openlit 1.16.2__py3-none-any.whl → 1.18.0__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (28) hide show
  1. openlit/__init__.py +3 -0
  2. openlit/instrumentation/anthropic/anthropic.py +28 -10
  3. openlit/instrumentation/anthropic/async_anthropic.py +27 -10
  4. openlit/instrumentation/bedrock/bedrock.py +13 -5
  5. openlit/instrumentation/cohere/cohere.py +33 -12
  6. openlit/instrumentation/elevenlabs/async_elevenlabs.py +6 -2
  7. openlit/instrumentation/elevenlabs/elevenlabs.py +6 -2
  8. openlit/instrumentation/gpt4all/gpt4all.py +30 -10
  9. openlit/instrumentation/groq/async_groq.py +31 -11
  10. openlit/instrumentation/groq/groq.py +31 -11
  11. openlit/instrumentation/mistral/async_mistral.py +33 -12
  12. openlit/instrumentation/mistral/mistral.py +33 -12
  13. openlit/instrumentation/ollama/async_ollama.py +57 -20
  14. openlit/instrumentation/ollama/ollama.py +57 -20
  15. openlit/instrumentation/openai/async_azure_openai.py +94 -35
  16. openlit/instrumentation/openai/async_openai.py +68 -27
  17. openlit/instrumentation/openai/azure_openai.py +89 -31
  18. openlit/instrumentation/openai/openai.py +68 -29
  19. openlit/instrumentation/transformers/transformers.py +20 -16
  20. openlit/instrumentation/vertexai/async_vertexai.py +104 -35
  21. openlit/instrumentation/vertexai/vertexai.py +104 -35
  22. openlit/instrumentation/vllm/__init__.py +43 -0
  23. openlit/instrumentation/vllm/vllm.py +143 -0
  24. openlit/semcov/__init__.py +4 -1
  25. {openlit-1.16.2.dist-info → openlit-1.18.0.dist-info}/METADATA +3 -1
  26. {openlit-1.16.2.dist-info → openlit-1.18.0.dist-info}/RECORD +28 -26
  27. {openlit-1.16.2.dist-info → openlit-1.18.0.dist-info}/LICENSE +0 -0
  28. {openlit-1.16.2.dist-info → openlit-1.18.0.dist-info}/WHEEL +0 -0
openlit/instrumentation/mistral/mistral.py CHANGED
@@ -102,7 +102,7 @@ def chat(gen_ai_endpoint, version, environment, application_name,
102
102
  span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
103
103
  False)
104
104
  span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_FINISH_REASON,
105
- response.choices[0].finish_reason)
105
+ [response.choices[0].finish_reason])
106
106
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
107
107
  response.usage.prompt_tokens)
108
108
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COMPLETION_TOKENS,
@@ -112,10 +112,19 @@ def chat(gen_ai_endpoint, version, environment, application_name,
112
112
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
113
113
  cost)
114
114
  if trace_content:
115
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
116
- prompt)
117
- # pylint: disable=line-too-long
118
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION, response.choices[0].message.content if response.choices[0].message.content else "")
115
+ span.add_event(
116
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
117
+ attributes={
118
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
119
+ },
120
+ )
121
+ span.add_event(
122
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
123
+ # pylint: disable=line-too-long
124
+ attributes={
125
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: response.choices[0].message.content if response.choices[0].message.content else "",
126
+ },
127
+ )
119
128
 
120
129
  span.set_status(Status(StatusCode.OK))
121
130
 
@@ -259,7 +268,7 @@ def chat_stream(gen_ai_endpoint, version, environment, application_name,
259
268
  span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
260
269
  True)
261
270
  span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_FINISH_REASON,
262
- finish_reason)
271
+ [finish_reason])
263
272
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
264
273
  prompt_tokens)
265
274
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COMPLETION_TOKENS,
@@ -269,10 +278,18 @@ def chat_stream(gen_ai_endpoint, version, environment, application_name,
269
278
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
270
279
  cost)
271
280
  if trace_content:
272
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
273
- prompt)
274
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
275
- llmresponse)
281
+ span.add_event(
282
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
283
+ attributes={
284
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
285
+ },
286
+ )
287
+ span.add_event(
288
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
289
+ attributes={
290
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: llmresponse,
291
+ },
292
+ )
276
293
 
277
294
  span.set_status(Status(StatusCode.OK))
278
295
 
@@ -377,8 +394,12 @@ def embeddings(gen_ai_endpoint, version, environment, application_name,
377
394
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
378
395
  cost)
379
396
  if trace_content:
380
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
381
- prompt)
397
+ span.add_event(
398
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
399
+ attributes={
400
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
401
+ },
402
+ )
382
403
 
