fix >100% in campaign-status; more precise ESA time computation

Campaign/views.py

@@ -264,10 +264,15 @@ def campaign_status_esa(campaign) -> str:
     """
     out_str += f"<h1>{campaign.campaignName}</h1>\n"
     out_str += "<table>\n"
-    out_str += "<tr>" + "".join(
-        f"<th>{x}</th>" for x in ["Username", "Progress", "First Modified", "Last Modified", "Time (Last-First)", "Time (Real)"]
-    ) + "</tr>\n"
-
+    out_str += """<tr>
+<th>Username</th>
+<th>Progress</th>
+<th>First Modified</th>
+<th>Last Modified</th>
+<th style="cursor: pointer" title="Very coarse upper bound estimate between the last and the first interaction with the system.">Time (Coarse ❔)</th>
+<th style="cursor: pointer" title="Sum of times between any two interactions that are not longer than 10 minutes.">Time (Real ❔)</th>
+</tr>\n
+""" 
     for team in campaign.teams.all():
         for user in team.members.all():
             if user.is_staff:
@@ -298,6 +303,7 @@ def campaign_status_esa(campaign) -> str:
             _data = DirectAssessmentDocumentResult.objects.filter(
                 createdBy=user, completed=True, task__campaign=campaign.id
             )
+            _data_uniq_len = len({item.id for item in _data})
 
             # If no data, show 0 progress or show that no task is assigned
             if not _data:
@@ -331,11 +337,11 @@ def campaign_status_esa(campaign) -> str:
                     continue
 
                 total_count = task.items.count()
-                if total_count == len(_data):
+                if total_count == _data_uniq_len:
                     out_str += f"<td>{user.username} ✅</td>"
                 else:
                     out_str += f"<td>{user.username} 🛠️</td>"
-                out_str += f"<td>{len(_data)}/{total_count} ({len(_data) / total_count:.0%})</td>"
+                out_str += f"<td>{_data_uniq_len}/{total_count} ({_data_uniq_len / total_count:.0%})</td>"
                 first_modified = min([x.start_time for x in _data])
                 last_modified = max([x.end_time for x in _data])
 
@@ -351,15 +357,9 @@ def campaign_status_esa(campaign) -> str:
                 annotation_time_upper = f'{int(floor(annotation_time_upper / 3600)):0>2d}h {int(floor((annotation_time_upper % 3600) / 60)):0>2d}m'
                 out_str += f"<td>{annotation_time_upper}</td>"
 
-                times = collections.defaultdict(list)
-                for item in _data:
-                    times[(item.item.documentID, item.item.targetID)].append((item.start_time, item.end_time))
-                times = [
-                    (min([x[0] for x in doc_v]), max([x[1] for x in doc_v]))
-                    for doc, doc_v in times.items()
-                ]
-
-                annotation_time = sum([b-a for a, b in times])
+                # consider time that's in any action within 10 minutes
+                times = sorted([item.start_time for item in _data] + [item.end_time for item in _data])
+                annotation_time = sum([b-a for a, b in zip(times, times[1:]) if (b-a) < 10*60])
                 annotation_time = f'{int(floor(annotation_time / 3600)):0>2d}h {int(floor((annotation_time % 3600) / 60)):0>2d}m'
 
                 out_str += f"<td>{annotation_time}</td>"

EvalData/models/direct_assessment_document.py

@@ -599,35 +599,25 @@ def get_hit_status_for_user(cls, user):
     @classmethod
     def get_time_for_user(cls, user):
         results = cls.objects.filter(createdBy=user, activated=False, completed=True)
+        campaign_opts = result.task.campaign.campaignOptions.lower().split(";")
         is_esa_or_mqm = any(
             [
-                "esa" in result.task.campaign.campaignOptions.lower().split(";")
-                or "mqm" in result.task.campaign.campaignOptions.lower().split(";")
+                "esa" in campaign_opts or "mqm" in campaign_opts
                 for result in results
             ]
         )
 
         if is_esa_or_mqm:
-            # for ESA or MQM, do minimum and maximum from each doc
-            import collections
-
-            timestamps = collections.defaultdict(list)
-            for result in results:
-                timestamps[
-                    result.item.documentID + " ||| " + result.item.targetID
-                ].append((result.start_time, result.end_time))
-
-            # timestamps are document-level now, but that does not change anything later on
-            timestamps = [
-                (min([x[0] for x in doc_v]), max([x[1] for x in doc_v]))
-                for doc, doc_v in timestamps.items()
-            ]
+            # consider time that's in any action within 10 minutes
+            times = sorted([item.start_time for item in results] + [item.end_time for item in results])
+            annotation_time = sum([b-a for a, b in zip(times, times[1:]) if (b-a) < 10*60])
+            return seconds_to_timedelta(annotation_time)
         else:
             timestamps = []
             for result in results:
                 timestamps.append((result.start_time, result.end_time))
 
-        return seconds_to_timedelta(_compute_user_total_annotation_time(timestamps))
+            return seconds_to_timedelta(_compute_user_total_annotation_time(timestamps))
 
     @classmethod
     def get_system_annotations(cls):