Use a global cache not thread-local one

Author: JBorrow

librarian_server/api/validate.py

@@ -5,6 +5,7 @@
 """
 
 import asyncio
+import datetime
 from functools import lru_cache
 from pathlib import Path
 from time import perf_counter
@@ -38,8 +39,44 @@
 
 router = APIRouter(prefix="/api/v2/validate")
 
+VALIDATION_TIMEOUT = datetime.timedelta(hours=8)
+VALIDATION_CACHE = {}
+
+
+async def cached_calculate_checksum_of_local_copy(
+    original_checksum: str,
+    original_size: int,
+    path_info_function: callable,
+    path: Path,
+    store_id: int,
+    instance_id: int,
+):
+    key = f"{original_checksum}-{instance_id}"
+
+    cached = VALIDATION_CACHE.get(key, None)
+
+    if cached is None or (
+        (datetime.datetime.now(datetime.timezone.utc) - cached[1]) > VALIDATION_TIMEOUT
+    ):
+        result = await asyncify(calculate_checksum_of_local_copy)(
+            original_checksum=original_checksum,
+            original_size=original_size,
+            path_info_function=path_info_function,
+            path=path,
+            store_id=store_id,
+            instance_id=instance_id,
+        )
+
+        VALIDATION_CACHE[key] = (result, datetime.datetime.now(datetime.timezone.utc))
+    else:
+        log.info(
+            f"Using cached result for instance {instance_id}", instance_id=instance_id
+        )
+        result = cached[0]
+
+    return result
+
 
-@lru_cache(maxsize=1024)
 def calculate_checksum_of_local_copy(
     original_checksum: str,
     original_size: int,
@@ -187,7 +224,7 @@ async def validate_file(
         if not instance.available:
             continue
 
-        this_checksum_info = asyncify(calculate_checksum_of_local_copy)(
+        this_checksum_info = cached_calculate_checksum_of_local_copy(
             original_checksum=file.checksum,
             original_size=file.size,
             path_info_function=instance.store.store_manager.path_info,

pyproject.toml

@@ -7,7 +7,7 @@ exclude=["*tests*"]
 [project]
 name="hera_librarian"
 requires-python = ">=3.10"
-version = "3.1.0"
+version = "3.1.1"
 dependencies = [
     "alembic",
     "argon2-cffi",

tests/integration_test/test_send_queue.py

@@ -371,6 +371,10 @@ def test_send_from_existing_file_row(
     # Should have _ours_ and _theirs_.
     assert len(instance_validations) == 2
 
+    # Check again (should use the cache)!
+    instance_validations = mocked_admin_client.validate_file(file_name=file_name)
+    assert len(instance_validations) == 2
+
     source_librarians_for_validations = {x.librarian for x in instance_validations}
 
     assert len(source_librarians_for_validations) == 2  # I.e. they are different