• fix(well-inventory): make CSV import reruns idempotent

- Detect previously imported well inventory rows before inserting related records
- Skip recreating field activity water-level samples and observations when the same row is reprocessed
- Return serializable existing-row results so CLI reruns report cleanly instead of crashing

Author: ksmuczynski

services/well_inventory_csv.py

@@ -438,6 +438,44 @@ def _generate_autogen_well_id(session, prefix: str, offset: int = 0) -> tuple[st
     return f"{prefix}{new_number:04d}", new_number
 
 
+def _find_existing_imported_well(
+    session: Session, model: WellInventoryRow
+) -> Thing | None:
+    if model.measurement_date_time is not None:
+        sample_name = (
+            f"{model.well_name_point_id}-WL-"
+            f"{model.measurement_date_time.strftime('%Y%m%d%H%M')}"
+        )
+        existing = session.scalars(
+            select(Thing)
+            .join(FieldEvent, FieldEvent.thing_id == Thing.id)
+            .join(FieldActivity, FieldActivity.field_event_id == FieldEvent.id)
+            .join(Sample, Sample.field_activity_id == FieldActivity.id)
+            .where(
+                Thing.name == model.well_name_point_id,
+                Thing.thing_type == "water well",
+                FieldActivity.activity_type == "well inventory",
+                Sample.sample_name == sample_name,
+            )
+            .order_by(Thing.id.asc())
+        ).first()
+        if existing is not None:
+            return existing
+
+    return session.scalars(
+        select(Thing)
+        .join(FieldEvent, FieldEvent.thing_id == Thing.id)
+        .join(FieldActivity, FieldActivity.field_event_id == FieldEvent.id)
+        .where(
+            Thing.name == model.well_name_point_id,
+            Thing.thing_type == "water well",
+            FieldEvent.event_date == model.date_time,
+            FieldActivity.activity_type == "well inventory",
+        )
+        .order_by(Thing.id.asc())
+    ).first()
+
+
 def _make_row_models(rows, session):
     models = []
     validation_errors = []
@@ -542,6 +580,10 @@ def _add_csv_row(session: Session, group: Group, model: WellInventoryRow, user)
     name = model.well_name_point_id
     date_time = model.date_time
 
+    existing_well = _find_existing_imported_well(session, model)
+    if existing_well is not None:
+        return existing_well.name
+
     # --------------------
     # Location and associated tables
     # --------------------

tests/test_well_inventory.py

@@ -494,6 +494,38 @@ def test_blank_depth_to_water_still_creates_water_level_records(tmp_path):
         assert observations[0].measuring_point_height == 2.5
 
 
+def test_rerunning_same_well_inventory_csv_is_idempotent():
+    """Re-importing the same CSV should not create duplicate well inventory records."""
+    file = Path("tests/features/data/well-inventory-valid.csv")
+    assert file.exists(), "Test data file does not exist."
+
+    first = well_inventory_csv(file)
+    assert first.exit_code == 0, first.stderr
+
+    with session_ctx() as session:
+        counts_after_first = {
+            "things": session.query(Thing).count(),
+            "field_events": session.query(FieldEvent).count(),
+            "field_activities": session.query(FieldActivity).count(),
+            "samples": session.query(Sample).count(),
+            "observations": session.query(Observation).count(),
+        }
+
+    second = well_inventory_csv(file)
+    assert second.exit_code == 0, second.stderr
+
+    with session_ctx() as session:
+        counts_after_second = {
+            "things": session.query(Thing).count(),
+            "field_events": session.query(FieldEvent).count(),
+            "field_activities": session.query(FieldActivity).count(),
+            "samples": session.query(Sample).count(),
+            "observations": session.query(Observation).count(),
+        }
+
+    assert counts_after_second == counts_after_first
+
+
 # =============================================================================
 # Error Handling Tests - Cover API error paths
 # =============================================================================