Tech Debt: Upgrading boto to boto3 for api & ingester

.github/workflows/actions.yml

@@ -3,14 +3,15 @@ on: [push, pull_request]
 
 jobs:
   test-client:
-    runs-on: ubuntu-20.04
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
     strategy:
       matrix:
         python: [3.8, 3.9, "3.10", 3.12]
         extras: ["test", "test,queuable,sentry"]
     steps:
       - name: Setup Python
-        uses: actions/setup-python@v2.2.2
+        uses: actions/setup-python@v5
         with:
           python-version: ${{ matrix.python }}
       - name: Check out repository code
@@ -20,13 +21,14 @@ jobs:
       - name: Test
         working-directory: ./client
         run: |
-          pip install -e .[${{ matrix.extras }}]
-          py.test
+          pip --version
+          pip install --verbose .[${{ matrix.extras }}]
+          python -m pytest --import-mode=importlib ./datalake/tests/
   test-docker:
     runs-on: ubuntu-latest
     steps:
       - name: Check out repository code
-        uses: actions/checkout@v2
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
       - name: Test

.gitignore

@@ -58,3 +58,6 @@ target/
 
 # build artifacts
 version.txt
+
+# Claude Code
+.claude/

Dockerfile

@@ -7,29 +7,33 @@ ENV	LC_ALL C.UTF-8
 
 # TODO: keep requirements in one place
 RUN pip install \
-    blinker>=1.4 \
-    boto3>=1.1.3 \
-    click>=5.1 \
-    Flask>=0.10.1 \
-    flask-swagger>=0.2.14 \
-    memoized_property>=1.0.1 \
-    python-dateutil>=2.4.2 \
-    python-dotenv>=0.1.3 \
-    pytz>=2015.4 \
-    sentry-sdk[flask]>=0.19.5 \
-    requests>=2.5 \
-    simplejson>=3.3.1 \
-    six>=1.10.0 \
+    'blinker>=1.4' \
+    'boto3==1.35.41' \
+    'botocore==1.35.64' \
+    'click>=5.1' \
+    'datalake<2' \
+    'Flask>=0.10.1' \
+    'flask-swagger>=0.2.14' \
+    'memoized_property>=1.0.1' \
+    'pyinotify>=0.9.4' \
+    'python-dateutil>=2.4.2' \
+    'python-dotenv>=0.1.3' \
+    'pytz>=2015.4' \
+    'raven>=5.0.0' \
+    'requests>=2.5' \
+    'sentry-sdk[flask]>=0.19.5' \
+    'simplejson>=3.3.1' \
+    'six>=1.10.0' \
     # test requirements
     'flake8>=2.5.0,<4.1' \
     'freezegun<1' \
-    'moto<3' \
+    'moto>=5,<6' \
     'pytest<8' \
+    'pytest-cov>=2.5.1,<4' \
+    'pytest-xdist' \
     'responses<0.22.0' \
-    pyinotify>=0.9.4, \
-    raven>=5.0.0 \
-    'tox>4,<5' \
-    'datalake<2'
+    'tox>4,<5'
+
 
 RUN mkdir -p /opt/
 COPY . /opt/

Makefile

@@ -7,18 +7,27 @@ IMAGE="$(REPO)/$(REPO_PATH):$(VERSION)"
 docker: version
 	docker build --build-arg VERSION=$(VERSION) -t $(IMAGE) .
 
-.PHONY: devshell  # Open a developer shell in the docker env
-devshell: docker
+.PHONY: dev  # Open a developer shell in the docker env
+dev: docker
 	docker run --rm -it -v $$PWD:/opt --entrypoint /bin/bash $(IMAGE)
 
 test-client: docker
-	docker run --rm --entrypoint tox $(IMAGE) -c /opt/client/tox.ini
+	docker run --rm -t --entrypoint tox $(IMAGE) -c /opt/client/tox.ini
 
 test-ingester: docker
-	docker run --rm --entrypoint py.test $(IMAGE) ingester
+	docker run --rm -t --entrypoint pytest $(IMAGE) /opt/ingester -svvx
 
 test-api: docker
-	docker run --rm --entrypoint py.test $(IMAGE) api
+	docker run --rm -t --entrypoint pytest $(IMAGE) /opt/api -svvx
+
+testp-client: docker
+	docker run --rm -t --entrypoint /bin/bash $(IMAGE) -c "cd /opt/client && pytest -svvx -n auto"
+
+testp-ingester: docker
+	docker run --rm -t --entrypoint pytest $(IMAGE) /opt/ingester -svvx -n auto
+
+testp-api: docker
+	docker run --rm -t --entrypoint pytest $(IMAGE) /opt/api -svvx -n auto
 
