Improve from_dfs error messages

Co-Authored-By: Florentin Dörre <florentin.dorre@neotechnology.com>

Author: adamnsch

python-wrapper/src/neo4j_viz/pandas.py

@@ -4,6 +4,7 @@
 from typing import Optional, Union
 
 from pandas import DataFrame
+from pydantic import BaseModel, ValidationError
 
 from .node import Node
 from .relationship import Relationship
@@ -12,6 +13,17 @@
 DFS_TYPE = Union[DataFrame, Iterable[DataFrame]]
 
 
+def _parse_validation_error(e: ValidationError, entity_type: type[BaseModel]) -> None:
+    for err in e.errors():
+        loc = err["loc"][0]
+        if err["type"] == "missing":
+            raise ValueError(
+                f"Mandatory column '{loc}' is missing. Expected one of {entity_type.model_fields[loc].validation_alias.choices} to be present"  # type: ignore
+            )
+        else:
+            raise ValueError(f"Error for column '{loc}' with provided input '{err['input']}'. Reason: {err['msg']}")
+
+
 def _from_dfs(
     node_dfs: DFS_TYPE,
     rel_dfs: DFS_TYPE,
@@ -41,7 +53,10 @@ def _from_dfs(
                         key = rename_properties[key]
                     properties[key] = value
 
-            nodes.append(Node(**top_level, properties=properties))
+                try:
+                    nodes.append(Node(**top_level, properties=properties))
+                except ValidationError as e:
+                    _parse_validation_error(e, Node)
 
     if isinstance(rel_dfs, DataFrame):
         rel_dfs_iter: Iterable[DataFrame] = [rel_dfs]
@@ -61,7 +76,10 @@ def _from_dfs(
                         key = rename_properties[key]
                     properties[key] = value
 
-            relationships.append(Relationship(**top_level, properties=properties))
+            try:
+                relationships.append(Relationship(**top_level, properties=properties))
+            except ValidationError as e:
+                _parse_validation_error(e, Relationship)
 
     VG = VisualizationGraph(nodes=nodes, relationships=relationships)
 

python-wrapper/tests/test_pandas.py

@@ -1,3 +1,4 @@
+import pytest
 from pandas import DataFrame
 from pydantic_extra_types.color import Color
 
@@ -104,3 +105,62 @@ def test_from_dfs() -> None:
     assert VG.relationships[1].source == 1
     assert VG.relationships[1].target == 0
     assert VG.relationships[1].caption == "REL2"
+
+
+def test_node_errors() -> None:
+    nodes = DataFrame(
+        {"caption": ["A", "B"], "size": [1337, 42], "color": "#FF0000", "instrument": ["piano", "guitar"]}
+    )
+    with pytest.raises(
+        ValueError,
+        match=r"Mandatory column 'id' is missing. Expected one of \['id', 'ID', 'id', 'nodeid', 'NODEID', 'nodeid', 'node_id', 'NODE_ID', 'nodeId'\] to be present",
+    ):
+        from_dfs(nodes, [])
+
+    nodes = DataFrame(
+        {
+            "id": [0, 1],
+            "caption": ["A", "B"],
+            "size": ["aaa", 42],
+            "color": "#FF0000",
+            "instrument": ["piano", "guitar"],
+        }
+    )
+    with pytest.raises(
+        ValueError,
+        match=r"Error for column 'size' with provided input 'aaa'. Reason: Input should be a valid integer, unable to parse string as an integer",
+    ):
+        from_dfs(nodes, [])
+
+
+def test_rel_errors() -> None:
+    nodes = DataFrame(
+        {"id": [0, 1], "caption": ["A", "B"], "size": [1337, 42], "color": "#FF0000", "instrument": ["piano", "guitar"]}
+    )
+    relationships = DataFrame(
+        {
+            "target": [1, 0],
+            "caption": ["REL", "REL2"],
+            "weight": [1.0, 2.0],
+        }
+    )
+    with pytest.raises(
+        ValueError,
+        match=r"Mandatory column 'source' is missing. Expected one of \['source', 'SOURCE', 'source', 'sourcenodeid', 'SOURCENODEID', 'sourcenodeid', 'source_node_id', 'SOURCE_NODE_ID', 'sourceNodeId', 'from', 'FROM', 'from'\] to be present",
+    ):
+        from_dfs(nodes, relationships)
+
+    relationships = DataFrame(
+        {
+            "source": [0, 1],
+            "target": [1, 0],
+            "caption": ["REL", "REL2"],
+            "caption_size": [1.0, -300],
+            "weight": [1.0, 2.0],
+        }
+    )
+    with pytest.raises(
+        ValueError,
+        match=r"Error for column 'caption_size' with provided input '-300.0'. Reason: Input should be greater than 0",
+    ):
+        from_dfs(nodes, relationships)