Hugging Face's logo

Spaces:
Shyamnath
/
inferencing-llm
Sleeping

App Files Community
inferencing-llm
File size: 12,147 Bytes 
c40c75a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
 
import React from 'react';
import { Card, Grid, Text, Title, Accordion, AccordionHeader, AccordionBody } from '@tremor/react';
import { AreaChart, BarChart } from '@tremor/react';
import { SpendMetrics, DailyData, ModelActivityData, MetricWithMetadata, KeyMetricWithMetadata } from './usage/types';
import { Collapse } from 'antd';

interface ActivityMetricsProps {
  modelMetrics: Record<string, ModelActivityData>;
}

const ModelSection = ({ modelName, metrics }: { modelName: string; metrics: ModelActivityData }) => {
  return (
    <div className="space-y-2">
      {/* Summary Cards */}
      <Grid numItems={4} className="gap-4">
        <Card>
          <Text>Total Requests</Text>
          <Title>{metrics.total_requests.toLocaleString()}</Title>
        </Card>
        <Card>
          <Text>Total Successful Requests</Text>
          <Title>{metrics.total_successful_requests.toLocaleString()}</Title>
        </Card>
        <Card>
          <Text>Total Tokens</Text>
          <Title>{metrics.total_tokens.toLocaleString()}</Title>
          <Text>{Math.round(metrics.total_tokens / metrics.total_successful_requests)} avg per successful request</Text>
        </Card>
        <Card>
          <Text>Total Spend</Text>
          <Title>${metrics.total_spend.toFixed(2)}</Title>
          <Text>${(metrics.total_spend / metrics.total_successful_requests).toFixed(3)} per successful request</Text>
        </Card>
      </Grid>

      {/* Charts */}
      <Grid numItems={2} className="gap-4">
        <Card>
          <Title>Total Tokens</Title>
          <AreaChart    
            data={metrics.daily_data}
            index="date"
            categories={["metrics.prompt_tokens", "metrics.completion_tokens", "metrics.total_tokens"]}
            colors={["blue", "cyan", "indigo"]}
            valueFormatter={(number: number) => number.toLocaleString()}
          />
        </Card>

        <Card>
          <Title>Requests per day</Title>
          <BarChart
            data={metrics.daily_data}
            index="date"
            categories={["metrics.api_requests"]}
            colors={["blue"]}
            valueFormatter={(number: number) => number.toLocaleString()}
          />
        </Card>

        <Card>
          <Title>Spend per day</Title>
          <BarChart
            data={metrics.daily_data}
            index="date"
            categories={["metrics.spend"]}
            colors={["green"]}
            valueFormatter={(value: number) => `$${value.toFixed(2)}`}
          />
        </Card>

        <Card>
          <Title>Success vs Failed Requests</Title>
          <AreaChart
            data={metrics.daily_data}
            index="date"
            categories={["metrics.successful_requests", "metrics.failed_requests"]}
            colors={["emerald", "red"]}
            valueFormatter={(number: number) => number.toLocaleString()}
            stack
          />
        </Card>
        
        <Card>
          <Title>Prompt Caching Metrics</Title>
          <div className="mb-2">
            <Text>Cache Read: {metrics.total_cache_read_input_tokens?.toLocaleString() || 0} tokens</Text>
            <Text>Cache Creation: {metrics.total_cache_creation_input_tokens?.toLocaleString() || 0} tokens</Text>
          </div>
          <AreaChart
            data={metrics.daily_data}
            index="date"
            categories={["metrics.cache_read_input_tokens", "metrics.cache_creation_input_tokens"]}
            colors={["cyan", "purple"]}
            valueFormatter={(number: number) => number.toLocaleString()}
          />
        </Card>
      </Grid>
    </div>
  );
};

export const ActivityMetrics: React.FC<ActivityMetricsProps> = ({ modelMetrics }) => {
  const modelNames = Object.keys(modelMetrics).sort((a, b) => {
    if (a === '') return 1;
    if (b === '') return -1;
    return modelMetrics[b].total_spend - modelMetrics[a].total_spend;
  });