383
404
  span.set_status(Status(StatusCode.OK))
384
405
 
openlit/instrumentation/ollama/async_ollama.py CHANGED
@@ -120,10 +120,18 @@ def async_chat(gen_ai_endpoint, version, environment, application_name,
120
120
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
121
121
  cost)
122
122
  if trace_content:
123
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
124
- prompt)
125
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
126
- llmresponse)
123
+ span.add_event(
124
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
125
+ attributes={
126
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
127
+ },
128
+ )
129
+ span.add_event(
130
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
131
+ attributes={
132
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: llmresponse,
133
+ },
134
+ )
127
135
 
128
136
  span.set_status(Status(StatusCode.OK))
129
137
 
@@ -198,10 +206,18 @@ def async_chat(gen_ai_endpoint, version, environment, application_name,
198
206
  span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
199
207
  False)
200
208
  if trace_content:
201
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
202
- prompt)
203
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
204
- response['message']['content'])
209
+ span.add_event(
210
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
211
+ attributes={
212
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
213
+ },
214
+ )
215
+ span.add_event(
216
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
217
+ attributes={
218
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: response['message']['content'],
219
+ },
220
+ )
205
221
 
206
222
  # Calculate cost of the operation
207
223
  cost = 0
@@ -216,7 +232,7 @@ def async_chat(gen_ai_endpoint, version, environment, application_name,
216
232
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
217
233
  total_tokens)
218
234
  span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_FINISH_REASON,
219
- response["done_reason"])
235
+ [response["done_reason"]])
220
236
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
221
237
  cost)
222
238
 
@@ -345,10 +361,19 @@ def async_generate(gen_ai_endpoint, version, environment, application_name,
345
361
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
346
362
  cost)
347
363
  if trace_content:
348
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
349
- kwargs.get("prompt", ""))
350
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
351
- llmresponse)
364
+ span.add_event(
365
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
366
+ attributes={
367
+ # pylint: disable=line-too-long
368
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: kwargs.get("prompt", ""),
369
+ },
370
+ )
371
+ span.add_event(
372
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
373
+ attributes={
374
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: llmresponse,
375
+ },
376
+ )
352
377
 
353
378
  span.set_status(Status(StatusCode.OK))
354
379
 
@@ -404,10 +429,18 @@ def async_generate(gen_ai_endpoint, version, environment, application_name,
404
429
  span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
405
430
  False)
406
431
  if trace_content:
407
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
408
- kwargs.get("prompt", ""))
409
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
410
- response['response'])
432
+ span.add_event(
433
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
434
+ attributes={
435
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: kwargs.get("prompt", ""),
436
+ },
437
+ )
438
+ span.add_event(
439
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
440
+ attributes={
441
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: response['response'],
442
+ },
443
+ )
411
444
 
412
445
  # Calculate cost of the operation
413
446
  cost = 0
@@ -422,7 +455,7 @@ def async_generate(gen_ai_endpoint, version, environment, application_name,
422
455
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
423
456
  total_tokens)
424
457
  span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_FINISH_REASON,
425
- response["done_reason"])
458
+ [response["done_reason"]])
426
459
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
427
460
  cost)
428
461
 
@@ -525,8 +558,12 @@ def async_embeddings(gen_ai_endpoint, version, environment, application_name,
525
558
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
526
559
  cost)
527
560
  if trace_content:
528
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
529
- kwargs.get('prompt', ""))
561
+ span.add_event(
562
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
563
+ attributes={
564
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: kwargs.get("prompt", ""),
565
+ },
566
+ )
530
567
 
531
568
  span.set_status(Status(StatusCode.OK))
532
569
 
openlit/instrumentation/ollama/ollama.py CHANGED
@@ -120,10 +120,18 @@ def chat(gen_ai_endpoint, version, environment, application_name,
120
120
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
121
121
  cost)
122
122
  if trace_content:
123
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
124
- prompt)
125
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
126
- llmresponse)
123
+ span.add_event(
124
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
125
+ attributes={
126
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
127
+ },
128
+ )
129
+ span.add_event(
130
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
131
+ attributes={
132
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: llmresponse,
133
+ },
134
+ )
127
135
 
128
136
  span.set_status(Status(StatusCode.OK))
129
137
 
@@ -198,10 +206,18 @@ def chat(gen_ai_endpoint, version, environment, application_name,
198
206
  span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
199
207
  False)
200
208
  if trace_content:
201
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
202
- prompt)
203
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
204
- response['message']['content'])
209
+ span.add_event(
210
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
211
+ attributes={
212
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
213
+ },
214
+ )
215
+ span.add_event(
216
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
217
+ attributes={
218
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: response['message']['content'],
219
+ },
220
+ )
205
221
 
