chore: set validation and typehint

Author: ion-elgreco

python/datafusion/context.py

@@ -63,6 +63,15 @@ def __arrow_c_array__(  # noqa: D105
     ) -> tuple[object, object]: ...
 
 
+class TableProviderExportable(Protocol):
+    """Type hint for object that has __datafusion_table_provider__ PyCapsule.
+
+    https://datafusion.apache.org/python/user-guide/io/table_provider.html
+    """
+
+    def __datafusion_table_provider__(self) -> object: ...  # noqa: D105
+
+
 class SessionConfig:
     """Session configuration options."""
 
@@ -685,7 +694,9 @@ def deregister_table(self, name: str) -> None:
         """Remove a table from the session."""
         self.ctx.deregister_table(name)
 
-    def register_table_provider(self, name: str, provider: Any) -> None:
+    def register_table_provider(
+        self, name: str, provider: TableProviderExportable
+    ) -> None:
         """Register a table provider.
 
         This table provider must have a method called ``__datafusion_table_provider__``

src/context.rs

@@ -43,7 +43,7 @@ use crate::store::StorageContexts;
 use crate::udaf::PyAggregateUDF;
 use crate::udf::PyScalarUDF;
 use crate::udwf::PyWindowUDF;
-use crate::utils::{get_tokio_runtime, wait_for_future};
+use crate::utils::{get_tokio_runtime, validate_pycapsule, wait_for_future};
 use datafusion::arrow::datatypes::{DataType, Schema, SchemaRef};
 use datafusion::arrow::pyarrow::PyArrowType;
 use datafusion::arrow::record_batch::RecordBatch;
@@ -576,7 +576,7 @@ impl PySessionContext {
         if provider.hasattr("__datafusion_table_provider__")? {
             let capsule = provider.getattr("__datafusion_table_provider__")?.call0()?;
             let capsule = capsule.downcast::<PyCapsule>()?;
-            // validate_pycapsule(capsule, "arrow_array_stream")?;
+            validate_pycapsule(capsule, "datafusion_table_provider")?;
 
             let provider = unsafe { capsule.reference::<FFI_TableProvider>() };
             let provider: ForeignTableProvider = provider.into();

src/dataframe.rs

@@ -44,7 +44,7 @@ use crate::expr::sort_expr::to_sort_expressions;
 use crate::physical_plan::PyExecutionPlan;
 use crate::record_batch::PyRecordBatchStream;
 use crate::sql::logical::PyLogicalPlan;
-use crate::utils::{get_tokio_runtime, wait_for_future};
+use crate::utils::{get_tokio_runtime, validate_pycapsule, wait_for_future};
 use crate::{
     errors::DataFusionError,
     expr::{sort_expr::PySortExpr, PyExpr},
@@ -724,22 +724,3 @@ fn record_batch_into_schema(
 
     RecordBatch::try_new(schema, data_arrays)
 }
-
-fn validate_pycapsule(capsule: &Bound<PyCapsule>, name: &str) -> PyResult<()> {
-    let capsule_name = capsule.name()?;
-    if capsule_name.is_none() {
-        return Err(PyValueError::new_err(
-            "Expected schema PyCapsule to have name set.",
-        ));
-    }
-
-    let capsule_name = capsule_name.unwrap().to_str()?;
-    if capsule_name != name {
-        return Err(PyValueError::new_err(format!(
-            "Expected name '{}' in PyCapsule, instead got '{}'",
-            name, capsule_name
-        )));
-    }
-
-    Ok(())
-}

src/utils.rs

@@ -18,7 +18,9 @@
 use crate::errors::DataFusionError;
 use crate::TokioRuntime;
 use datafusion::logical_expr::Volatility;
+use pyo3::exceptions::PyValueError;
 use pyo3::prelude::*;
+use pyo3::types::PyCapsule;
 use std::future::Future;
 use std::sync::OnceLock;
 use tokio::runtime::Runtime;
@@ -58,3 +60,22 @@ pub(crate) fn parse_volatility(value: &str) -> Result<Volatility, DataFusionErro
         }
     })
 }
+
+pub(crate) fn validate_pycapsule(capsule: &Bound<PyCapsule>, name: &str) -> PyResult<()> {
+    let capsule_name = capsule.name()?;
+    if capsule_name.is_none() {
+        return Err(PyValueError::new_err(
+            "Expected schema PyCapsule to have name set.",
+        ));
+    }
+
+    let capsule_name = capsule_name.unwrap().to_str()?;
+    if capsule_name != name {
+        return Err(PyValueError::new_err(format!(
+            "Expected name '{}' in PyCapsule, instead got '{}'",
+            name, capsule_name
+        )));
+    }
+
+    Ok(())
+}