feat(cursor-origin): Track pull requests

src/sentry/integrations/cursor_origin/authors.py

@@ -0,0 +1,17 @@
+from __future__ import annotations
+
+from sentry.models.commitauthor import CommitAuthor
+
+MAX_AUTHOR_EMAIL_LENGTH = 75
+
+
+def get_or_create_commit_author(organization_id: int, email: str, name: str) -> CommitAuthor | None:
+    if not email or len(email) > MAX_AUTHOR_EMAIL_LENGTH:
+        return None
+
+    author, _ = CommitAuthor.objects.get_or_create(
+        organization_id=organization_id,
+        email=email,
+        defaults={"name": name[:128]},
+    )
+    return author

src/sentry/integrations/cursor_origin/pull_request.py

@@ -0,0 +1,96 @@
+from __future__ import annotations
+
+import logging
+from collections.abc import Mapping, Sequence
+from typing import Any
+
+from sentry.integrations.cursor_origin.authors import get_or_create_commit_author
+from sentry.integrations.cursor_origin.handlers import WebhookEventHandler
+from sentry.integrations.cursor_origin.repository import active_repositories
+from sentry.integrations.cursor_origin.webhook_types import PullRequestEvent
+from sentry.integrations.services.integration.model import (
+    RpcIntegration,
+    RpcOrganizationIntegration,
+)
+from sentry.integrations.types import IntegrationProviderSlug
+from sentry.integrations.utils.metrics import IntegrationWebhookEventType
+from sentry.models.pullrequest import PullRequestLifecycleState
+from sentry.models.repository import Repository
+from sentry.pr_metrics.lifecycle_mapping import update_pull_request_from_scm_snapshot
+from sentry.utils import metrics
+
+logger = logging.getLogger("sentry.integrations.cursor_origin")
+
+
+def lifecycle_state(event: PullRequestEvent) -> str:
+    if event.pull_request.merged:
+        return PullRequestLifecycleState.MERGED
+    if event.pull_request.state == "closed":
+        return PullRequestLifecycleState.CLOSED
+    return PullRequestLifecycleState.OPEN
+
+
+class PullRequestLifecycleHandler(WebhookEventHandler):
+    """Record a pull request from any of Origin's snapshot events."""
+
+    EVENT_TYPE = IntegrationWebhookEventType.MERGE_REQUEST
+
+    def __call__(
+        self,
+        payload: Mapping[str, Any],
+        delivery_id: str,
+        integration: RpcIntegration,
+        org_integrations: Sequence[RpcOrganizationIntegration],
+    ) -> None:
+        event = PullRequestEvent.from_payload(payload)
+
+        repositories = active_repositories(event.repository_id, org_integrations)
+        if not repositories:
+            logger.info(
+                "cursor_origin.pull_request.unknown_repository",
+                extra={"delivery_id": delivery_id, "external_id": event.repository_id},
+            )
+            metrics.incr("cursor_origin.pull_request.unknown_repository", sample_rate=1.0)
+            return
+
+        for repo in repositories:
+            self._record(repo, event, delivery_id)
+
+    def _record(self, repo: Repository, event: PullRequestEvent, delivery_id: str) -> None:
+        pull_request = event.pull_request
+        state = lifecycle_state(event)
+        email, name = pull_request.author.email_and_name()
+        author = get_or_create_commit_author(repo.organization_id, email, name)
+
+        _, created = update_pull_request_from_scm_snapshot(
+            provider=IntegrationProviderSlug.CURSOR_ORIGIN.value,
+            organization_id=repo.organization_id,
+            repository_id=repo.id,
+            key=pull_request.number,
+            defaults={
+                "organization_id": repo.organization_id,
+                "title": pull_request.title,
+                "author": author,
+                "message": pull_request.body,
+                "merge_commit_sha": pull_request.merge_commit_sha or None,
+                "head_commit_sha": pull_request.head.sha,
+                "opened_at": pull_request.created_at,
+                "closed_at": pull_request.closed_at,
+                "merged_at": pull_request.merged_at,
+                "provider_updated_at": pull_request.updated_at,
+                "state": state,
+                "draft": pull_request.draft,
+            },
+            event_state=state,
+            event_updated_at=pull_request.updated_at,
+        )
+
+        if created:
+            logger.info(
+                "cursor_origin.pull_request.created",
+                extra={
+                    "delivery_id": delivery_id,
+                    "repository_id": repo.id,
+                    "key": pull_request.number,
+                },
+            )

src/sentry/integrations/cursor_origin/push.py

@@ -9,6 +9,7 @@
 
 from sentry import options
 from sentry.constants import ObjectStatus
+from sentry.integrations.cursor_origin.authors import get_or_create_commit_author
 from sentry.integrations.cursor_origin.client import CursorOriginApiClient
 from sentry.integrations.cursor_origin.handlers import WebhookEventHandler
 from sentry.integrations.cursor_origin.repository import (
@@ -23,7 +24,6 @@
 )
 from sentry.integrations.utils.metrics import IntegrationWebhookEventType
 from sentry.models.commit import Commit
-from sentry.models.commitauthor import CommitAuthor
 from sentry.models.commitfilechange import CommitFileChange, post_bulk_create
 from sentry.models.repository import Repository
 from sentry.plugins.providers.integration_repository import IntegrationRepositoryProvider
@@ -33,8 +33,6 @@
 
 EMPTY_SHA = "0" * 40
 
-MAX_AUTHOR_EMAIL_LENGTH = 75
-
 
 class RepositoryPushedHandler(WebhookEventHandler):
     """Record the commits a push added.
@@ -155,7 +153,9 @@ def _record(
         if Commit.objects.filter(repository_id=repo.id, key=commit.sha).exists():
             return
 