206
222
  # Calculate cost of the operation
207
223
  cost = 0
@@ -216,7 +232,7 @@ def chat(gen_ai_endpoint, version, environment, application_name,
216
232
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
217
233
  total_tokens)
218
234
  span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_FINISH_REASON,
219
- response["done_reason"])
235
+ [response["done_reason"]])
220
236
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
221
237
  cost)
222
238
 
@@ -345,10 +361,19 @@ def generate(gen_ai_endpoint, version, environment, application_name,
345
361
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
346
362
  cost)
347
363
  if trace_content:
348
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
349
- kwargs.get("prompt", ""))
350
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
351
- llmresponse)
364
+ span.add_event(
365
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
366
+ attributes={
367
+ # pylint: disable=line-too-long
368
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: kwargs.get("prompt", ""),
369
+ },
370
+ )
371
+ span.add_event(
372
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
373
+ attributes={
374
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: llmresponse,
375
+ },
376
+ )
352
377
 
353
378
  span.set_status(Status(StatusCode.OK))
354
379
 
@@ -404,10 +429,18 @@ def generate(gen_ai_endpoint, version, environment, application_name,
404
429
  span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
405
430
  False)
406
431
  if trace_content:
407
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
408
- kwargs.get("prompt", ""))
409
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
410
- response['response'])
432
+ span.add_event(
433
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
434
+ attributes={
435
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: kwargs.get("prompt", ""),
436
+ },
437
+ )
438
+ span.add_event(
439
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
440
+ attributes={
441
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: response['response'],
442
+ },
443
+ )
411
444
 
412
445
  # Calculate cost of the operation
413
446
  cost = 0
@@ -422,7 +455,7 @@ def generate(gen_ai_endpoint, version, environment, application_name,
422
455
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
423
456
  total_tokens)
424
457
  span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_FINISH_REASON,
425
- response["done_reason"])
458
+ [response["done_reason"]])
426
459
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
427
460
  cost)
428
461
 
@@ -525,8 +558,12 @@ def embeddings(gen_ai_endpoint, version, environment, application_name,
525
558
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
526
559
  cost)
527
560
  if trace_content:
528
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
529
- kwargs.get('prompt', ""))
561
+ span.add_event(
562
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
563
+ attributes={
564
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: kwargs.get("prompt", ""),
565
+ },
566
+ )
530
567
 
531
568
  span.set_status(Status(StatusCode.OK))
532
569
 
openlit/instrumentation/openai/async_azure_openai.py CHANGED
@@ -143,10 +143,18 @@ def azure_async_chat_completions(gen_ai_endpoint, version, environment, applicat
143
143
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
144
144
  cost)
145
145
  if trace_content:
146
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
147
- prompt)
148
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
149
- llmresponse)
146
+ span.add_event(
147
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
148
+ attributes={
149
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
150
+ },
151
+ )
152
+ span.add_event(
153
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
154
+ attributes={
155
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: llmresponse,
156
+ },
157
+ )
150
158
 
151
159
  span.set_status(Status(StatusCode.OK))
152
160
 
@@ -238,8 +246,12 @@ def azure_async_chat_completions(gen_ai_endpoint, version, environment, applicat
238
246
  span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
239
247
  False)
240
248
  if trace_content:
241
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
242
- prompt)
249
+ span.add_event(
250
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
251
+ attributes={
252
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
253
+ },
254
+ )
243
255
 
244
256
  # Set span attributes when tools is not passed to the function call
245
257
  if "tools" not in kwargs:
@@ -255,23 +267,31 @@ def azure_async_chat_completions(gen_ai_endpoint, version, environment, applicat
255
267
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
256
268
  response.usage.total_tokens)
257
269
  span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_FINISH_REASON,
258
- response.choices[0].finish_reason)
270
+ [response.choices[0].finish_reason])
259
271
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
260
272
  cost)
261
273
 
262
274
  # Set span attributes for when n = 1 (default)
263
275
  if "n" not in kwargs or kwargs["n"] == 1:
264
276
  if trace_content:
265
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
266
- response.choices[0].message.content)
277
+ span.add_event(
278
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
279
+ attributes={
280
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: response.choices[0].message.content,
281
+ },
282
+ )
267
283
 
268
284
  # Set span attributes for when n > 0
269
285
  else:
270
286
  i = 0
271
287
  while i < kwargs["n"] and trace_content is True:
272
- attribute_name = f"gen_ai.completion.{i}"
273
- span.set_attribute(attribute_name,
274
- response.choices[i].message.content)
288
+ attribute_name = f"gen_ai.content.completion.{i}"
289
+ span.add_event(
290
+ name=attribute_name,
291
+ attributes={
292
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: response.choices[i].message.content,
293
+ },
294
+ )
275
295
  i += 1
276
296
 
