Add source root - cherry pick apache#1252

Author: ntjohnson1

src/catalog.rs

@@ -36,19 +36,19 @@ use std::any::Any;
 use std::collections::HashSet;
 use std::sync::Arc;
 
-#[pyclass(name = "RawCatalog", module = "datafusion.catalog", subclass)]
+#[pyclass(frozen, name = "RawCatalog", module = "datafusion.catalog", subclass)]
 #[derive(Clone)]
 pub struct PyCatalog {
     pub catalog: Arc<dyn CatalogProvider>,
 }
 
-#[pyclass(name = "RawSchema", module = "datafusion.catalog", subclass)]
+#[pyclass(frozen, name = "RawSchema", module = "datafusion.catalog", subclass)]
 #[derive(Clone)]
 pub struct PySchema {
     pub schema: Arc<dyn SchemaProvider>,
 }
 
-#[pyclass(name = "RawTable", module = "datafusion.catalog", subclass)]
+#[pyclass(frozen, name = "RawTable", module = "datafusion.catalog", subclass)]
 #[derive(Clone)]
 pub struct PyTable {
     pub table: Arc<dyn TableProvider>,

src/config.rs

@@ -1,4 +1,7 @@
-// Licensed to the Apache Software Foundation (ASF) under one
+// Licensed to the Apache use parking_lot::RwLock;
+
+#[pyclass(name = "Config", module = "datafusion", subclass, frozen)]
+#[derive(Clone)]Foundation (ASF) under one
 // or more contributor license agreements.  See the NOTICE file
 // distributed with this work for additional information
 // regarding copyright ownership.  The ASF licenses this file

src/context.rs

@@ -77,7 +77,7 @@ use pyo3::IntoPyObjectExt;
 use tokio::task::JoinHandle;
 
 /// Configuration options for a SessionContext
-#[pyclass(name = "SessionConfig", module = "datafusion", subclass)]
+#[pyclass(frozen, name = "SessionConfig", module = "datafusion", subclass)]
 #[derive(Clone, Default)]
 pub struct PySessionConfig {
     pub config: SessionConfig,
@@ -170,7 +170,7 @@ impl PySessionConfig {
 }
 
 /// Runtime options for a SessionContext
-#[pyclass(name = "RuntimeEnvBuilder", module = "datafusion", subclass)]
+#[pyclass(frozen, name = "RuntimeEnvBuilder", module = "datafusion", subclass)]
 #[derive(Clone)]
 pub struct PyRuntimeEnvBuilder {
     pub builder: RuntimeEnvBuilder,
@@ -257,7 +257,7 @@ impl PyRuntimeEnvBuilder {
 }
 
 /// `PySQLOptions` allows you to specify options to the sql execution.
-#[pyclass(name = "SQLOptions", module = "datafusion", subclass)]
+#[pyclass(frozen, name = "SQLOptions", module = "datafusion", subclass)]
 #[derive(Clone)]
 pub struct PySQLOptions {
     pub options: SQLOptions,

src/dataframe.rs

@@ -68,7 +68,7 @@ type SharedCachedBatches = Arc<Mutex<CachedBatches>>;
 // https://github.com/apache/datafusion-python/pull/1016#discussion_r1983239116
 // - we have not decided on the table_provider approach yet
 // this is an interim implementation
-#[pyclass(name = "TableProvider", module = "datafusion")]
+#[pyclass(frozen, name = "TableProvider", module = "datafusion")]
 pub struct PyTableProvider {
     provider: Arc<dyn TableProvider + Send>,
 }
@@ -195,7 +195,7 @@ fn build_formatter_config_from_python(formatter: &Bound<'_, PyAny>) -> PyResult<
 }
 