  // Calculate total metrics across all models
  const totalMetrics = {
    total_requests: 0,
    total_successful_requests: 0,
    total_tokens: 0,
    total_spend: 0,
    total_cache_read_input_tokens: 0,
    total_cache_creation_input_tokens: 0,
    daily_data: {} as Record<string, {
      prompt_tokens: number;
      completion_tokens: number;
      total_tokens: number;
      api_requests: number;
      spend: number;
      successful_requests: number;
      failed_requests: number;
      cache_read_input_tokens: number;
      cache_creation_input_tokens: number;
    }>
  };

  // Aggregate data
  Object.values(modelMetrics).forEach(model => {
    totalMetrics.total_requests += model.total_requests;
    totalMetrics.total_successful_requests += model.total_successful_requests;
    totalMetrics.total_tokens += model.total_tokens;
    totalMetrics.total_spend += model.total_spend;
    totalMetrics.total_cache_read_input_tokens += model.total_cache_read_input_tokens || 0;
    totalMetrics.total_cache_creation_input_tokens += model.total_cache_creation_input_tokens || 0;

    // Aggregate daily data
    model.daily_data.forEach(day => {
      if (!totalMetrics.daily_data[day.date]) {
        totalMetrics.daily_data[day.date] = {
          prompt_tokens: 0,
          completion_tokens: 0,
          total_tokens: 0,
          api_requests: 0,
          spend: 0,
          successful_requests: 0,
          failed_requests: 0,
          cache_read_input_tokens: 0,
          cache_creation_input_tokens: 0
        };
      }
      totalMetrics.daily_data[day.date].prompt_tokens += day.metrics.prompt_tokens;
      totalMetrics.daily_data[day.date].completion_tokens += day.metrics.completion_tokens;
      totalMetrics.daily_data[day.date].total_tokens += day.metrics.total_tokens;
      totalMetrics.daily_data[day.date].api_requests += day.metrics.api_requests;
      totalMetrics.daily_data[day.date].spend += day.metrics.spend;
      totalMetrics.daily_data[day.date].successful_requests += day.metrics.successful_requests;
      totalMetrics.daily_data[day.date].failed_requests += day.metrics.failed_requests;
      totalMetrics.daily_data[day.date].cache_read_input_tokens += day.metrics.cache_read_input_tokens || 0;
      totalMetrics.daily_data[day.date].cache_creation_input_tokens += day.metrics.cache_creation_input_tokens || 0;
    });
  });

  // Convert daily_data object to array and sort by date
  const sortedDailyData = Object.entries(totalMetrics.daily_data)
    .map(([date, metrics]) => ({ date, metrics }))
    .sort((a, b) => new Date(a.date).getTime() - new Date(b.date).getTime());

  return (
    <div className="space-y-8">
      {/* Global Summary */}
      <div className="border rounded-lg p-4">
        <Title>Overall Usage</Title>
        <Grid numItems={4} className="gap-4 mb-4">
          <Card>
            <Text>Total Requests</Text>
            <Title>{totalMetrics.total_requests.toLocaleString()}</Title>
          </Card>
          <Card>
            <Text>Total Successful Requests</Text>
            <Title>{totalMetrics.total_successful_requests.toLocaleString()}</Title>
          </Card>
          <Card>
            <Text>Total Tokens</Text>
            <Title>{totalMetrics.total_tokens.toLocaleString()}</Title>
          </Card>
          <Card>
            <Text>Total Spend</Text>
            <Title>${totalMetrics.total_spend.toFixed(2)}</Title>
          </Card>
        </Grid>

        <Grid numItems={2} className="gap-4">
          <Card>
            <Title>Total Tokens Over Time</Title>
            <AreaChart    
              data={sortedDailyData}
              index="date"
              categories={["metrics.prompt_tokens", "metrics.completion_tokens", "metrics.total_tokens"]}
              colors={["blue", "cyan", "indigo"]}
              valueFormatter={(number: number) => number.toLocaleString()}
            />
          </Card>
          <Card>
            <Title>Total Requests Over Time</Title>
            <AreaChart
              data={sortedDailyData}
              index="date"
              categories={["metrics.successful_requests", "metrics.failed_requests"]}
              colors={["emerald", "red"]}
              valueFormatter={(number: number) => number.toLocaleString()}
              stack
            />
          </Card>
        </Grid>
      </div>

