Skip to content

fix(langchain): Record reasoning tokens as metric #2393

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

Open
Dylan Pulver (dylanpulver) wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fix(langchain): Record reasoning tokens as metric #2393

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
5 changes: 5 additions & 0 deletions .changeset/langchain-reasoning-tokens-metric.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,5 @@
---
"braintrust": patch
---

fix(langchain): Record reasoning tokens as metric
1 change: 1 addition & 0 deletions ...narios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.log-payloads.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -266,6 +266,7 @@
"tags": []
},
"metrics": {
"completion_reasoning_tokens": "<number>",
"completion_tokens": "<number>",
"end": 0,
"prompt_cached_tokens": "<number>",
Expand Down
1 change: 1 addition & 0 deletions ...scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.span-tree.json
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

1 change: 1 addition & 0 deletions e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.span-tree.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -113,6 +113,7 @@ span_tree:
│ "model": "gpt-4o-mini-2024-07-18"
│ }
│ metric_keys: [
│ "completion_reasoning_tokens",
│ "completion_tokens",
│ "prompt_cached_tokens",
│ "prompt_tokens",
Expand Down
1 change: 1 addition & 0 deletions e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.log-payloads.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -265,6 +265,7 @@
"tags": []
},
"metrics": {
"completion_reasoning_tokens": "<number>",
"completion_tokens": "<number>",
"end": 0,
"prompt_cached_tokens": "<number>",
Expand Down
1 change: 1 addition & 0 deletions e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.span-tree.json
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

1 change: 1 addition & 0 deletions e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.span-tree.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -113,6 +113,7 @@ span_tree:
│ "model": "gpt-4o-mini-2024-07-18"
│ }
│ metric_keys: [
│ "completion_reasoning_tokens",
│ "completion_tokens",
│ "prompt_cached_tokens",
│ "prompt_tokens",
Expand Down
6 changes: 6 additions & 0 deletions .../wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1-latest.span-tree.json
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

6 changes: 6 additions & 0 deletions ...s/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1-latest.span-tree.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -155,6 +155,7 @@ span_tree:
│ "tags": []
│ }
│ metrics: {
│ "completion_reasoning_tokens": 0,
│ "completion_tokens": 2,
│ "prompt_cached_tokens": 0,
│ "prompt_tokens": 12,
Expand Down Expand Up @@ -553,6 +554,7 @@ span_tree:
│ ]
│ }
│ metrics: {
│ "completion_reasoning_tokens": 0,
│ "completion_tokens": 2,
│ "prompt_cached_tokens": 0,
│ "prompt_tokens": 18,
Expand Down Expand Up @@ -716,6 +718,7 @@ span_tree:
│ "tags": []
│ }
│ metrics: {
│ "completion_reasoning_tokens": 0,
│ "completion_tokens": 6,
│ "prompt_cached_tokens": 0,
│ "prompt_tokens": 22,
Expand Down Expand Up @@ -940,6 +943,7 @@ span_tree:
│ "tags": []
│ }
│ metrics: {
│ "completion_reasoning_tokens": 0,
│ "completion_tokens": 16,
│ "prompt_cached_tokens": 0,
│ "prompt_tokens": 72,
Expand Down Expand Up @@ -1191,6 +1195,7 @@ span_tree:
│ "tags": []
│ }
│ metrics: {
│ "completion_reasoning_tokens": 0,
│ "completion_tokens": 21,
│ "prompt_cached_tokens": 0,
│ "prompt_tokens": 76,
Expand Down Expand Up @@ -1457,6 +1462,7 @@ span_tree:
"tags": []
}
metrics: {
"completion_reasoning_tokens": 0,
"completion_tokens": 11,
"prompt_cached_tokens": 0,
"prompt_tokens": 106,
Expand Down
6 changes: 6 additions & 0 deletions ...enarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1.span-tree.json
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

