument name pattern in docs.

* Deleted the xml driver for st::api::lims.

* Simplified code for insert size and restored is tests.

* Restored previously available tests.

Used existing test fixtures, which had to be changed
in a few places.

* Improved and extended basic tests for tag zero

* Extended tests for study and sample properties

Deleted low-level LIMS XML API

Remove lims xml fixtures

Changes are due to the MySQL database server upgrade to v8.+

Updated DBIx classes from the prod database.

st::api::lims - drop deprecated methods

Dropped unused constants

Dropped project_id, project_name, request_id accessors
from st::api::lims since this data is not represented
in ml warehouse. project_ids method is dropped as well.

Dropped lims accessors for which no data exist.

In practice we never automatically merged rapid runs,
this flag was reverted to false for them by the pipeline.
Instruments on which rapid runs were performed are no
longer in use.

Restricted automatically meargeable to NovaSeq.

This was used for Tradis runs when no LIMS support
was available. Was not needed for the last 7 years.

Deleted the code for inline indexes.

#761)

* Removed fallback to study ref genome for complex tag zero and lane-level objects

* Fixed a typo in the name of a variable

* Corrected the scope of the variable

Record all recent changes

Update devel with bugfix release from master

Limit DRAGEN analysis to human samples only.

Moved tests for aggregation across lanes to a new test
scripts. Tests for a new merge by library method will
be added to this new test script.

Also refactored the method itself.

We use the name of the Application in RunParameters.xml
to detect that the instrument belongs to the NovaSeq
type. Newer NovaSeqX instrument we have has the following
definition for the application:
  <Application>NovaSeqXSeries Control Software</Application>
This pull request fixes the reg exp we use to ensure that
the above definition does not result in the instrument being
classified as NovaSeq.

Distinguish between NovaSeq and NovaSeqX RunParams.

Ensure insert size is returned for compositions.

Added a method for aggregation by library

Added records for undocumented changed.

Fixed incorrect string comparison

Name variables consistently.
Source dependencies from the branch that is the target.

Correct propagation of the CI target branch.

prep release 98.0.0