Download example databases for use in CI tests (#608)

* 💽 Create SQL test data and dockerfile spec

* Remove large binary file from history w/git-filter

* Rely on externally provided example data.

* Remove slow-test code, no longer needed.

* Skip test if example database does not exist.

* Docker container no longer hosted here

* Download, get and mount 💾 .sql data (empty)

* Correct mountpoint 🐳

* Namespaced test database deployment 📇

* Pin Postgres image to ensure compat with .sql dump.

* Use curl instead of wget; make script idempotent.

* Download sqlite data.

* Test against SQLite with example data

---------

Co-authored-by: Dan Allan <[email protected]>

.github/workflows/ci.yml

@@ -63,7 +63,7 @@ jobs:
       shell: bash -l {0}
       run: |
         set -vxeuo pipefail
-        coverage run -m pytest -v -m "not slow"
+        coverage run -m pytest -v
         coverage report
       env:
         # Provide test suite with a PostgreSQL database to use.

continuous_integration/scripts/download_sqlite_data.sh

@@ -0,0 +1,4 @@
+#!/bin/bash
+set -e
+
+curl -fsSLO https://github.com/bluesky/tiled-example-database/releases/latest/download/tiled_test_db_sqlite.db

continuous_integration/scripts/start_postgres.sh

@@ -1,5 +1,8 @@
 #!/bin/bash
 set -e
 
-docker run -d --rm --name tiled-test-postgres -p 5432:5432 -e POSTGRES_PASSWORD=secret -d docker.io/postgres
+# Try and get the latest postgres test data from the tiled-example-database repository
+curl -fsSLO https://github.com/bluesky/tiled-example-database/releases/latest/download/tiled_test_db_pg.sql
+
+docker run -d --rm --name tiled-test-postgres -p 5432:5432 -e POSTGRES_PASSWORD=secret -e POSTGRES_DB=tiled-example-data -v ./tiled_test_db_pg.sql:/docker-entrypoint-initdb.d/tiled_test_db_pg.sql -d docker.io/postgres:16
 docker ps

pytest.ini

@@ -5,6 +5,3 @@ log_cli = 1
 log_cli_level = WARNING
 log_cli_format = %(asctime)s [%(levelname)8s] %(message)s (%(filename)s:%(lineno)s)
 log_cli_date_format=%Y-%m-%d %H:%M:%S
-addopts = --strict-markers -m 'not slow'
-markers =
-    slow: marks tests as slow (deselect with '-m "not slow"')

tiled/_tests/conftest.py

@@ -3,8 +3,10 @@
 import tempfile
 from pathlib import Path
 
+import asyncpg
 import pytest
 import pytest_asyncio
+from sqlalchemy.ext.asyncio import create_async_engine
 
 from .. import profiles
 from ..catalog import from_uri, in_memory
@@ -97,37 +99,113 @@ def poll_enumerate():
 
 # To test with postgres, start a container like:
 #
-# docker run --name tiled-test-postgres -p 5432:5432 -e POSTGRES_PASSWORD=secret -d docker.io/postgres
+# docker run --name tiled-test-postgres -p 5432:5432 -e POSTGRES_PASSWORD=secret -d docker.io/postgres:16
 # and set this env var like:
 #
 # TILED_TEST_POSTGRESQL_URI=postgresql+asyncpg://postgres:secret@localhost:5432
 
 TILED_TEST_POSTGRESQL_URI = os.getenv("TILED_TEST_POSTGRESQL_URI")
 
 
-@pytest_asyncio.fixture(params=["sqlite", "postgresql"])
-async def adapter(request, tmpdir):
+@pytest_asyncio.fixture
+async def postgresql_adapter(request, tmpdir):
     """
     Adapter instance
 
