track llmReqId wherever possible

src/dokumetry/anthropic.py

@@ -45,6 +45,7 @@ def patched_completions_create(*args, **kwargs):
         completion_tokens = llm.count_tokens(response.completion)
 
         data = {
+                "llmReqId": response.id,
                 "environment": environment,
                 "applicationName": application_name,
                 "sourceLanguage": "python",

src/dokumetry/cohere.py

@@ -64,7 +64,6 @@ def stream_generator():
                 for event in original_generate(*args, **kwargs):
                     accumulated_content += event.text
                     yield event
-
                 end_time = time.time()
                 duration = end_time - start_time
                 prompt = kwargs.get('prompt')
@@ -96,6 +95,7 @@ def stream_generator():
 
             for generation in response:
                 data = {
+                    "llmReqId": generation.id,
                     "environment": environment,
                     "applicationName": application_name,
                     "sourceLanguage": "python",
@@ -168,14 +168,16 @@ def chat_generate(*args, **kwargs):
             def stream_generator():
                 accumulated_content = ""
                 for event in original_chat(*args, **kwargs):
+                    if event.event_type == "stream-start":
+                        responseId = event.generation_id
                     if event.event_type == "text-generation":
                         accumulated_content += event.text
                     yield event
-
                 end_time = time.time()
                 duration = end_time - start_time
                 prompt = kwargs.get('message')
                 data = {
+                    "llmReqId": responseId,
                     "environment": environment,
                     "applicationName": application_name,
                     "sourceLanguage": "python",
@@ -201,6 +203,7 @@ def stream_generator():
             model = kwargs.get('model', "command")
             prompt = kwargs.get('message')
             data = {
+                "llmReqId": response.generation_id,
                 "environment": environment,
                 "applicationName": application_name,
                 "sourceLanguage": "python",
@@ -239,6 +242,7 @@ def summarize_generate(*args, **kwargs):
         prompt = kwargs.get('text')
 
         data = {
+                "llmReqId": response.id,
                 "environment": environment,
                 "applicationName": application_name,
                 "sourceLanguage": "python",

src/dokumetry/openai.py

@@ -53,6 +53,7 @@ def stream_generator():
                         if content:
                             accumulated_content += content
                     yield chunk
+                    responseId = chunk.id
                 end_time = time.time()
                 duration = end_time - start_time
                 message_prompt = kwargs.get('messages', "No prompt provided")
@@ -74,6 +75,7 @@ def stream_generator():
 
                 prompt = "\n".join(formatted_messages)
                 data = {
+                    "llmReqId": responseId,
                     "environment": environment,
                     "applicationName": application_name,
                     "sourceLanguage": "python",
@@ -114,6 +116,7 @@ def stream_generator():
             prompt = "\n".join(formatted_messages)
 
             data = {
+                "llmReqId": response.id,
                 "environment": environment,
                 "applicationName": application_name,
                 "sourceLanguage": "python",
@@ -172,10 +175,12 @@ def stream_generator():
                         if content:
                             accumulated_content += content
                     yield chunk
+                    responseId = chunk.id
                 end_time = time.time()
                 duration = end_time - start_time
                 prompt = kwargs.get('prompt', "No prompt provided")
                 data = {
+                    "llmReqId": responseId,
                     "environment": environment,
                     "applicationName": application_name,
                     "sourceLanguage": "python",
@@ -199,6 +204,7 @@ def stream_generator():
             prompt = kwargs.get('prompt', "No prompt provided")
 
             data = {
+                "llmReqId": response.id,
                 "environment": environment,
                 "applicationName": application_name,
                 "sourceLanguage": "python",
@@ -296,7 +302,7 @@ def patched_fine_tuning_create(*args, **kwargs):
             "skipResp": skip_resp,
             "requestDuration": duration,
             "model": model,
-            "finetuneJobId": response.id,
+            "llmReqId": response.id,
             "finetuneJobStatus": response.status,
         }
 
@@ -339,6 +345,7 @@ def patched_image_create(*args, **kwargs):
 
         for items in response.data:
             data = {
+                "llmReqId": response.created,
                 "environment": environment,
                 "applicationName": application_name,
                 "sourceLanguage": "python",
@@ -387,6 +394,7 @@ def patched_image_create_variation(*args, **kwargs):
         for items in response.data:
 
             data = {
+                "llmReqId": response.created,
                 "environment": environment,
                 "applicationName": application_name,
                 "sourceLanguage": "python",