      {/* Individual Model Sections */}
      <Collapse defaultActiveKey={modelNames[0]}>
        {modelNames.map((modelName) => (
          <Collapse.Panel 
            key={modelName} 
            header={
              <div className="flex justify-between items-center w-full">
                <Title>{modelMetrics[modelName].label || 'Unknown Item'}</Title>
                <div className="flex space-x-4 text-sm text-gray-500">
                  <span>${modelMetrics[modelName].total_spend.toFixed(2)}</span>
                  <span>{modelMetrics[modelName].total_requests.toLocaleString()} requests</span>
                </div>
              </div>
            }
          >
            <ModelSection 
              modelName={modelName || 'Unknown Model'} 
              metrics={modelMetrics[modelName]} 
            />
          </Collapse.Panel>
        ))}
      </Collapse>
    </div>
  );
};

// Helper function to format key label
const formatKeyLabel = (modelData: KeyMetricWithMetadata, model: string): string => {
  const keyAlias = modelData.metadata.key_alias || `key-hash-${model}`;
  const teamId = modelData.metadata.team_id;
  return teamId ? `${keyAlias} (team_id: ${teamId})` : keyAlias;
};

// Process data function
export const processActivityData = (dailyActivity: { results: DailyData[] }, key: "models" | "api_keys"): Record<string, ModelActivityData> => {
  const modelMetrics: Record<string, ModelActivityData> = {};

  dailyActivity.results.forEach((day) => {
    Object.entries(day.breakdown[key] || {}).forEach(([model, modelData]) => {
      if (!modelMetrics[model]) {
        modelMetrics[model] = {
          label: key === 'api_keys' 
            ? formatKeyLabel(modelData as KeyMetricWithMetadata, model)
            : model,
          total_requests: 0,
          total_successful_requests: 0,
          total_failed_requests: 0,
          total_tokens: 0,
          prompt_tokens: 0,
          completion_tokens: 0,
          total_spend: 0,
          total_cache_read_input_tokens: 0,
          total_cache_creation_input_tokens: 0,
          daily_data: []
        };
      }

      // Update totals
      modelMetrics[model].total_requests += modelData.metrics.api_requests;
      modelMetrics[model].prompt_tokens += modelData.metrics.prompt_tokens;
      modelMetrics[model].completion_tokens += modelData.metrics.completion_tokens;
      modelMetrics[model].total_tokens += modelData.metrics.total_tokens;
      modelMetrics[model].total_spend += modelData.metrics.spend;
      modelMetrics[model].total_successful_requests += modelData.metrics.successful_requests;
      modelMetrics[model].total_failed_requests += modelData.metrics.failed_requests;
      modelMetrics[model].total_cache_read_input_tokens += modelData.metrics.cache_read_input_tokens || 0;
      modelMetrics[model].total_cache_creation_input_tokens += modelData.metrics.cache_creation_input_tokens || 0;

      // Add daily data
      modelMetrics[model].daily_data.push({
        date: day.date,
        metrics: {
          prompt_tokens: modelData.metrics.prompt_tokens,
          completion_tokens: modelData.metrics.completion_tokens,
          total_tokens: modelData.metrics.total_tokens,
          api_requests: modelData.metrics.api_requests,
          spend: modelData.metrics.spend,
          successful_requests: modelData.metrics.successful_requests,
          failed_requests: modelData.metrics.failed_requests,
          cache_read_input_tokens: modelData.metrics.cache_read_input_tokens || 0,
          cache_creation_input_tokens: modelData.metrics.cache_creation_input_tokens || 0
        }
      });
    });
  });

  // Sort daily data
  Object.values(modelMetrics).forEach(metrics => {
    metrics.daily_data.sort((a, b) => new Date(a.date).getTime() - new Date(b.date).getTime());
  });

  return modelMetrics;
};