 .PHONY: test  # Run the tests
 test:
@@ -27,6 +36,18 @@ test:
 	$(MAKE) test-ingester
 	$(MAKE) test-api
 
+.PHONY: testp  # Run all tests in parallel with pytest-xdist
+testp: docker
+	docker run --rm -t --entrypoint /bin/bash $(IMAGE) -c "\
+		set -e && \
+		echo '==> Running client tests...' && \
+		cd /opt/client && pytest -svvx -n auto && \
+		echo '==> Running ingester tests...' && \
+		cd /opt/ingester && pytest -svvx -n auto && \
+		echo '==> Running API tests...' && \
+		cd /opt/api && pytest -svvx -n auto && \
+		echo '==> All tests passed!'"
+
 .PHONY: push
 push:
 	docker push $(IMAGE)

api/datalake_api/v0.py

@@ -18,6 +18,8 @@
 from flask import current_app as app
 import os
 import simplejson as json
+from datetime import datetime, timezone
+import decimal
 from .querier import ArchiveQuerier, Cursor, InvalidCursor, \
     DEFAULT_LOOKBACK_DAYS
 from .fetcher import ArchiveFileFetcher
@@ -29,6 +31,38 @@
 
 _archive_querier = None
 
+
+def unix_ms_to_utc_iso(unix_ms):
+    if unix_ms is None:
+        return unix_ms
+    unix_ms_to_iso = unix_ms
+    if isinstance(unix_ms_to_iso, decimal.Decimal):
+        unix_ms_to_iso = float(unix_ms_to_iso)
+    iso = datetime.fromtimestamp(
+        unix_ms_to_iso / 1000.0, tz=timezone.utc
+    ).isoformat(timespec='milliseconds').replace('+00:00', 'Z')
+    return iso
+
+
+def add_utc_metadata(metadata):
+    """Add ISO-8601 UTC timestamp fields to metadata dict
+
+    This function takes a metadata dict and adds start_iso and end_iso fields
+    based on existing start and end epoch timestamps
+    iso precision is set to milliseconds
+    Can be expanded to add any api-level metadata
+    """
+    if not metadata:
+        return metadata
+
+    start_iso = unix_ms_to_utc_iso(metadata['start'])
+    end_iso = unix_ms_to_utc_iso(metadata['end'])
+
+    metadata['start_iso'] = start_iso
+    metadata['end_iso'] = end_iso
+    return metadata
+
+
 def _get_aws_kwargs():
     kwargs = dict(
         region_name=app.config.get('AWS_REGION'),
@@ -305,6 +339,14 @@ def files_get():
                                   type: string
                                   description: 16-byte blake2 hash of the file
                                                content
+                                start_iso:
+                                  type: string
+                                  description: the start time of the file in ISO
+                                               format UTC iso timezone
+                                end_iso:
+                                  type: string
+                                  description: the end time of the file in ISO
+                                               format UTC iso timezone
 
               next:
                   type: string
@@ -349,7 +391,10 @@ def files_get():
                                    where=params.get('where'),
                                    cursor=params.get('cursor'))
 
-    [r.update(http_url=_get_canonical_http_url(r)) for r in results]
+    for r in results:
+        r.update(http_url=_get_canonical_http_url(r))
+        r['metadata'] = add_utc_metadata(r['metadata'])
+
     response = {
         'records': results,
         'next': _get_next_url(flask.request, results),
@@ -476,6 +521,7 @@ def file_get_metadata(file_id):
           id: DatalakeAPIError
     '''
     f = _get_file(file_id)
+    f.metadata = add_utc_metadata(f.metadata)
     return Response(json.dumps(f.metadata), content_type='application/json')
 
 
@@ -542,6 +588,7 @@ def latest_get(what, where):
     params = _validate_latest_params(params)
     f = _get_latest(what, where, params.get('lookback', DEFAULT_LOOKBACK_DAYS))
     f.update(http_url=_get_canonical_http_url(f))
+    f['metadata'] = add_utc_metadata(f['metadata'])
     return Response(json.dumps(f), content_type='application/json')
 
