Implement web-poet serializers

scrapy_zyte_api/__init__.py

@@ -5,6 +5,9 @@
 
     install_reactor("twisted.internet.asyncioreactor.AsyncioSelectorReactor")
 
+# Register web-poet serializers
+from . import _serialization  # noqa: F401
+
 from ._annotations import ExtractFrom, actions, custom_attrs
 from ._middlewares import (
     ScrapyZyteAPIDownloaderMiddleware,

scrapy_zyte_api/_page_inputs.py

@@ -18,6 +18,7 @@ class Actions:
     results: Optional[List[_ActionResult]]
 
 
+@attrs.define
 class Geolocation:
     """A page input that forces a given geolocation for all other page inputs.
 

scrapy_zyte_api/_serialization.py

@@ -0,0 +1,37 @@
+import json
+
+from ._page_inputs import Actions, Geolocation, Screenshot
+
+try:
+    from web_poet.serialization import SerializedLeafData, register_serialization
+except ImportError:
+    pass
+else:
+
+    def _serialize_Actions(o: Actions) -> SerializedLeafData:
+        return {"results.json": json.dumps(o.results).encode()}
+
+    def _deserialize_Actions(cls: type[Actions], data: SerializedLeafData) -> Actions:
+        return cls(results=json.loads(data["results.json"]))
+
+    register_serialization(_serialize_Actions, _deserialize_Actions)
+
+    def _serialize_Geolocation(o: Geolocation) -> SerializedLeafData:
+        return {}
+
+    def _deserialize_Geolocation(
+        cls: type[Geolocation], data: SerializedLeafData
+    ) -> Geolocation:
+        return cls()
+
+    register_serialization(_serialize_Geolocation, _deserialize_Geolocation)
+
+    def _serialize_Screenshot(o: Screenshot) -> SerializedLeafData:
+        return {"body": o.body}
+
+    def _deserialize_Screenshot(
+        cls: type[Screenshot], data: SerializedLeafData
+    ) -> Screenshot:
+        return cls(body=data["body"])
+
+    register_serialization(_serialize_Screenshot, _deserialize_Screenshot)

tests/test_middlewares.py

@@ -3,10 +3,10 @@
 
 import pytest
 from packaging.version import Version
-from scrapy.utils.defer import deferred_f_from_coro_f
 from scrapy import Request, Spider
 from scrapy.http.response import Response
 from scrapy.item import Item
+from scrapy.utils.defer import deferred_f_from_coro_f
 from scrapy.utils.test import get_crawler
 
 from scrapy_zyte_api import (

tests/test_serialization.py

@@ -0,0 +1,24 @@
+import pytest
+
+from scrapy_zyte_api._page_inputs import Actions, Geolocation, Screenshot
+
+
+@pytest.mark.parametrize(
+    "case",
+    [
+        {"cls": Actions, "kwargs": {"results": [{"action": "click", "id": "x"}]}},
+        {"cls": Geolocation, "kwargs": {}},
+        {"cls": Screenshot, "kwargs": {"body": b"PNGDATA"}},
+    ],
+)
+def test(case):
+    wp = pytest.importorskip("web_poet.serialization")
+    cls = case["cls"]
+    kwargs = case["kwargs"]
+
+    obj = cls(**kwargs)
+    data = wp.serialize_leaf(obj)
+    reconstructed = wp.deserialize_leaf(cls, data)
+
+    assert reconstructed == obj
+    assert id(reconstructed) != id(obj)