github/langchain
1
0
Fork 0
mirror of https://github.com/hwchase17/langchain.git synced 2025-09-01 11:02:37 +00:00
Code Issues Packages Projects Releases Wiki Activity

anthropic[patch]: cache tokens nit (#31484)

Browse Source 
if you pass in beta headers directly cache_creation is a dict
This commit is contained in:
Bagatur
2025-06-05 16:15:03 -04:00
committed by GitHub
parent 14c561e15d
commit f7f52cab12
1 changed files with 6 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
libs/partners/anthropic/langchain_anthropic/chat_models.py
View File

@@ -2131,11 +2131,12 @@ def _create_usage_metadata(anthropic_usage: BaseModel) -> UsageMetadata:
}
# Add (beta) cache TTL information if available
cache_creation = getattr(anthropic_usage, "cache_creation", None)
cache_creation_keys = ("ephemeral_1h_input_tokens", "ephemeral_5m_input_tokens")
if cache_creation:
for k in ["ephemeral_1h_input_tokens", "ephemeral_5m_input_tokens"]:
v = getattr(cache_creation, k, None)
if v:
input_token_details[k] = v
if isinstance(cache_creation, BaseModel):
cache_creation = cache_creation.model_dump()
for k in cache_creation_keys:
input_token_details[k] = cache_creation.get(k)
# Anthropic input_tokens exclude cached token counts.
input_tokens = (
@@ -2149,6 +2150,6 @@ def _create_usage_metadata(anthropic_usage: BaseModel) -> UsageMetadata:
output_tokens=output_tokens,
total_tokens=input_tokens + output_tokens,
input_token_details=InputTokenDetails(
**{k: v for k, v in input_token_details.items() if v is not None}
**{k: v for k, v in input_token_details.items() if v is not None},
),
)