 

api/setup.py

@@ -30,7 +30,7 @@ def get_version_from_pyver():
         if 'sdist' in sys.argv or 'bdist_wheel' in sys.argv:
             raise ImportError('You must install pyver to create a package')
         else:
-            return 'noversion'
+            return '0.0.0'
     version, version_info = pyver.get_version(pkg="datalake_api",
                                               public=True)
     return version

api/tests/conftest.py

@@ -18,7 +18,7 @@
     ClientError as BotoClientError,
     NoCredentialsError
 )
-from moto import mock_dynamodb2
+from moto import mock_aws
 
 from datalake_api import app as datalake_api
 from datalake.tests import *  # noqa
@@ -33,30 +33,27 @@
 # This will cause moto to fail
 # But more critically, may impact production systems
 # So we test for real credentials and fail hard if they exist
-sts = boto3.client('sts')
-try:
-    sts.get_caller_identity()
-    pytest.exit("Real AWS credentials detected, aborting", 3)
-except NoCredentialsError:
-    pass  # no credentials are good
+# Session fixture runs in pytest setup rather than at import time
+@pytest.fixture(scope='session', autouse=True)
+def verify_no_aws_credentials():
+    sts = boto3.client('sts')
+    try:
+        sts.get_caller_identity()
+        pytest.exit("Real AWS credentials detected, aborting", 3)
+    except NoCredentialsError:
+        pass
 
 
 def get_client():
     from datalake_api import settings
     datalake_api.app.config.from_object(settings)
 
     datalake_api.app.config['TESTING'] = True
+    datalake_api.app.config['AWS_REGION'] = 'us-east-1'
     datalake_api.app.config['AWS_ACCESS_KEY_ID'] = 'abc'
     datalake_api.app.config['AWS_SECRET_ACCESS_KEY'] = '123'
 
-    # TODO: Sigh. The api caches the archive_fetcher and s3_bucket, which is
-    # the right thing. However, because moto<3 still uses httpretty, and
-    # because httpretty wreaks havoc on the python socket code, these cached
-    # parts end up in a bad state after their first use. The right thing to do
-    # here is to upgrade moto. But for that we will also have to move
-    # everything from boto to boto3. This is a near-term goal. But first lets
-    # get everything off of python2.
-    for a in ('archive_fetcher', 's3_bucket'):
+    for a in ('archive_fetcher', 's3_bucket', 'dynamodb'):
         try:
             delattr(datalake_api.app, a)
         except AttributeError:
@@ -71,15 +68,15 @@ def client():
 
 @pytest.fixture
 def dynamodb(request):
-    mock = mock_dynamodb2()
+    mock = mock_aws()
     mock.start()
 
     def tear_down():
         mock.stop()
     request.addfinalizer(tear_down)
 
     return boto3.resource('dynamodb',
-                          region_name='us-west-2',
+                          region_name='us-east-1',
                           aws_secret_access_key='123',
                           aws_access_key_id='abc')
 

api/tests/test_metadata.py

@@ -11,6 +11,8 @@
 # WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
 # License for the specific language governing permissions and limitations under
 # the License.
+from datetime import datetime, timezone
+from decimal import Decimal
 import pytest
 import simplejson as json
 
@@ -32,7 +34,20 @@ def test_get_metadata(metadata_getter, s3_file_maker, random_metadata):
     res = metadata_getter('12345')
     assert res.status_code == 200
     assert res.content_type == 'application/json'
-    assert json.loads(res.data) == random_metadata
+    res_data = json.loads(res.data)
+    for k, v in res_data.items():
+        if k == 'start_iso' or k == 'end_iso':
+            k_epoch = k.replace('_iso','')
+            v_epoch = res_data[k_epoch]
+            if v is None:
+                assert v == v_epoch
+
+            expected_v_iso = datetime.fromtimestamp(
+                v_epoch / 1000.0, tz=timezone.utc
+            ).isoformat(timespec='milliseconds').replace('+00:00', 'Z')
+            assert v == expected_v_iso
+        else:
+            assert v == random_metadata[k]
 
 
 def test_no_such_metadata(s3_bucket_maker, metadata_getter):

client/datalake/tests/conftest.py

@@ -20,7 +20,7 @@
 
 
 try:
-    from moto import mock_s3
+    from moto import mock_aws
     import boto3
     from six.moves.urllib.parse import urlparse
     import json
@@ -125,24 +125,8 @@ def get_tmpfile(content):
 
 
 @pytest.fixture
-def aws_connector(request):
-
-    def create_connection(mocker, connector):
-        mock = mocker()
-        mock.start()
-
-        def tear_down():
-            mock.stop()
-        request.addfinalizer(tear_down)
-
-        return connector()
-
-    return create_connection
-
-
-@pytest.fixture
-def s3_connection(aws_connector):
-    with mock_s3():
+def s3_connection():
+    with mock_aws():
         yield boto3.resource('s3')