Add get_queue_info helper function

tpv/core/helpers.py

@@ -8,6 +8,8 @@
 import random
 from functools import reduce
 from galaxy import model
+from sqlalchemy import func, text, Float
+
 
 GIGABYTES = 1024.0**3
 
@@ -125,3 +127,42 @@ def get_dataset_attributes(datasets):
             'size': get_dataset_size(i.dataset.dataset)}
         for i in datasets or {}
     }
+
+def get_queue_info(app):
+    # query the Galaxy DB and return a dictionary of sums of cores/mem/gpus
+    # for queued and running jobs. This requires 'tpv_mem', 'tpv_cores' and
+    # 'tpv_gpus' params to be stored for all non-terminal jobs. These params
+    # can be inherited from the default tool in the TPV shared database.
+    query = app.model.context.query(
+        model.Job.destination_id,
+        model.Job.state,
+        func.count(model.Job.id).label('job_count'),
+        func.sum(
+            func.cast(text("encode(destination_params, 'escape')::json ->> 'tpv_cores'"), Float)
+        ).label('sum_cores'),
+        func.sum(
+            func.cast(text("encode(destination_params, 'escape')::json ->> 'tpv_mem'"), Float)
+        ).label('sum_mem'),
+        func.sum(
+            func.cast(text("encode(destination_params, 'escape')::json ->> 'tpv_gpus'"), Float)
+        ).label('sum_gpus'),
+    ).filter(
+        model.Job.state.in_(['queued', 'running'])
+    ).group_by(
+        model.Job.destination_id,
+        model.Job.state
+    )
+    results = query.all()
+
+    db_queue_info = {}
+    for row in results:
+        destination_id, state, job_count, sum_cores, sum_mem, sum_gpus = row
+        if not destination_id in db_queue_info:
+            db_queue_info[destination_id] = {}
+        db_queue_info[destination_id][state] = {
+            "sum_cores": sum_cores,
+            "sum_mem": sum_mem,
+            "sum_gpus": sum_gpus,
+            "job_count": job_count,
+        }
+    return db_queue_info