Dependencies: Make MongoDB subsystem not strictly depend on Zyp

.github/workflows/tests.yml

@@ -106,7 +106,7 @@ jobs:
           pip install "setuptools>=64" --upgrade
 
           # Install package in editable mode.
-          pip install --use-pep517 --prefer-binary --editable=.[mongodb,develop,test]
+          pip install --use-pep517 --prefer-binary --editable=.[develop,test]
 
       - name: Run linters and software tests
         run: poe test -- -m 'dynamodb'
@@ -161,7 +161,7 @@ jobs:
           pip install "setuptools>=64" --upgrade
 
           # Install package in editable mode.
-          pip install --use-pep517 --prefer-binary --editable=.[mongodb,develop,test]
+          pip install --use-pep517 --prefer-binary --editable=.[mongodb,zyp,develop,test]
 
       - name: Run linters and software tests
         run: poe test -- -m 'mongodb'

CHANGES.md

@@ -7,6 +7,7 @@
 - MongoDB: Use `bson` package to parse BSON CANONICAL representation
 - MongoDB: Complete and verify BSON data type mapping end-to-end
 - MongoDB: Use improved decoding machinery also for `MongoDBCDCTranslator`
+- Dependencies: Make MongoDB subsystem not strictly depend on Zyp
 
 ## 2024/09/10 v0.0.15
 - Added Zyp Treatments, a slightly tailored transformation subsystem

src/commons_codec/transform/mongodb.py

@@ -15,16 +15,14 @@
 from sqlalchemy_cratedb.support import quote_relation_name
 
 from commons_codec.model import SQLOperation
-from zyp.model.base import DictOrList
-from zyp.model.collection import CollectionTransformation
-
-logger = logging.getLogger(__name__)
-
 
 Document = t.Mapping[str, t.Any]
 DocumentCollection = t.List[Document]
 
 
+logger = logging.getLogger(__name__)
+
+
 def date_converter(value):
     if isinstance(value, int):
         return value
@@ -45,15 +43,16 @@ class MongoDBCrateDBConverter:
     Extracted from cratedb-toolkit, earlier migr8.
     """
 
-    transformation: t.Union[CollectionTransformation, None] = None
+    transformation: t.Any = None
 
     def decode_documents(self, data: t.Iterable[Document]) -> Iterable[Document]:
         """
         Decode MongoDB Extended JSON, considering CrateDB specifics.
         """
         data = map(self.decode_value, data)
+        # TODO: This is currently untyped. Types are defined in Zyp, see `zyp.model.base`.
         if self.transformation is not None:
-            data = self.transformation.apply(t.cast(DictOrList, data))
+            data = self.transformation.apply(data)
         return data
 
     def decode_document(self, data: Document) -> Document:

tests/transform/mongodb/__init__.py

@@ -0,0 +1,3 @@
+import pytest
+
+pytest.importorskip("bson", reason="Skipping MongoDB/BSON tests because 'bson' package is not installed")

tests/transform/mongodb/test_mongodb_cdc.py

@@ -5,12 +5,9 @@
 
 import datetime
 
-from commons_codec.model import SQLOperation
-
-pytest.importorskip("pymongo")
-
 from bson import ObjectId, Timestamp
 
+from commons_codec.model import SQLOperation
 from commons_codec.transform.mongodb import MongoDBCDCTranslator
 
 MSG_OPERATION_UNKNOWN = {

tests/transform/mongodb/test_mongodb_convert.py

@@ -1,12 +1,13 @@
+# ruff: noqa: E402
 import pytest
 
+pytestmark = pytest.mark.mongodb
+
 from commons_codec.transform.mongodb import MongoDBCrateDBConverter, date_converter
 from zyp.model.bucket import BucketTransformation, ValueConverter
 from zyp.model.collection import CollectionTransformation
 from zyp.model.treatment import Treatment
 
-pytestmark = pytest.mark.mongodb
-
 
 def test_date_converter_int():
     """

tests/transform/mongodb/test_mongodb_full.py

@@ -1,5 +1,8 @@
+# ruff: noqa: E402
 import pytest
 
+pytestmark = pytest.mark.mongodb
+
 from commons_codec.model import SQLOperation
 from commons_codec.transform.mongodb import MongoDBFullLoadTranslator
 from tests.transform.mongodb.data import RECORD_IN_ALL_TYPES, RECORD_OUT_ALL_TYPES