Deal with unexpected problems in data when calculating statistics, and fix some bugs

Author: apdavison

api/dashboard/js/cumulative-user-count.js

@@ -51,7 +51,7 @@ d3.json("/statistics/cumulative-user-count", function (error, data) {
 });
 
 d3.json(
-  "/statistics/cumulative-user-count?platform=BrainScaleS",
+  "/statistics/cumulative-user-count?hardware_platform=BrainScaleS",
   function (error, data) {
     if (error) return console.warn(error);
 
@@ -104,7 +104,7 @@ d3.json(
 );
 
 d3.json(
-  "/statistics/cumulative-user-count?platform=BrainScaleS-2",
+  "/statistics/cumulative-user-count?hardware_platform=BrainScaleS-2",
   function (error, data) {
     if (error) return console.warn(error);
 
@@ -157,7 +157,7 @@ d3.json(
 );
 
 d3.json(
-  "/statistics/cumulative-user-count?platform=SpiNNaker",
+  "/statistics/cumulative-user-count?hardware_platform=SpiNNaker",
   function (error, data) {
     if (error) return console.warn(error);
 
@@ -210,7 +210,7 @@ d3.json(
 );
 
 d3.json(
-  "/statistics/cumulative-user-count?platform=Spikey",
+  "/statistics/cumulative-user-count?hardware_platform=Spikey",
   function (error, data) {
     if (error) return console.warn(error);
 

api/simqueue/db.py

@@ -291,7 +291,8 @@ async def update_log(job_id, log, append=False):
 
 
 def get_list_filter(attr, value):
-    if len(value) > 0:
+    assert isinstance(value, str)
+    if len(value) > 1:
         return attr.in_(value)
     else:
         return attr == value[0]

api/simqueue/resources/statistics.py

@@ -58,6 +58,7 @@ async def job_count(start: date = None, end: date = None, interval: int = 7):
             [
                 (timestamp["timestamp_completion"].date() - start).days
                 for timestamp in completion_timestamps
+                if timestamp["timestamp_completion"]
             ],
             dtype=int,
         )
@@ -265,7 +266,11 @@ async def resource_usage(start: date = None, end: date = None, interval: int = 7
             fields=["timestamp_completion", "resource_usage"],
         )
         completed = np.array(
-            [(job["timestamp_completion"].date() - start).days for job in completed_jobs],
+            [
+                (job["timestamp_completion"].date() - start).days
+                for job in completed_jobs
+                if job["timestamp_completion"]
+            ],
             dtype=int,
         )
         usage_per_job = np.array([job["resource_usage"] for job in completed_jobs])