Skip to content

Commit

Permalink
added error handling
Browse files Browse the repository at this point in the history
  • Loading branch information
@kasprovav
kasprovav committed Nov 14, 2024
1 parent 6114c4b commit 8212dbb
Show file tree
Hide file tree
Showing 2 changed files with 113 additions and 77 deletions.
86 changes: 54 additions & 32 deletions ai_ta_backend/database/sql.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -159,42 +159,64 @@ def getPreAssignedAPIKeys(self, email: str):
return self.supabase_client.table("pre_authorized_api_keys").select("*").contains("emails", '["' + email + '"]').execute()

def getConversationsCreatedAtByCourse(self, course_name: str):
count_response = self.supabase_client.table("llm-convo-monitor")\
.select("created_at", count="exact")\
.eq("course_name", course_name)\
.execute()

total_count = count_response.count

if total_count <= 0:
return [], 0

all_data = []
batch_size = 1000
start = 0

while start < total_count:
end = min(start + batch_size - 1, total_count - 1)

response = self.supabase_client.table("llm-convo-monitor")\
.select("created_at")\
try:
count_response = self.supabase_client.table("llm-convo-monitor")\
.select("created_at", count="exact")\
.eq("course_name", course_name)\
.range(start, end)\
.execute()

if not response.data:
print(f"No data returned for range {start} to {end}.")
break

all_data.extend(response.data)
start += batch_size

return all_data, total_count

total_count = count_response.count if hasattr(count_response, 'count') else 0

if total_count <= 0:
print(f"No conversations found for course: {course_name}")
return [], 0

all_data = []
batch_size = 1000
start = 0

while start < total_count:
end = min(start + batch_size - 1, total_count - 1)

try:
response = self.supabase_client.table("llm-convo-monitor")\
.select("created_at")\
.eq("course_name", course_name)\
.range(start, end)\
.execute()

if not response or not hasattr(response, 'data') or not response.data:
print(f"No data returned for range {start} to {end}.")
break

all_data.extend(response.data)
start += batch_size

except Exception as batch_error:
print(f"Error fetching batch {start}-{end}: {str(batch_error)}")
continue

if not all_data:
print(f"No conversation data could be retrieved for course: {course_name}")
return [], 0

return all_data, len(all_data)

except Exception as e:
print(f"Error in getConversationsCreatedAtByCourse for {course_name}: {str(e)}")
return [], 0

def getProjectStats(self, project_name: str):
response = self.supabase_client.table("project_stats").select("total_messages, total_conversations, unique_users")\
.eq("project_name", project_name).execute()
try:
response = self.supabase_client.table("project_stats").select("total_messages, total_conversations, unique_users")\
.eq("project_name", project_name).execute()

if response and hasattr(response, 'data') and response.data:
return response.data[0]
except Exception as e:
print(f"Error fetching project stats for {project_name}: {str(e)}")

return response.data[0] if response.data else {"total_messages": 0, "total_conversations": 0, "unique_users": 0}
# Return default values if anything fails
return {"total_messages": 0, "total_conversations": 0, "unique_users": 0}


104 changes: 59 additions & 45 deletions ai_ta_backend/service/retrieval_service.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -526,54 +526,68 @@ def format_for_json(self, found_docs: List[Document]) -> List[Dict]:
def getConversationStats(self, course_name: str):
"""
Fetches conversation data from the database and groups them by day, hour, and weekday.
Args:
course_name (str)
Returns:
dict: A dictionary containing:
- 'per_day': Counts of conversations by date (YYYY-MM-DD).
- 'per_hour': Counts of conversations by hour (0-23).
- 'per_weekday': Counts of conversations by weekday (Monday-Sunday).
- 'heatmap': A nested dictionary for heatmap data (days of the week as keys, hours as inner keys).
"""
# print(f"Received course_name: {course_name}")

conversations, total_count = self.sqlDb.getConversationsCreatedAtByCourse(course_name)

central_tz = pytz.timezone('America/Chicago')

grouped_data = {
'per_day': defaultdict(int),
'per_hour': defaultdict(int),
'per_weekday': defaultdict(int),
'heatmap': defaultdict(lambda: defaultdict(int)),
}
try:
conversations, total_count = self.sqlDb.getConversationsCreatedAtByCourse(course_name)

# Initialize with empty data (all zeros)
response_data = {
'per_day': {},
'per_hour': {str(hour): 0 for hour in range(24)}, # Convert hour to string for consistency
'per_weekday': {day: 0 for day in ['Monday', 'Tuesday', 'Wednesday', 'Thursday', 'Friday', 'Saturday', 'Sunday']},
'heatmap': {day: {str(hour): 0 for hour in range(24)} for day in ['Monday', 'Tuesday', 'Wednesday', 'Thursday', 'Friday', 'Saturday', 'Sunday']},
'total_count': 0
}

if not conversations:
return response_data

central_tz = pytz.timezone('America/Chicago')
grouped_data = {
'per_day': defaultdict(int),
'per_hour': defaultdict(int),
'per_weekday': defaultdict(int),
'heatmap': defaultdict(lambda: defaultdict(int)),
}

if conversations:
for record in conversations:
created_at = record['created_at']
parsed_date = parser.parse(created_at).astimezone(central_tz)

day = parsed_date.date()
hour = parsed_date.hour
day_of_week = parsed_date.strftime('%A')

grouped_data['per_day'][str(day)] += 1
grouped_data['per_hour'][hour] += 1
grouped_data['per_weekday'][day_of_week] += 1
grouped_data['heatmap'][day_of_week][hour] += 1

else:
print("No valid response data. Check if the query is correct or if the response is empty.")
return {}

return {
'per_day': dict(grouped_data['per_day']),
'per_hour': dict(grouped_data['per_hour']),
'per_weekday': dict(grouped_data['per_weekday']),
'heatmap': {day: dict(hours) for day, hours in grouped_data['heatmap'].items()},
}
try:
created_at = record['created_at']
parsed_date = parser.parse(created_at).astimezone(central_tz)

day = parsed_date.date()
hour = parsed_date.hour
day_of_week = parsed_date.strftime('%A')

grouped_data['per_day'][str(day)] += 1
grouped_data['per_hour'][str(hour)] += 1 # Convert hour to string
grouped_data['per_weekday'][day_of_week] += 1
grouped_data['heatmap'][day_of_week][str(hour)] += 1 # Convert hour to string
except Exception as e:
print(f"Error processing record: {str(e)}")
continue

return {
'per_day': dict(grouped_data['per_day']),
'per_hour': {str(k): v for k, v in grouped_data['per_hour'].items()},
'per_weekday': dict(grouped_data['per_weekday']),
'heatmap': {day: {str(h): count for h, count in hours.items()}
for day, hours in grouped_data['heatmap'].items()},
'total_count': total_count
}

except Exception as e:
print(f"Error in getConversationStats for course {course_name}: {str(e)}")
self.sentry.capture_exception(e)
# Return empty data structure on error
return {
'per_day': {},
'per_hour': {str(hour): 0 for hour in range(24)},
'per_weekday': {day: 0 for day in ['Monday', 'Tuesday', 'Wednesday', 'Thursday', 'Friday', 'Saturday', 'Sunday']},
'heatmap': {day: {str(hour): 0 for hour in range(24)}
for day in ['Monday', 'Tuesday', 'Wednesday', 'Thursday', 'Friday', 'Saturday', 'Sunday']},
'total_count': 0
}

def getProjectStats(self, project_name: str) -> Dict[str, int]:
"""
Expand Down

0 comments on commit 8212dbb

Please sign in to comment.