ref: Remove log and use on_conflict_do_nothing

src/seer/grouping/grouping.py

@@ -11,7 +11,6 @@
 from pydantic import BaseModel, ValidationInfo, field_validator
 from sentence_transformers import SentenceTransformer
 from sqlalchemy.dialects.postgresql import insert
-from sqlalchemy.exc import IntegrityError
 from torch.cuda import OutOfMemoryError
 
 from seer.db import DbGroupingRecord, Session
@@ -481,48 +480,20 @@ def insert_new_grouping_record(self, issue: GroupingRequest, embedding: np.ndarr
  :param embedding: The embedding of the stacktrace.
  """
  with Session() as session:
- existing_record = (
- session.query(DbGroupingRecord)
- .filter_by(hash=issue.hash, project_id=issue.project_id)
- .first()
+ insert_stmt = insert(DbGroupingRecord).values(
+ project_id=issue.project_id,
+ message=issue.message,
+ stacktrace_embedding=embedding,
+ hash=issue.hash,
+ error_type=issue.exception_type,
  )
 
- extra = {
- "project_id": issue.project_id,
- "stacktrace_length": len(issue.stacktrace),
- "input_hash": issue.hash,
- }
-
- if existing_record is None:
- new_record = GroupingRecord(
- project_id=issue.project_id,
- message=issue.message,
- stacktrace_embedding=embedding,
- hash=issue.hash,
- error_type=issue.exception_type,
- ).to_db_model()
- session.add(new_record)
-
- try:
- session.commit()
- except IntegrityError:
- session.rollback()
- existing_record = (
- session.query(DbGroupingRecord)
- .filter_by(hash=issue.hash, project_id=issue.project_id)
- .first()
- )
- extra["existing_hash"] = existing_record.hash
- logger.info(
- "group_already_exists_in_seer_db",
- extra=extra,
- )
- else:
- extra["existing_hash"] = existing_record.hash
- logger.info(
- "group_already_exists_in_seer_db",
- extra=extra,
+ session.execute(
+ insert_stmt.on_conflict_do_nothing(
+ index_elements=(DbGroupingRecord.project_id, DbGroupingRecord.hash)
  )
+ )
+ session.commit()
 
  @sentry_sdk.tracing.trace
  def delete_grouping_records_for_project(self, project_id: int) -> bool: