Skip to content

A
aisbf
Commit 6130a6db authored by Stefy Lanza (nextime / spora )'s avatar Stefy Lanza (nextime / spora )
Browse files
Options

Fix analytics

parent 006441ea
Hide whitespace changes
Inline Side-by-side
Showing with 78 additions and 55 deletions
aisbf/analytics.py
View file @ 6130a6db
......@@ -562,6 +562,34 @@ class Analytics:
)
return total_cost
def _normalize_time_window(
self,
time_range: Optional[str] = None,
from_datetime: Optional[datetime] = None,
to_datetime: Optional[datetime] = None
) -> tuple[datetime, datetime, str]:
now = datetime.now()
if from_datetime and to_datetime:
normalized_range = 'custom' if time_range not in ['yesterday'] else time_range
return from_datetime, to_datetime, normalized_range
if time_range == '1h':
return now - timedelta(hours=1), now, '1h'
if time_range == '6h':
return now - timedelta(hours=6), now, '6h'
if time_range == 'yesterday':
today = now.replace(hour=0, minute=0, second=0, microsecond=0)
return today - timedelta(days=1), today - timedelta(microseconds=1), 'yesterday'
if time_range == '7d':
return now - timedelta(days=7), now, '7d'
if time_range == '30d':
return now - timedelta(days=30), now, '30d'
if time_range == '90d':
return now - timedelta(days=90), now, '90d'
return now - timedelta(hours=24), now, '24h'
def get_token_usage_by_date_range(
self,
provider_id: Optional[str] = None,
......@@ -584,8 +612,7 @@ class Analytics:
Returns:
Dictionary with token counts and cost estimates
"""
start = from_datetime or (datetime.now() - timedelta(days=1))
end = to_datetime or datetime.now()
start, end, _ = self._normalize_time_window('custom' if from_datetime or to_datetime else '24h', from_datetime, to_datetime)
with self.db._get_connection() as conn:
cursor = conn.cursor()
......@@ -686,10 +713,7 @@ class Analytics:
"""
Get statistics for all providers.
"""
# Use local time to match how MySQL CURRENT_TIMESTAMP stores data
now_local = datetime.now()
start = (from_datetime if from_datetime else (now_local - timedelta(days=1)))
end = (to_datetime if to_datetime else now_local)
start, end, _ = self._normalize_time_window('custom' if from_datetime or to_datetime else '24h', from_datetime, to_datetime)
with self.db._get_connection() as conn:
cursor = conn.cursor()
......@@ -824,10 +848,10 @@ class Analytics:
Returns:
List of time-series data points
"""
# Determine time range
if time_range == 'custom' and from_datetime and to_datetime:
cutoff = from_datetime
end_time = to_datetime
cutoff, end_time, normalized_time_range = self._normalize_time_window(time_range, from_datetime, to_datetime)
# Determine time range bucket size
if normalized_time_range == 'custom':
# Calculate bucket size based on range
total_minutes = (end_time - cutoff).total_seconds() / 60
if total_minutes <= 60:
......@@ -842,39 +866,21 @@ class Analytics:
bucket_minutes = 60 * 24 # daily
else: # > 30 days
bucket_minutes = 60 * 24 * 7 # weekly
elif time_range == '1h':
cutoff = datetime.now() - timedelta(hours=1)
end_time = datetime.now()
elif normalized_time_range == '1h':
bucket_minutes = 5
elif time_range == '6h':
cutoff = datetime.now() - timedelta(hours=6)
end_time = datetime.now()
elif normalized_time_range == '6h':
bucket_minutes = 15
elif time_range == '24h':
cutoff = datetime.now() - timedelta(hours=24)
end_time = datetime.now()
elif normalized_time_range == '24h':
bucket_minutes = 30
elif time_range == 'yesterday':
# Yesterday: from 00:00:00 to 23:59:59 of previous day
today = datetime.now().replace(hour=0, minute=0, second=0, microsecond=0)
cutoff = today - timedelta(days=1)
end_time = today - timedelta(microseconds=1)
