GitLab

  -debug <level> : turn on debug messages at <level>
  -no_cleanup    : don't perform global_cleanup when worker exits

better debug system to allow worker to control analysis debug level
new batch_size method will use avg_msec_per_job to calculate batch_size on the fly
  if analysis_stat.batch_size==0

keep analysis_job.input_id as varchar(255) to allow UNIQUE(analysis_id,input_id)
but in adaptor added logic so that if input_id in AnalysisJob object exceeds
the 255 char limit to store/fetch from the analysis_data table.  The input_id
in the analysis_job table becomes '_ext_input_analysis_data_id ##' which is a unique
internal variable to trigger the fetch routine to know to get the 'real' input_id
from the analysis_data table.
NO MORE 255 char limit on input_id and completely transparent to API user.

if the hash keys/values are the same.

to AnalysisStats object and store/fetch of avg_msec_per_job from analysis_stats table.
Corresponds to schema change.  Old hive databases need to do this sql to bring them up to spec;
  alter table analysis_stats add column avg_msec_per_job int(10) default 0 NOT NULL after batch_size;
  alter table analysis_stats modify column last_update datetime NOT NULL;

  analysis_stats.avg_msec_per_job int
modified
  analysis_stats.last_update to a datetime so that it is only changed when the API tells it to.

In run_autonomously method, updated a queen method call to print_running_worker_status. It was wrongly print_worker_status

to 'LOADING' to trigger sync so system knows that something changed

  -failed_jobs           : show all failed jobs
  -reset_job_id <num>    : reset a job back to READY so it can be rerun
  -reset_all_jobs_for_analysis_id <num>

added method reset_all_jobs_for_analysis_id to facilitate re-flowing data through new dataflow rules.
extended perldoc
changed the 'retry count' to 7 (so runs 1 + 7 retrys)

thus allowing the RunnableDBs to do tracking of output data to the specific analysis_job_id

Initially used to manually re-run a job with runWorker.pl -job_id

and Analysis::RunnableDB superclasses

properly handle RunnableDBs that throw exceptions in the fetch_input stage.

changed INSERT syntax to be more SQL compliant

changed INSERT syntax to be more SQL compliant

changed INSERT syntax to be more SQL compliant

Set the OUTPUT_AUTOFLUSH=1 to get info immediately when in -loop mode. Also print out the time when the next loop will occur

In synchronize_AnalysisStats method, added a POSIX::ceil when setting the num_required_workers for an AnalysisStats object

used to always store a full URL

in AnalysisStatsAdaptor::fetch_by_analysis_id

modified  Bio::EnsEMBL::Analysis::stats to not do any exception catching
if  AnalysisStatsAdaptor->fetch_by_analysis_id fails there is something
very wrong and the exception should propogate out and cause the program to fail

Bio::EnsEMBL::Analysis::RunnableDB via namespace extension syntax
so that hive system can use analysis.modules that inherit from Bio::EnsEMBL::Analysis::RunnableDB