Skip to content

VertexAI usage metadata parsing #1171

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
ReneFabricius wants to merge 2 commits into from
Closed

VertexAI usage metadata parsing #1171

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3775292
Don't parse empty usage_metadata
ReneFabricius Apr 24, 2025
10ceef6
Parse vertexai usage details
ReneFabricius Apr 25, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
35 changes: 34 additions & 1 deletion langfuse/callback/langchain.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1066,6 +1066,7 @@ def _parse_usage_model(usage: typing.Union[pydantic.BaseModel, dict]):
# https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/get-token-count
("prompt_token_count", "input"),
("candidates_token_count", "output"),
("total_token_count", "total"),
# Bedrock: https://docs.aws.amazon.com/bedrock/latest/userguide/monitoring-cw.html#runtime-cloudwatch-metrics
("inputTokenCount", "input"),
("outputTokenCount", "output"),
Expand Down Expand Up @@ -1114,6 +1115,38 @@ def _parse_usage_model(usage: typing.Union[pydantic.BaseModel, dict]):
if "output" in usage_model:
usage_model["output"] = max(0, usage_model["output"] - value)

if "prompt_tokens_details" in usage_model:
prompt_tokens_details = usage_model.pop("prompt_tokens_details", [])

for entry in prompt_tokens_details:
if (
isinstance(entry, dict)
and "modality" in entry
and "token_count" in entry
):
value = entry["token_count"]
usage_model[f"input_{entry['modality']}"] = value

if "input" in usage_model:
usage_model["input"] = max(0, usage_model["input"] - value)

if "candidates_tokens_details" in usage_model:
candidates_tokens_details = usage_model.pop("candidates_tokens_details", [])

for entry in candidates_tokens_details:
if (
isinstance(entry, dict)
and "modality" in entry
and "token_count" in entry
):
value = entry["token_count"]
usage_model[f"output_{entry['modality']}"] = value

if "output" in usage_model:
usage_model["output"] = max(0, usage_model["output"] - value)

_ = usage_model.pop("cache_tokens_details", [])

return usage_model if usage_model else None


Expand All @@ -1131,7 +1164,7 @@ def _parse_usage(response: LLMResult):
for generation in response.generations:
for generation_chunk in generation:
if generation_chunk.generation_info and (
"usage_metadata" in generation_chunk.generation_info
generation_chunk.generation_info.get("usage_metadata", None)
):
llm_usage = _parse_usage_model(
generation_chunk.generation_info["usage_metadata"]
Expand Down