elif normalized_time_range == 'yesterday':
bucket_minutes = 30
elif time_range == '7d':
cutoff = datetime.now() - timedelta(days=7)
end_time = datetime.now()
elif normalized_time_range == '7d':
bucket_minutes = 60 * 24 # Daily
elif time_range == '30d':
cutoff = datetime.now() - timedelta(days=30)
end_time = datetime.now()
elif normalized_time_range == '30d':
bucket_minutes = 60 * 24 # Daily
elif time_range == '90d':
cutoff = datetime.now() - timedelta(days=90)
end_time = datetime.now()
elif normalized_time_range == '90d':
bucket_minutes = 60 * 24 # Daily
else: # Default 24h
cutoff = datetime.now() - timedelta(hours=24)
end_time = datetime.now()
bucket_minutes = 30
# Query database for token usage in time range
......@@ -1228,8 +1234,7 @@ class Analytics:
rotation_filter: Optional[str] = None,
autoselect_filter: Optional[str] = None
) -> Dict[str, Any]:
start = from_datetime or (datetime.now() - timedelta(days=1))
end = to_datetime or datetime.now()
start, end, _ = self._normalize_time_window('custom' if from_datetime or to_datetime else '24h', from_datetime, to_datetime)
provider_stats = self.get_all_providers_stats(
from_datetime,
......@@ -1266,6 +1271,9 @@ class Analytics:
if not tier_info:
continue
if tier_info.get('source') == 'default':
continue
cached_usage = None
try:
cached_row = self.db.get_provider_usage(None if user_filter == -1 else user_filter, provider_id)
......@@ -1275,19 +1283,22 @@ class Analytics:
cached_usage = None
runtime_quota = self._derive_quota_from_usage(provider_id, cached_usage)
free_limit = max(int((runtime_quota or {}).get('limit') or tier_info.get('limit', 0) or 0), 1)
free_limit = int((runtime_quota or {}).get('limit') or tier_info.get('limit', 0) or 0)
limit_type = (runtime_quota or {}).get('limit_type') or tier_info.get('limit_type') or 'requests'
period = (runtime_quota or {}).get('period') or tier_info.get('period') or 'month'
if free_limit <= 0:
continue
if limit_type == 'tokens':
usage_amount = usage.get('tokens_used', 0) or 0
else:
usage_amount = (runtime_quota or {}).get('used') or usage.get('request_count', 0) or 0
usage_amount = usage.get('request_count', 0) or 0
if usage_amount <= free_limit:
continue
extra_free_tiers = usage_amount // free_limit
extra_free_tiers = max((usage_amount - 1) // free_limit, 0)
if extra_free_tiers <= 0:
continue
......@@ -1352,9 +1363,7 @@ class Analytics:
Returns:
Dictionary with cost estimates
"""
# Use date range for token usage if specified
start = from_datetime or (datetime.now() - timedelta(days=1))
end = to_datetime or datetime.now()
start, end, _ = self._normalize_time_window('custom' if from_datetime or to_datetime else '24h', from_datetime, to_datetime)
# Get providers that have data
providers = self.get_all_providers_stats(from_datetime, to_datetime, user_filter,
......@@ -1558,9 +1567,7 @@ class Analytics:
Per-rotation breakdown: which provider/model received what share of hits and tokens.
Returns list of {rotation_id, entries: [{provider_id, model_name, requests, tokens, hit_pct, token_pct, avg_latency_ms}]}
"""
now = datetime.now()
start = from_datetime or (now - timedelta(days=1))
end = to_datetime or now
start, end, _ = self._normalize_time_window('custom' if from_datetime or to_datetime else '24h', from_datetime, to_datetime)
with self.db._get_connection() as conn:
cursor = conn.cursor()
......@@ -1628,9 +1635,7 @@ class Analytics:
Returns list of {autoselect_id, entries: [{model_name, requests, tokens, hit_pct, token_pct, avg_latency_ms}]}
Selection latency is stored as latency_ms on the 'autoselect' provider_id rows.