     Note that startup() and shutdown() are not called, and must be run
     either manually (as in the fixture 'a') or via the app (as in the fixture 'client').
     """
-    if request.param == "sqlite":
-        adapter = in_memory(writable_storage=str(tmpdir))
+    if not TILED_TEST_POSTGRESQL_URI:
+        raise pytest.skip("No TILED_TEST_POSTGRESQL_URI configured")
+    # Create temporary database.
+    async with temp_postgres(TILED_TEST_POSTGRESQL_URI) as uri_with_database_name:
+        # Build an adapter on it, and initialize the database.
+        adapter = from_uri(
+            uri_with_database_name,
+            writable_storage=str(tmpdir),
+            init_if_not_exists=True,
+        )
         yield adapter
-    elif request.param == "postgresql":
-        if not TILED_TEST_POSTGRESQL_URI:
-            raise pytest.skip("No TILED_TEST_POSTGRESQL_URI configured")
-        # Create temporary database.
-        async with temp_postgres(TILED_TEST_POSTGRESQL_URI) as uri_with_database_name:
-            # Build an adapter on it, and initialize the database.
-            adapter = from_uri(
-                uri_with_database_name,
-                writable_storage=str(tmpdir),
-                init_if_not_exists=True,
-            )
-            yield adapter
-            await adapter.shutdown()
-    else:
-        assert False
+
+
+@pytest_asyncio.fixture
+async def postgresql_with_example_data_adapter(request, tmpdir):
+    """
+    Adapter instance
+
+    Note that startup() and shutdown() are not called, and must be run
+    either manually (as in the fixture 'a') or via the app (as in the fixture 'client').
+    """
+    if not TILED_TEST_POSTGRESQL_URI:
+        raise pytest.skip("No TILED_TEST_POSTGRESQL_URI configured")
+    DATABASE_NAME = "tiled-example-data"
+    uri = TILED_TEST_POSTGRESQL_URI
+    if uri.endswith("/"):
+        uri = uri[:-1]
+    uri_with_database_name = f"{uri}/{DATABASE_NAME}"
+    engine = create_async_engine(uri_with_database_name)
+    try:
+        async with engine.connect():
+            pass
+    except asyncpg.exceptions.InvalidCatalogNameError:
+        raise pytest.skip(
+            f"PostgreSQL instance contains no database named {DATABASE_NAME!r}"
+        )
+    adapter = from_uri(
+        uri_with_database_name,
+        writable_storage=str(tmpdir),
+    )
+    yield adapter
+
+
+@pytest_asyncio.fixture
+async def sqlite_adapter(request, tmpdir):
+    """
+    Adapter instance
+
+    Note that startup() and shutdown() are not called, and must be run
+    either manually (as in the fixture 'a') or via the app (as in the fixture 'client').
+    """
+    yield in_memory(writable_storage=str(tmpdir))
+
+
+@pytest_asyncio.fixture
+async def sqlite_with_example_data_adapter(request, tmpdir):
+    """
+    Adapter instance
+
+    Note that startup() and shutdown() are not called, and must be run
+    either manually (as in the fixture 'a') or via the app (as in the fixture 'client').
+    """
+    SQLITE_DATABASE_PATH = Path("./tiled_test_db_sqlite.db")
+    if not SQLITE_DATABASE_PATH.is_file():
+        raise pytest.skip(f"Could not find {SQLITE_DATABASE_PATH}")
+    adapter = from_uri(
+        f"sqlite+aiosqlite:///{SQLITE_DATABASE_PATH}",
+        writable_storage=str(tmpdir),
+    )
+    yield adapter
+
+
+@pytest.fixture(params=["sqlite_adapter", "postgresql_adapter"])
+def adapter(request):
+    """
+    Adapter instance
+
+    Note that startup() and shutdown() are not called, and must be run
+    either manually (as in the fixture 'a') or via the app (as in the fixture 'client').
+    """
+    yield request.getfixturevalue(request.param)
+
+
+@pytest.fixture(
+    params=["sqlite_with_example_data_adapter", "postgresql_with_example_data_adapter"]
+)
+def example_data_adapter(request):
+    """
+    Adapter instance
+
+    Note that startup() and shutdown() are not called, and must be run
+    either manually (as in the fixture 'a') or via the app (as in the fixture 'client').
+    """
+    yield request.getfixturevalue(request.param)

tiled/_tests/test_catalog.py

@@ -156,24 +156,14 @@ async def test_search(a):
     assert await d.search(Eq("number", 12)).keys_range(0, 5) == ["c"]
 
 
-@pytest.mark.slow
 @pytest.mark.asyncio
-async def test_metadata_index_is_used(a):
-    for i in range(10000):
-        await a.create_node(
-            metadata={
-                "number": i,
-                "number_as_string": str(i),
-                "nested": {"number": i, "number_as_string": str(i), "bool": bool(i)},
-                "bool": bool(i),
-            },
-            specs=[],
-            structure_family="array",
-        )
-    # Check that an index (specifically the 'top_level_metdata' index) is used
+async def test_metadata_index_is_used(example_data_adapter):
+    a = example_data_adapter  # for succinctness below
+    # Check that an index (specifically the 'top_level_metadata' index) is used
     # by inspecting the content of an 'EXPLAIN ...' query. The exact content
     # is intended for humans and is not an API, but we can coarsely check
     # that the index of interest is mentioned.
+    await a.startup()
     with record_explanations() as e:
         results = await a.search(Key("number_as_string") == "3").keys_range(0, 5)
         assert len(results) == 1
@@ -200,6 +190,7 @@ async def test_metadata_index_is_used(a):
         )
         assert len(results) == 1
         assert "top_level_metadata" in str(e)
+    await a.shutdown()
 
 
 @pytest.mark.asyncio