-        author = self._author(repo, commit)
+        author = get_or_create_commit_author(
+            repo.organization_id, commit.author_email, commit.author_name
+        )
         changes = file_changes_from(client.get_commit_files(repo.config["name"], commit.sha))
 
         with transaction.atomic(router.db_for_write(Commit)):
@@ -184,14 +184,3 @@ def _record(
             if rows:
                 CommitFileChange.objects.bulk_create(rows, ignore_conflicts=True)
                 post_bulk_create(rows)
-
-    def _author(self, repo: Repository, commit: PushedCommit) -> CommitAuthor | None:
-        if not commit.author_email or len(commit.author_email) > MAX_AUTHOR_EMAIL_LENGTH:
-            return None
-
-        commit_author, _ = CommitAuthor.objects.get_or_create(
-            organization_id=repo.organization_id,
-            email=commit.author_email,
-            defaults={"name": commit.author_name[:128]},
-        )
-        return commit_author

src/sentry/integrations/cursor_origin/webhook.py

@@ -32,6 +32,7 @@
     WebhookEventHandler,
 )
 from sentry.integrations.cursor_origin.keys import signing_keys_for
+from sentry.integrations.cursor_origin.pull_request import PullRequestLifecycleHandler
 from sentry.integrations.cursor_origin.push import RepositoryPushedHandler
 from sentry.integrations.cursor_origin.repository_events import RepositoryMetadataUpdatedHandler
 from sentry.integrations.cursor_origin.webhook_types import OriginPayloadError
@@ -140,6 +141,12 @@ def verify_delivery(request: HttpRequest, body: bytes) -> Verification:
     "installation.suspended": InstallationRemovedHandler,
     "installation.unsuspended": InstallationRestoredHandler,
     "installation.updated": InstallationUpdatedHandler,
+    "pull_request.closed": PullRequestLifecycleHandler,
+    "pull_request.created": PullRequestLifecycleHandler,
+    "pull_request.merged": PullRequestLifecycleHandler,
+    "pull_request.metadata.updated": PullRequestLifecycleHandler,
+    "pull_request.published": PullRequestLifecycleHandler,
+    "pull_request.reopened": PullRequestLifecycleHandler,
     "repository.metadata.updated": RepositoryMetadataUpdatedHandler,
     "repository.pushed": RepositoryPushedHandler,
 }

src/sentry/integrations/cursor_origin/webhook_types.py

@@ -2,7 +2,7 @@
 
 from collections.abc import Mapping
 from datetime import datetime
-from typing import Any
+from typing import Any, Literal
 
 from pydantic import BaseModel, Field, ValidationError, validator
 
@@ -111,3 +111,77 @@ def from_payload(cls, payload: Mapping[str, Any]) -> RepositoryMetadataEvent:
             return cls.parse_obj(payload)
         except ValidationError as e:
             raise OriginPayloadError(str(e)) from e
+
+
+class PullRequestHead(OriginModel):
+    sha: str = Field(min_length=1)
+
+
+class PullRequestUser(OriginModel):
+    email: str = Field(min_length=1)
+    display_name: str = Field(default="", alias="displayName")
+
+
+class PullRequestApp(OriginModel):
+    id: str = Field(min_length=1)
+    display_name: str = Field(default="", alias="displayName")
+
+
+class PullRequestServiceAccount(OriginModel):
+    id: str = Field(min_length=1)
+
+
+class PullRequestAuthor(OriginModel):
+    """Exactly one of a user, an app or a service account."""
+
+    user: PullRequestUser | None = None
+    app: PullRequestApp | None = None
+    service_account: PullRequestServiceAccount | None = Field(default=None, alias="serviceAccount")
+
+    def email_and_name(self) -> tuple[str, str]:
+        if self.user is not None:
+            return self.user.email, self.user.display_name
+        if self.app is not None:
+            return f"{self.app.id}@localhost", self.app.display_name or self.app.id
+        assert self.service_account is not None
+        return f"{self.service_account.id}@localhost", self.service_account.id
+
+
+class PullRequest(OriginModel):
+    """Origin's own id is not kept: it is a prefixed string, and
+    `PullRequest.external_id` is an integer column."""
+
+    number: str = Field(min_length=1)
+    title: str
+    body: str
+    state: Literal["open", "closed"]
+    draft: bool
+    merged: bool
+    head: PullRequestHead
+    merge_commit_sha: str = Field(alias="mergeCommitSha")
+    author: PullRequestAuthor
+    created_at: datetime | None = Field(..., alias="createdAt")
+    updated_at: datetime | None = Field(..., alias="updatedAt")
+    closed_at: datetime | None = Field(..., alias="closedAt")
+    merged_at: datetime | None = Field(..., alias="mergedAt")
+
+    @validator("created_at", "updated_at", "closed_at", "merged_at", pre=True)
+    def _absent_date(cls, value: Any) -> Any:
+        """Origin sends an empty string for an unset date."""
+        return value or None
+
+
+class PullRequestEvent(OriginModel):
+    repository: Repository
+    pull_request: PullRequest = Field(alias="pullRequest")
+
+    @property
+    def repository_id(self) -> str:
+        return self.repository.id
+
+    @classmethod
+    def from_payload(cls, payload: Mapping[str, Any]) -> PullRequestEvent:
+        try:
+            return cls.parse_obj(payload)
+        except ValidationError as e:
+            raise OriginPayloadError(str(e)) from e