"""
now = datetime.now()
start = from_datetime or (now - timedelta(days=1))
end = to_datetime or now
start, end, _ = self._normalize_time_window('custom' if from_datetime or to_datetime else '24h', from_datetime, to_datetime)
with self.db._get_connection() as conn:
cursor = conn.cursor()
......
aisbf/routes/dashboard/providers.py
View file @ 6130a6db
......@@ -880,8 +880,11 @@ async def dashboard_providers_save(request: Request, config: str = Form(...)):
"request": request,
"session": request.session,
"__version__": __version__,
"user_providers_data": providers_data,
"user_providers_bootstrap_json": _json_parse_bootstrap(providers_data),
"user_providers_data": list(providers_data.values()),
"user_providers_bootstrap_json": _json_parse_bootstrap([
{"provider_id": provider_id, "config": provider_config}
for provider_id, provider_config in providers_data.items()
]),
"studio_capability_choices": serialize_studio_capability_choices(),
"studio_adapter_choices": serialize_studio_adapter_choices(),
"studio_adapter_profile_choices": serialize_studio_adapter_profile_choices(),
......@@ -2293,6 +2296,21 @@ async def dashboard_analytics(
today = datetime.now().replace(hour=0, minute=0, second=0, microsecond=0)
from_datetime = today - timedelta(days=1)
to_datetime = today - timedelta(microseconds=1)
elif time_range == '1h':
from_datetime = datetime.now() - timedelta(hours=1)
to_datetime = datetime.now()
elif time_range == '6h':
from_datetime = datetime.now() - timedelta(hours=6)
to_datetime = datetime.now()
elif time_range == '7d':
from_datetime = datetime.now() - timedelta(days=7)
to_datetime = datetime.now()
elif time_range == '30d':
from_datetime = datetime.now() - timedelta(days=30)
to_datetime = datetime.now()
elif time_range == '90d':
from_datetime = datetime.now() - timedelta(days=90)
to_datetime = datetime.now()
elif time_range == 'custom':
if not from_datetime or not to_datetime:
time_range = '24h'
......
templates/dashboard/analytics.html
View file @ 6130a6db
......@@ -607,9 +607,9 @@ window.onclick = function(event) {
<p style="font-size: 28px; font-weight: bold;" id="savings-amount">{{ currency_symbol }}{{ "%.2f"|format((optimization_savings.total_cost_saved if optimization_savings else 0) or 0) }}</p>
<small style="color: rgba(255,255,255,0.8);">
{% if optimization_savings and optimization_savings.provider_equivalents %}
Provider free-tier equivalents + optimization
Configured free-tier equivalents in selected period
{% else %}
Cache hits & optimization
No configured free-tier savings in selected period
{% endif %}
</small>
</div>
......@@ -646,8 +646,8 @@ window.onclick = function(event) {
<div style="background: var(--bg-page); padding: 16px; border-radius: 8px; margin-bottom: 20px;">
<h4 style="margin: 0 0 10px 0;">Free-Tier Equivalent Savings</h4>
<p style="margin: 0 0 10px 0; color: var(--color-muted);">
The table below estimates provider-side free-tier equivalents using known upstream free limits such as weekly or monthly request allowances.
When the filtered usage exceeds a provider's base free allowance, the overflow is counted as additional free-tier-equivalent capacity and translated into token and premium-value equivalents.
The table below only uses explicitly configured provider free-tier limits. Generic defaults are excluded to avoid overstated savings on subscription or bundled plans.
Extra free-tier equivalents are counted only for complete additional quota blocks beyond the first included tier within the selected period.
</p>
<table>
<tr>
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment