move event_content_is_seer_eligible to ingest seer module

Author: lobsterkatie

src/sentry/grouping/ingest/seer.py

@@ -9,16 +9,17 @@
 from sentry import options
 from sentry import ratelimits as ratelimiter
 from sentry.conf.server import SEER_SIMILARITY_MODEL_VERSION
-from sentry.eventstore.models import Event
+from sentry.eventstore.models import Event, GroupEvent
 from sentry.grouping.grouping_info import get_grouping_info_from_variants
 from sentry.grouping.variants import BaseVariant
 from sentry.models.grouphash import GroupHash
 from sentry.models.project import Project
 from sentry.seer.similarity.similar_issues import get_similarity_data_from_seer
 from sentry.seer.similarity.types import SimilarIssuesEmbeddingsRequest
 from sentry.seer.similarity.utils import (
+    SEER_INELIGIBLE_EVENT_PLATFORMS,
     ReferrerOptions,
-    event_content_is_seer_eligible,
+    event_content_has_stacktrace,
     filter_null_from_string,
     get_stacktrace_string_with_metrics,
     killswitch_enabled,
@@ -40,7 +41,7 @@ def should_call_seer_for_grouping(event: Event, variants: Mapping[str, BaseVaria
     project = event.project
 
     # Check both of these before returning based on either so we can gather metrics on their results
-    content_is_eligible = event_content_is_seer_eligible(event)
+    content_is_eligible = _event_content_is_seer_eligible(event)
     seer_enabled_for_project = _project_has_similarity_grouping_enabled(project)
     if not (content_is_eligible and seer_enabled_for_project):
         return False
@@ -67,6 +68,34 @@ def should_call_seer_for_grouping(event: Event, variants: Mapping[str, BaseVaria
     return True
 
 
+def _event_content_is_seer_eligible(event: GroupEvent | Event) -> bool:
+    """
+    Determine if an event's contents makes it fit for using with Seer's similar issues model.
+    """
+    if not event_content_has_stacktrace(event):
+        metrics.incr(
+            "grouping.similarity.event_content_seer_eligible",
+            sample_rate=options.get("seer.similarity.metrics_sample_rate"),
+            tags={"eligible": False, "blocker": "no-stacktrace"},
+        )
+        return False
+
+    if event.platform in SEER_INELIGIBLE_EVENT_PLATFORMS:
+        metrics.incr(
+            "grouping.similarity.event_content_seer_eligible",
+            sample_rate=options.get("seer.similarity.metrics_sample_rate"),
+            tags={"eligible": False, "blocker": "unsupported-platform"},
+        )
+        return False
+
+    metrics.incr(
+        "grouping.similarity.event_content_seer_eligible",
+        sample_rate=options.get("seer.similarity.metrics_sample_rate"),
+        tags={"eligible": True, "blocker": "none"},
+    )
+    return True
+
+
 def _project_has_similarity_grouping_enabled(project: Project) -> bool:
     # TODO: This is a hack to get ingest to turn on for projects as soon as they're backfilled. When
     # the backfill script completes, we turn on this option, enabling ingest immediately rather than

src/sentry/seer/similarity/utils.py

@@ -319,35 +319,6 @@ def event_content_has_stacktrace(event: GroupEvent | Event) -> bool:
     return exception_stacktrace or threads_stacktrace or only_stacktrace
 
 
-def event_content_is_seer_eligible(event: GroupEvent | Event) -> bool:
-    """
-    Determine if an event's contents makes it fit for using with Seer's similar issues model.
-    """
-    # TODO: Determine if we want to filter out non-sourcemapped events
-    if not event_content_has_stacktrace(event):
-        metrics.incr(
-            "grouping.similarity.event_content_seer_eligible",
-            sample_rate=options.get("seer.similarity.metrics_sample_rate"),
-            tags={"eligible": False, "blocker": "no-stacktrace"},
-        )
-        return False
-
-    if event.platform in SEER_INELIGIBLE_EVENT_PLATFORMS:
-        metrics.incr(
-            "grouping.similarity.event_content_seer_eligible",
-            sample_rate=options.get("seer.similarity.metrics_sample_rate"),
-            tags={"eligible": False, "blocker": "unsupported-platform"},
-        )
-        return False
-
-    metrics.incr(
-        "grouping.similarity.event_content_seer_eligible",
-        sample_rate=options.get("seer.similarity.metrics_sample_rate"),
-        tags={"eligible": True, "blocker": "none"},
-    )
-    return True
-
-
 def record_did_call_seer_metric(*, call_made: bool, blocker: str) -> None:
     metrics.incr(
         "grouping.similarity.did_call_seer",

tests/sentry/event_manager/grouping/test_seer_grouping.py

@@ -67,7 +67,7 @@ def test_obeys_seer_similarity_flags(self):
                 return_value=[seer_result_data],
             ),
             patch(
-                "sentry.grouping.ingest.seer.event_content_is_seer_eligible",
+                "sentry.grouping.ingest.seer._event_content_is_seer_eligible",
                 return_value=True,
             ),
         ):

tests/sentry/grouping/ingest/test_seer.py

@@ -1,18 +1,21 @@
 from dataclasses import asdict
 from time import time
+from typing import Any
 from unittest.mock import ANY, MagicMock, Mock, call, patch
+from uuid import uuid1
 
 from sentry import options
 from sentry.conf.server import SEER_SIMILARITY_MODEL_VERSION
 from sentry.eventstore.models import Event
 from sentry.grouping.ingest.seer import (
+    _event_content_is_seer_eligible,
     get_seer_similar_issues,
     maybe_check_seer_for_matching_grouphash,
     should_call_seer_for_grouping,
 )
 from sentry.models.grouphash import GroupHash
 from sentry.seer.similarity.types import SeerSimilarIssueData
-from sentry.seer.similarity.utils import MAX_FRAME_COUNT
+from sentry.seer.similarity.utils import MAX_FRAME_COUNT, SEER_INELIGIBLE_EVENT_PLATFORMS
 from sentry.testutils.cases import TestCase
 from sentry.testutils.helpers.eventprocessing import save_new_event
 from sentry.testutils.helpers.options import override_options
@@ -65,7 +68,7 @@ def test_obeys_content_filter(self) -> None:
 
         for content_eligibility, expected_result in [(True, True), (False, False)]:
             with patch(
-                "sentry.grouping.ingest.seer.event_content_is_seer_eligible",
+                "sentry.grouping.ingest.seer._event_content_is_seer_eligible",
                 return_value=content_eligibility,
             ):
                 assert should_call_seer_for_grouping(self.event, self.variants) is expected_result
@@ -603,3 +606,68 @@ def test_too_many_only_system_frames_maybe_check_seer_for_matching_group_hash_in
                 "use_reranking": True,
             }
         )
+
+
+class EventContentIsSeerEligibleTest(TestCase):
+    def get_eligible_event_data(self) -> dict[str, Any]:
+        return {
+            "title": "FailedToFetchError('Charlie didn't bring the ball back')",
+            "exception": {
+                "values": [
+                    {
+                        "type": "FailedToFetchError",
+                        "value": "Charlie didn't bring the ball back",
+                        "stacktrace": {
+                            "frames": [
+                                {
+                                    "function": "play_fetch",
+                                    "filename": "dogpark.py",
+                                    "context_line": "raise FailedToFetchError('Charlie didn't bring the ball back')",
+                                }
+                            ]
+                        },
+                    }
+                ]
+            },
+            "platform": "python",
+        }
+
+    def test_no_stacktrace(self):
+        good_event_data = self.get_eligible_event_data()
+        good_event = Event(
+            project_id=self.project.id,
+            event_id=uuid1().hex,
+            data=good_event_data,
+        )
+
+        bad_event_data = self.get_eligible_event_data()
+        del bad_event_data["exception"]
+        bad_event = Event(
+            project_id=self.project.id,
+            event_id=uuid1().hex,
+            data=bad_event_data,
+        )
+
+        assert _event_content_is_seer_eligible(good_event) is True
+        assert _event_content_is_seer_eligible(bad_event) is False
+
+    def test_platform_filter(self):
+        good_event_data = self.get_eligible_event_data()
+        good_event = Event(
+            project_id=self.project.id,
+            event_id=uuid1().hex,
+            data=good_event_data,
+        )
+
+        bad_event_data = self.get_eligible_event_data()
+        bad_event_data["platform"] = "other"
+        bad_event = Event(
+            project_id=self.project.id,
+            event_id=uuid1().hex,
+            data=bad_event_data,
+        )
+
+        assert good_event_data["platform"] not in SEER_INELIGIBLE_EVENT_PLATFORMS
+        assert bad_event_data["platform"] in SEER_INELIGIBLE_EVENT_PLATFORMS
+        assert _event_content_is_seer_eligible(good_event) is True
+        assert _event_content_is_seer_eligible(bad_event) is False

tests/sentry/seer/similarity/test_utils.py

@@ -2,19 +2,15 @@
 from collections.abc import Callable
 from typing import Any, Literal, cast
 from unittest.mock import patch
-from uuid import uuid1
 
 import pytest
 
-from sentry.eventstore.models import Event
 from sentry.seer.similarity.utils import (
     BASE64_ENCODED_PREFIXES,
     MAX_FRAME_COUNT,
-    SEER_INELIGIBLE_EVENT_PLATFORMS,
     ReferrerOptions,
     TooManyOnlySystemFramesException,
     _is_snipped_context_line,
-    event_content_is_seer_eligible,
     filter_null_from_string,
     get_stacktrace_string,
     get_stacktrace_string_with_metrics,
@@ -875,71 +871,6 @@ def test_no_header_one_frame_no_filename(self, mock_metrics):
         assert get_stacktrace_string(exception) == ""
 
 
-class EventContentIsSeerEligibleTest(TestCase):
-    def get_eligible_event_data(self) -> dict[str, Any]:
-        return {
-            "title": "FailedToFetchError('Charlie didn't bring the ball back')",
-            "exception": {
-                "values": [
-                    {
-                        "type": "FailedToFetchError",
-                        "value": "Charlie didn't bring the ball back",
-                        "stacktrace": {
-                            "frames": [
-                                {
-                                    "function": "play_fetch",
-                                    "filename": "dogpark.py",
-                                    "context_line": "raise FailedToFetchError('Charlie didn't bring the ball back')",
-                                }
-                            ]
-                        },
-                    }
-                ]
-            },
-            "platform": "python",
-        }
-
-    def test_no_stacktrace(self):
-        good_event_data = self.get_eligible_event_data()
-        good_event = Event(
-            project_id=self.project.id,
-            event_id=uuid1().hex,
-            data=good_event_data,
-        )
-
-        bad_event_data = self.get_eligible_event_data()
-        del bad_event_data["exception"]
-        bad_event = Event(
-            project_id=self.project.id,
-            event_id=uuid1().hex,
-            data=bad_event_data,
-        )
-
-        assert event_content_is_seer_eligible(good_event) is True
-        assert event_content_is_seer_eligible(bad_event) is False
-
-    def test_platform_filter(self):
-        good_event_data = self.get_eligible_event_data()
-        good_event = Event(
-            project_id=self.project.id,
-            event_id=uuid1().hex,
-            data=good_event_data,
-        )
-
-        bad_event_data = self.get_eligible_event_data()
-        bad_event_data["platform"] = "other"
-        bad_event = Event(
-            project_id=self.project.id,
-            event_id=uuid1().hex,
-            data=bad_event_data,
-        )
-
-        assert good_event_data["platform"] not in SEER_INELIGIBLE_EVENT_PLATFORMS
-        assert bad_event_data["platform"] in SEER_INELIGIBLE_EVENT_PLATFORMS
-        assert event_content_is_seer_eligible(good_event) is True
-        assert event_content_is_seer_eligible(bad_event) is False
-
-
 class SeerUtilsTest(TestCase):
     def test_filter_null_from_string(self):
         string_with_null = 'String with null \x00, "\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00" is null'