fix: (Temporarily) Re-add suport for pre-2.6.0 YAMLs with PyPDFConverter (#8443)

Author: shadeMe

haystack/components/converters/pypdf.py

@@ -12,6 +12,7 @@
 from haystack.dataclasses import ByteStream
 from haystack.lazy_imports import LazyImport
 from haystack.utils.base_serialization import deserialize_class_instance, serialize_class_instance
+from haystack.utils.type_serialization import deserialize_type
 
 with LazyImport("Run 'pip install pypdf'") as pypdf_import:
     from pypdf import PdfReader
@@ -118,7 +119,12 @@ def from_dict(cls, data):
         """
         custom_converter_data = data["init_parameters"]["converter"]
         if custom_converter_data is not None:
-            data["init_parameters"]["converter"] = deserialize_class_instance(custom_converter_data)
+            if "data" in custom_converter_data:
+                data["init_parameters"]["converter"] = deserialize_class_instance(custom_converter_data)
+            else:
+                # TODO: Remove in 2.7.0
+                converter_class = deserialize_type(custom_converter_data["type"])
+                data["init_parameters"]["converter"] = converter_class.from_dict(custom_converter_data)
         return default_from_dict(cls, data)
 
     def _default_convert(self, reader: "PdfReader") -> Document:

releasenotes/notes/pypdf-serde-fixes-revert-22fcdad91951d3c2.yaml

@@ -0,0 +1,4 @@
+---
+fixes:
+  - |
+    Revert change to PyPDFConverter that broke the deserialization of pre 2.6.0 YAMLs.

test/components/converters/test_pypdf_to_document.py

@@ -7,7 +7,7 @@
 import pytest
 
 from haystack import Document, default_from_dict, default_to_dict
-from haystack.components.converters.pypdf import PyPDFToDocument
+from haystack.components.converters.pypdf import PyPDFToDocument, DefaultConverter
 from haystack.dataclasses import ByteStream
 
 
@@ -79,6 +79,15 @@ def test_from_dict_custom_converter(self):
         assert isinstance(instance, PyPDFToDocument)
         assert isinstance(instance.converter, CustomConverter)
 
+    def test_from_dict_pre_2_6_0(self):
+        data = {
+            "type": "haystack.components.converters.pypdf.PyPDFToDocument",
+            "init_parameters": {"converter": {"type": "haystack.components.converters.pypdf.DefaultConverter"}},
+        }
+        instance = PyPDFToDocument.from_dict(data)
+        assert isinstance(instance, PyPDFToDocument)
+        assert isinstance(instance.converter, DefaultConverter)
+
     @pytest.mark.integration
     def test_run(self, test_files_path, pypdf_converter):
         """