feat: Build for Python 3.14 (#383)

Author: mattkjames7

.github/workflows/build-and-test.yml

@@ -43,7 +43,7 @@ jobs:
           python-version: "3.12"
 
       - name: Install uv
-        uses: astral-sh/setup-uv@38f3f104447c67c051c4a08e39b64a148898af3a # v4
+        uses: astral-sh/setup-uv@cec208311dfd045dd5311c1add060b2062131d57 # v8.0.0
         with:
           version: "0.11.3"
 
@@ -83,7 +83,7 @@ jobs:
           python-version: ${{ matrix.python-version }}
 
       - name: Install uv
-        uses: astral-sh/setup-uv@38f3f104447c67c051c4a08e39b64a148898af3a # v4
+        uses: astral-sh/setup-uv@cec208311dfd045dd5311c1add060b2062131d57 # v8.0.0
         with:
           version: "0.11.3"
 

.github/workflows/build-packages.yml

@@ -126,7 +126,7 @@ jobs:
           python-version: 3.12
 
       - name: Install uv
-        uses: astral-sh/setup-uv@38f3f104447c67c051c4a08e39b64a148898af3a # v4
+        uses: astral-sh/setup-uv@cec208311dfd045dd5311c1add060b2062131d57 # v8.0.0
         with:
           version: "0.11.3"
 

.github/workflows/release.yml

@@ -17,7 +17,7 @@ An Object Graph Mapper or OGM provides a developer-friendly workflow that allows
 
 ### Prerequisites
 
-- **Python 3.10 - 3.12**
+- **Python 3.10+**
 - [`pymgclient`](https://github.com/memgraph/pymgclient):
 
   - Install `pymgclient` [build prerequisites](https://memgraph.github.io/pymgclient/introduction.html#build-prerequisites)
@@ -27,8 +27,10 @@ An Object Graph Mapper or OGM provides a developer-friendly workflow that allows
   pip install --user pymgclient
   ```
 
-> [!WARNING]  
-> Python 3.11 users: On Windows, GQLAlchemy is not yet compatible with this Python version. Linux users can install GQLAlchemy **without** the DGL extra (due to its dependencies not supporting Python 3.11 yet). If this is currently a blocker for you, please let us know by [opening an issue](https://github.com/memgraph/gqlalchemy/issues).
+> [!NOTE]
+> GQLAlchemy is tested on Python `3.10` through `3.14` in CI. Some optional extras
+> (for example TensorFlow/TF-GNN stacks) are available only for a subset of Python
+> versions due to upstream wheel availability.
 
 ### Install GQLAlchemy
 

README.md

@@ -231,6 +231,12 @@ class Streamer(User):
 
 The above class definition is not ideal, since it is not enforcing schema as before. To do that, [add constraints](#create-constraints).
 
+!!! note
+    In the current Pydantic-compatible OGM behavior, omitted nullable fields
+    (`Optional[...]`) can be treated as `None` during model initialization.
+    This preserves historical GQLAlchemy usage patterns, but differs from strict
+    "required but nullable" validation semantics some users may expect.
+
 If you try loading the node again, the following code:
 
 ```python

docs/how-to-guides/ogm.md

@@ -13,8 +13,10 @@ An Object Graph Mapper or OGM provides a developer-friendly workflow for writing
 Either install GQLAlchemy through [pip](installation.md#pip) or [build it from
 source](installation.md#source). If you are using [Conda](https://docs.conda.io/en/latest/) for Python environment management, you can install GQLAlchemy through [pip](installation.md#pip).
 
-!!! danger
-    Python 3.11 users: On Windows, GQLAlchemy is not yet compatible with this Python version. Linux users can install GQLAlchemy **without** the DGL extra (due to its dependencies not supporting Python 3.11 yet). If this is currently a blocker for you, please let us know by [opening an issue](https://github.com/memgraph/gqlalchemy/issues).
+!!! note
+    GQLAlchemy supports modern Python versions (including Python `3.14`), with CI
+    coverage for `3.10` through `3.14`. Some optional extras may have narrower
+    version support based on upstream dependencies.
 
 ### 2. Connect to Memgraph
 

docs/index.md

@@ -7,7 +7,7 @@ as pip and uv, and by building it from source.
 
