Skip to content

fix(anthropic): Gate gen_ai.response.tool_calls on outputs, not inputs #7207

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
ericapisani wants to merge 1 commit into from
+9 −48
Open

fix(anthropic): Gate gen_ai.response.tool_calls on outputs, not inputs #7207

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
9 changes: 3 additions & 6 deletions sentry_sdk/integrations/anthropic.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -608,16 +608,13 @@ def _set_output_data(
set_on_span(SPANDATA.GEN_AI_RESPONSE_FINISH_REASONS, [finish_reason])

client = sentry_sdk.get_client()
record_inputs = False
record_outputs = False
if has_data_collection_enabled(client.options):
record_inputs = client.options["data_collection"]["gen_ai"]["inputs"]
record_outputs = client.options["data_collection"]["gen_ai"]["outputs"]
elif should_send_default_pii() and integration.include_prompts:
record_inputs = True
record_outputs = True

if record_inputs or record_outputs:
if record_outputs:
output_messages: "dict[str, list[Any]]" = {
"response": [],
"tool": [],
Expand All @@ -629,15 +626,15 @@ def _set_output_data(
elif output["type"] == "tool_use":
output_messages["tool"].append(output)

if record_inputs and len(output_messages["tool"]) > 0:
if len(output_messages["tool"]) > 0:
set_data_normalized(
span,
SPANDATA.GEN_AI_RESPONSE_TOOL_CALLS,
output_messages["tool"],
unpack=False,
)

if record_outputs and len(output_messages["response"]) > 0:
if len(output_messages["response"]) > 0:
set_data_normalized(
span, SPANDATA.GEN_AI_RESPONSE_TEXT, output_messages["response"]
)
Expand Down
48 changes: 6 additions & 42 deletions tests/integrations/anthropic/test_anthropic.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -639,62 +639,48 @@ async def test_nonstreaming_create_message_data_collection_async(
@pytest.mark.parametrize("span_streaming", [True, False])
@pytest.mark.parametrize("stream_gen_ai_spans", [True, False])
@pytest.mark.parametrize(
"data_collection,send_default_pii,include_prompts,outputs_collected,tool_calls_collected",
"data_collection,send_default_pii,include_prompts,outputs_collected",
[
pytest.param(
{"gen_ai": {"outputs": True}},
False,
False,
True,
True,
id="gen-ai-outputs-enabled-overrides-pii-and-include-prompts",
),
pytest.param(
{"gen_ai": {"outputs": False}},
True,
True,
False,
True,
id="gen-ai-outputs-disabled-still-collects-tool-calls-gated-on-inputs",
),
pytest.param(
{"gen_ai": {"inputs": False}},
True,
True,
True,
False,
id="gen-ai-inputs-disabled-drops-tool-calls-only",
id="gen-ai-outputs-disabled-overrides-pii-and-include-prompts",
),
pytest.param(
{"gen_ai": {"inputs": False, "outputs": False}},
True,
True,
False,
False,
id="gen-ai-inputs-and-outputs-disabled-overrides-pii-and-include-prompts",
),
pytest.param(
{"gen_ai": {}},
False,
False,
True,
True,
id="gen-ai-inputs-and-outputs-omitted-defaults-to-enabled",
),
pytest.param(
None,
True,
True,
True,
True,
id="legacy-pii-and-include-prompts-enabled",
),
pytest.param(
None,
False,
True,
False,
False,
id="legacy-pii-disabled",
),
],
Expand All @@ -707,7 +693,6 @@ def test_nonstreaming_create_message_data_collection_outputs(
send_default_pii,
include_prompts,
outputs_collected,
tool_calls_collected,
stream_gen_ai_spans,
span_streaming,
):
Expand Down Expand Up @@ -765,15 +750,12 @@ def test_nonstreaming_create_message_data_collection_outputs(
span_data[SPANDATA.GEN_AI_RESPONSE_TEXT]
== DATA_COLLECTION_EXPECTED_RESPONSE_TEXT
)
else:
assert SPANDATA.GEN_AI_RESPONSE_TEXT not in span_data

if tool_calls_collected:
assert (
json.loads(span_data[SPANDATA.GEN_AI_RESPONSE_TOOL_CALLS])
== DATA_COLLECTION_EXPECTED_TOOL_CALLS
)
else:
assert SPANDATA.GEN_AI_RESPONSE_TEXT not in span_data
assert SPANDATA.GEN_AI_RESPONSE_TOOL_CALLS not in span_data


Expand All @@ -785,62 +767,48 @@ def test_nonstreaming_create_message_data_collection_outputs(
@pytest.mark.parametrize("stream_gen_ai_spans", [True, False])
@pytest.mark.asyncio
@pytest.mark.parametrize(
"data_collection,send_default_pii,include_prompts,outputs_collected,tool_calls_collected",
"data_collection,send_default_pii,include_prompts,outputs_collected",
[
pytest.param(
{"gen_ai": {"outputs": True}},
False,
False,
True,
True,
id="gen-ai-outputs-enabled-overrides-pii-and-include-prompts",
),
pytest.param(
{"gen_ai": {"outputs": False}},
True,
True,
False,
True,
id="gen-ai-outputs-disabled-still-collects-tool-calls-gated-on-inputs",
),
pytest.param(
{"gen_ai": {"inputs": False}},
True,
True,
True,
False,
id="gen-ai-inputs-disabled-drops-tool-calls-only",
id="gen-ai-outputs-disabled-overrides-pii-and-include-prompts",
),
pytest.param(
{"gen_ai": {"inputs": False, "outputs": False}},
True,
True,
False,
False,
id="gen-ai-inputs-and-outputs-disabled-overrides-pii-and-include-prompts",
),
pytest.param(
{"gen_ai": {}},
False,
False,
True,
True,
id="gen-ai-inputs-and-outputs-omitted-defaults-to-enabled",
),
pytest.param(
None,
True,
True,
True,
True,
id="legacy-pii-and-include-prompts-enabled",
),
pytest.param(
None,
False,
True,
False,
False,
id="legacy-pii-disabled",
),
],
Expand All @@ -853,7 +821,6 @@ async def test_nonstreaming_create_message_data_collection_outputs_async(
send_default_pii,
include_prompts,
outputs_collected,
tool_calls_collected,
stream_gen_ai_spans,
span_streaming,
):
Expand Down Expand Up @@ -911,15 +878,12 @@ async def test_nonstreaming_create_message_data_collection_outputs_async(
span_data[SPANDATA.GEN_AI_RESPONSE_TEXT]
== DATA_COLLECTION_EXPECTED_RESPONSE_TEXT
)
else:
assert SPANDATA.GEN_AI_RESPONSE_TEXT not in span_data

if tool_calls_collected:
assert (
json.loads(span_data[SPANDATA.GEN_AI_RESPONSE_TOOL_CALLS])
== DATA_COLLECTION_EXPECTED_TOOL_CALLS
)
else:
assert SPANDATA.GEN_AI_RESPONSE_TEXT not in span_data
assert SPANDATA.GEN_AI_RESPONSE_TOOL_CALLS not in span_data


Expand Down
Loading