277
297
  # Return original response
@@ -284,8 +304,12 @@ def azure_async_chat_completions(gen_ai_endpoint, version, environment, applicat
284
304
  response.usage.prompt_tokens,
285
305
  response.usage.completion_tokens)
286
306
 
287
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
288
- "Function called with tools")
307
+ span.add_event(
308
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
309
+ attributes={
310
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: "Function called with tools",
311
+ },
312
+ )
289
313
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
290
314
  response.usage.prompt_tokens)
291
315
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COMPLETION_TOKENS,
@@ -443,10 +467,18 @@ def azure_async_completions(gen_ai_endpoint, version, environment, application_n
443
467
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
444
468
  cost)
445
469
  if trace_content:
446
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
447
- prompt)
448
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
449
- llmresponse)
470
+ span.add_event(
471
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
472
+ attributes={
473
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
474
+ },
475
+ )
476
+ span.add_event(
477
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
478
+ attributes={
479
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: llmresponse,
480
+ },
481
+ )
450
482
 
451
483
  span.set_status(Status(StatusCode.OK))
452
484
 
@@ -519,8 +551,12 @@ def azure_async_completions(gen_ai_endpoint, version, environment, application_n
519
551
  span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
520
552
  False)
521
553
  if trace_content:
522
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
523
- kwargs.get("prompt", ""))
554
+ span.add_event(
555
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
556
+ attributes={
557
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: kwargs.get("prompt", ""),
558
+ },
559
+ )
524
560
 
525
561
  # Set span attributes when tools is not passed to the function call
526
562
  if "tools" not in kwargs:
@@ -536,23 +572,31 @@ def azure_async_completions(gen_ai_endpoint, version, environment, application_n
536
572
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
537
573
  response.usage.total_tokens)
538
574
  span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_FINISH_REASON,
539
- response.choices[0].finish_reason)
575
+ [response.choices[0].finish_reason])
540
576
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
541
577
  cost)
542
578
 
543
579
  # Set span attributes for when n = 1 (default)
544
580
  if "n" not in kwargs or kwargs["n"] == 1:
545
581
  if trace_content:
546
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
547
- response.choices[0].text)
582
+ span.add_event(
583
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
584
+ attributes={
585
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: response.choices[0].text,
586
+ },
587
+ )
548
588
 
549
589
  # Set span attributes for when n > 0
550
590
  else:
551
591
  i = 0
552
592
  while i < kwargs["n"] and trace_content is True:
553
- attribute_name = f"gen_ai.completion.{i}"
554
- span.set_attribute(attribute_name,
555
- response.choices[i].text)
593
+ attribute_name = f"gen_ai.content.completion.{i}"
594
+ span.add_event(
595
+ name=attribute_name,
596
+ attributes={
597
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: response.choices[i].text,
598
+ },
599
+ )
556
600
  i += 1
557
601
  return response
558
602
 
@@ -563,8 +607,12 @@ def azure_async_completions(gen_ai_endpoint, version, environment, application_n
563
607
  response.usage.prompt_tokens,
564
608
  response.usage.completion_tokens)
565
609
 
566
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
567
- "Function called with tools")
610
+ span.add_event(
611
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
612
+ attributes={
613
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: "Function called with tools",
614
+ },
615
+ )
568
616
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
569
617
  response.usage.prompt_tokens)
570
618
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COMPLETION_TOKENS,
@@ -679,8 +727,12 @@ def azure_async_embedding(gen_ai_endpoint, version, environment, application_nam
679
727
  response.usage.total_tokens)
680
728
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST, cost)
681
729
  if trace_content:
682
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
683
- kwargs.get("input", ""))
730
+ span.add_event(
731
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
732
+ attributes={
733
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: kwargs.get("input", ""),
734
+ },
735
+ )
684
736
 
685
737
  span.set_status(Status(StatusCode.OK))
686
738
 
@@ -796,12 +848,19 @@ def azure_async_image_generate(gen_ai_endpoint, version, environment, applicatio
796
848
  span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_USER,
797
849
  kwargs.get("user", ""))
798
850
  if trace_content:
799
- span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_PROMPT,
800
- kwargs.get("prompt", ""))
801
-
851
+ span.add_event(
852
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
853
+ attributes={
854
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: kwargs.get("prompt", ""),
855
+ },
856
+ )
802
857
  attribute_name = f"gen_ai.response.image.{images_count}"
803
- span.set_attribute(attribute_name,
804
- getattr(items, image))
858
+ span.add_event(
859
+ name=attribute_name,
860
+ attributes={
861
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: getattr(items, image),
862
+ },
863
+ )
805
864
 
806
865
  images_count+=1
807
866