From 60d549a6a96a3bfbf12442fa1e85c3429b0cf71d Mon Sep 17 00:00:00 2001
From: ivanbelenky <ivandietert@gmail.com>
Date: Mon, 21 Oct 2024 15:24:47 -0300
Subject: [PATCH] prompt details and completion tokens details

---
 instructor/retry.py | 7 +++++--
 instructor/utils.py | 6 ++++++
 2 files changed, 11 insertions(+), 2 deletions(-)

diff --git a/instructor/retry.py b/instructor/retry.py
index ffc41194d..205c7843f 100644
--- a/instructor/retry.py
+++ b/instructor/retry.py
@@ -14,7 +14,7 @@
 from instructor.utils import update_total_usage
 from instructor.validators import AsyncValidationError
 from openai.types.chat import ChatCompletion
-from openai.types.completion_usage import CompletionUsage
+from openai.types.completion_usage import CompletionUsage, CompletionTokensDetails, PromptTokensDetails
 from pydantic import BaseModel, ValidationError
 from tenacity import (
     AsyncRetrying,
@@ -71,7 +71,10 @@ def initialize_usage(mode: Mode) -> CompletionUsage | Any:
     Returns:
         CompletionUsage | Any: Initialized usage object.
     """
-    total_usage = CompletionUsage(completion_tokens=0, prompt_tokens=0, total_tokens=0)
+    total_usage = CompletionUsage(completion_tokens=0, prompt_tokens=0, total_tokens=0,
+        completion_tokens_details = CompletionTokensDetails(audio_tokens=0, reasoning_tokens=0),
+        prompt_token_details = PromptTokensDetails(audio_tokens=0, cached_tokens=0)
+    )
     if mode in {Mode.ANTHROPIC_TOOLS, Mode.ANTHROPIC_JSON}:
         from anthropic.types import Usage as AnthropicUsage
 
diff --git a/instructor/utils.py b/instructor/utils.py
index e9cacac25..e477f53ad 100644
--- a/instructor/utils.py
+++ b/instructor/utils.py
@@ -140,6 +140,12 @@ def update_total_usage(
         total_usage.completion_tokens += response_usage.completion_tokens or 0
         total_usage.prompt_tokens += response_usage.prompt_tokens or 0
         total_usage.total_tokens += response_usage.total_tokens or 0
+        if (rtd := response_usage.completion_tokens_details) and (ttd := total_usage.completion_tokens_details):
+            ttd.audio_tokens = (ttd.audio_tokens or 0) + (rtd.audio_tokens or 0)
+            ttd.reasoning_tokens = (ttd.reasoning_tokens or 0) + (rtd.reasoning_tokens or 0)
+        if (rpd := response_usage.prompt_tokens_details) and (tpd := total_usage.prompt_tokens_details):
+            tpd.audio_tokens = (tpd.audio_tokens or 0) + (rpd.audio_tokens or 0)
+            tpd.cached_tokens = (tpd.cached_tokens or 0) + (rpd.cached_tokens or 0)
         response.usage = total_usage  # Replace each response usage with the total usage
         return response