 To install GQLAlchemy, you will need the following:
 
-- **Python 3.10 - 3.12**
+- **Python 3.10+**
 - [`pymgclient`](https://github.com/memgraph/pymgclient):
 
     * Install `pymgclient` [build prerequisites](https://memgraph.github.io/pymgclient/introduction.html#build-prerequisites)
@@ -17,8 +17,10 @@ To install GQLAlchemy, you will need the following:
     pip install --user pymgclient
     ```
 
-!!! danger
-    Python 3.11 users: On Windows, GQLAlchemy is not yet compatible with this Python version. If this is currently a blocker for you, please let us know by [opening an issue](https://github.com/memgraph/gqlalchemy/issues).
+!!! note
+    GQLAlchemy is tested on Python `3.10` through `3.14` in CI. Some optional extras
+    (for example TensorFlow/TF-GNN stacks) are available only for a subset of Python
+    versions due to upstream wheel availability.
 
 ## Install with pip
 

docs/installation.md

@@ -156,6 +156,9 @@ def parse_obj(cls, obj)
 Used to convert a dictionary object into the appropriate
 GraphObject.
 
+Compatibility note: `parse_obj` is retained for backward compatibility.
+For Pydantic v2-style usage, prefer `model_validate`.
+
 ## NodeMetaclass Objects
 
 ```python

docs/reference/gqlalchemy/models.md

@@ -13,8 +13,7 @@
 # limitations under the License.
 
 import warnings
-
-from pydantic.v1 import validator  # noqa F401
+from pydantic import validator as _validator  # noqa F401
 
 from gqlalchemy.models import (  # noqa F401
     MemgraphConstraintExists,
@@ -64,6 +63,7 @@
 
 warnings.filterwarnings("once", category=GQLAlchemyWarning)
 __all__ = ["Memgraph"]
+_validator_deprecation_warning_shown = False
 
 call = Call
 create = Create
@@ -75,3 +75,18 @@
 return_ = Return
 load_csv = LoadCsv
 MemgraphQueryBuilder = QueryBuilder
+
+
+def validator(*args, **kwargs):
+    global _validator_deprecation_warning_shown
+    if not _validator_deprecation_warning_shown:
+        warnings.warn(
+            "Importing 'validator' from this module is deprecated and will be "
+            "removed in a future release. For Pydantic v2, migrate to "
+            "'field_validator' where applicable, or keep importing "
+            "'validator' from 'pydantic' for v1 compatibility.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+        _validator_deprecation_warning_shown = True
+    return _validator(*args, **kwargs)

gqlalchemy/__init__.py

@@ -119,7 +119,7 @@ def _create_connection(self) -> Connection:
 def _convert_memgraph_value(value: Any) -> Any:
     """Converts Memgraph objects to custom Node/Relationship objects."""
     if isinstance(value, mgclient.Relationship):
-        return Relationship.parse_obj(
+        return Relationship.model_validate(
             {
                 "_type": value.type,
                 "_id": value.id,
@@ -130,7 +130,7 @@ def _convert_memgraph_value(value: Any) -> Any:
         )
 
     if isinstance(value, mgclient.Node):
-        return Node.parse_obj(
+        return Node.model_validate(
             {
                 "_id": value.id,
                 "_labels": set(value.labels),
@@ -139,7 +139,7 @@ def _convert_memgraph_value(value: Any) -> Any:
         )
 
     if isinstance(value, mgclient.Path):
-        return Path.parse_obj(
+        return Path.model_validate(
             {
                 "_nodes": list([_convert_memgraph_value(node) for node in value.nodes]),
                 "_relationships": list([_convert_memgraph_value(rel) for rel in value.relationships]),
@@ -192,7 +192,7 @@ def _create_connection(self):
 def _convert_neo4j_value(value: Any) -> Any:
     """Converts Neo4j objects to custom Node/Relationship objects."""
     if isinstance(value, Neo4jRelationship):
-        return Relationship.parse_obj(
+        return Relationship.model_validate(
             {
                 "_type": value.type,
                 "_id": value.id,
@@ -203,7 +203,7 @@ def _convert_neo4j_value(value: Any) -> Any:
         )
 
     if isinstance(value, Neo4jNode):
-        return Node.parse_obj(
+        return Node.model_validate(
             {
                 "_id": value.id,
                 "_labels": set(value.labels),
@@ -212,7 +212,7 @@ def _convert_neo4j_value(value: Any) -> Any:
         )
 
     if isinstance(value, Neo4jPath):
-        return Path.parse_obj(
+        return Path.model_validate(
             {
                 "_nodes": list([_convert_neo4j_value(node) for node in value.nodes]),
                 "_relationships": list([_convert_neo4j_value(rel) for rel in value.relationships]),