6 changes: 6 additions & 0 deletions ...cenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1.span-tree.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -155,6 +155,7 @@ span_tree:
│ "tags": []
│ }
│ metrics: {
│ "completion_reasoning_tokens": 0,
│ "completion_tokens": 2,
│ "prompt_cached_tokens": 0,
│ "prompt_tokens": 12,
Expand Down Expand Up @@ -553,6 +554,7 @@ span_tree:
│ ]
│ }
│ metrics: {
│ "completion_reasoning_tokens": 0,
│ "completion_tokens": 2,
│ "prompt_cached_tokens": 0,
│ "prompt_tokens": 18,
Expand Down Expand Up @@ -716,6 +718,7 @@ span_tree:
│ "tags": []
│ }
│ metrics: {
│ "completion_reasoning_tokens": 0,
│ "completion_tokens": 6,
│ "prompt_cached_tokens": 0,
│ "prompt_tokens": 22,
Expand Down Expand Up @@ -940,6 +943,7 @@ span_tree:
│ "tags": []
│ }
│ metrics: {
│ "completion_reasoning_tokens": 0,
│ "completion_tokens": 16,
│ "prompt_cached_tokens": 0,
│ "prompt_tokens": 72,
Expand Down Expand Up @@ -1191,6 +1195,7 @@ span_tree:
│ "tags": []
│ }
│ metrics: {
│ "completion_reasoning_tokens": 0,
│ "completion_tokens": 21,
│ "prompt_cached_tokens": 0,
│ "prompt_tokens": 76,
Expand Down Expand Up @@ -1457,6 +1462,7 @@ span_tree:
"tags": []
}
metrics: {
"completion_reasoning_tokens": 0,
"completion_tokens": 11,
"prompt_cached_tokens": 0,
"prompt_tokens": 106,
Expand Down
62 changes: 62 additions & 0 deletions js/src/wrappers/langchain/callback-handler.test.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -128,4 +128,66 @@ describe("BraintrustLangChainCallbackHandler metrics", () => {
tokens: 12,
});
});

it("preserves reasoning metrics from message usage metadata", async () => {
const { endLog } = await finishChatModelRun({
generations: [
[
{
message: {
usage_metadata: {
input_tokens: 10,
output_tokens: 20,
total_tokens: 30,
output_token_details: {
reasoning: 16,
},
},
},
},
],
],
});

expect(endLog.metrics).toEqual({
prompt_tokens: 10,
completion_tokens: 20,
completion_reasoning_tokens: 16,
tokens: 30,
});
});

it("reports zero reasoning tokens rather than dropping the field", async () => {
const { endLog } = await finishChatModelRun({
generations: [
[
{
message: {
usage_metadata: {
input_tokens: 12,
output_tokens: 2,
total_tokens: 14,
input_token_details: {
audio: 0,
cache_read: 0,
},
output_token_details: {
audio: 0,
reasoning: 0,
},
},
},
},
],
],
});

expect(endLog.metrics).toEqual({
prompt_tokens: 12,
completion_tokens: 2,
prompt_cached_tokens: 0,
completion_reasoning_tokens: 0,
tokens: 14,
});
});
});
4 changes: 4 additions & 0 deletions js/src/wrappers/langchain/callback-handler.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -523,6 +523,7 @@ function getMetricsFromResponse(
}

const inputTokenDetails = usageMetadata.input_token_details;
const outputTokenDetails = usageMetadata.output_token_details;
return normalizeTokenMetrics({
total_tokens: usageMetadata.total_tokens,
prompt_tokens: usageMetadata.input_tokens,
Expand All @@ -533,6 +534,9 @@ function getMetricsFromResponse(
prompt_cached_tokens: isRecord(inputTokenDetails)
? inputTokenDetails.cache_read
: undefined,
completion_reasoning_tokens: isRecord(outputTokenDetails)
? outputTokenDetails.reasoning
: undefined,
});
}

Expand Down