 /// Python mapping of `ParquetOptions` (includes just the writer-related options).
-#[pyclass(name = "ParquetWriterOptions", module = "datafusion", subclass)]
+#[pyclass(frozen, name = "ParquetWriterOptions", module = "datafusion", subclass)]
 #[derive(Clone, Default)]
 pub struct PyParquetWriterOptions {
     options: ParquetOptions,
@@ -256,7 +256,7 @@ impl PyParquetWriterOptions {
 }
 
 /// Python mapping of `ParquetColumnOptions`.
-#[pyclass(name = "ParquetColumnOptions", module = "datafusion", subclass)]
+#[pyclass(frozen, name = "ParquetColumnOptions", module = "datafusion", subclass)]
 #[derive(Clone, Default)]
 pub struct PyParquetColumnOptions {
     options: ParquetColumnOptions,

src/expr.rs

@@ -115,7 +115,7 @@ pub mod window;
 use sort_expr::{to_sort_expressions, PySortExpr};
 
 /// A PyExpr that can be used on a DataFrame
-#[pyclass(name = "RawExpr", module = "datafusion.expr", subclass)]
+#[pyclass(frozen, name = "RawExpr", module = "datafusion.expr", subclass)]
 #[derive(Debug, Clone)]
 pub struct PyExpr {
     pub expr: Expr,
@@ -637,7 +637,7 @@ impl PyExpr {
     }
 }
 
-#[pyclass(name = "ExprFuncBuilder", module = "datafusion.expr", subclass)]
+#[pyclass(frozen, name = "ExprFuncBuilder", module = "datafusion.expr", subclass)]
 #[derive(Debug, Clone)]
 pub struct PyExprFuncBuilder {
     pub builder: ExprFuncBuilder,

src/physical_plan.rs

@@ -24,7 +24,7 @@ use pyo3::{exceptions::PyRuntimeError, prelude::*, types::PyBytes};
 
 use crate::{context::PySessionContext, errors::PyDataFusionResult};
 
-#[pyclass(name = "ExecutionPlan", module = "datafusion", subclass)]
+#[pyclass(frozen, name = "ExecutionPlan", module = "datafusion", subclass)]
 #[derive(Debug, Clone)]
 pub struct PyExecutionPlan {
     pub plan: Arc<dyn ExecutionPlan>,

src/store.rs

@@ -36,7 +36,12 @@ pub enum StorageContexts {
     HTTP(PyHttpContext),
 }
 
-#[pyclass(name = "LocalFileSystem", module = "datafusion.store", subclass)]
+#[pyclass(
+    frozen,
+    name = "LocalFileSystem",
+    module = "datafusion.store",
+    subclass
+)]
 #[derive(Debug, Clone)]
 pub struct PyLocalFileSystemContext {
     pub inner: Arc<LocalFileSystem>,
@@ -62,7 +67,7 @@ impl PyLocalFileSystemContext {
     }
 }
 
-#[pyclass(name = "MicrosoftAzure", module = "datafusion.store", subclass)]
+#[pyclass(frozen, name = "MicrosoftAzure", module = "datafusion.store", subclass)]
 #[derive(Debug, Clone)]
 pub struct PyMicrosoftAzureContext {
     pub inner: Arc<MicrosoftAzure>,
@@ -134,7 +139,7 @@ impl PyMicrosoftAzureContext {
     }
 }
 
-#[pyclass(name = "GoogleCloud", module = "datafusion.store", subclass)]
+#[pyclass(frozen, name = "GoogleCloud", module = "datafusion.store", subclass)]
 #[derive(Debug, Clone)]
 pub struct PyGoogleCloudContext {
     pub inner: Arc<GoogleCloudStorage>,
@@ -164,7 +169,7 @@ impl PyGoogleCloudContext {
     }
 }
 
-#[pyclass(name = "AmazonS3", module = "datafusion.store", subclass)]
+#[pyclass(frozen, name = "AmazonS3", module = "datafusion.store", subclass)]
 #[derive(Debug, Clone)]
 pub struct PyAmazonS3Context {
     pub inner: Arc<AmazonS3>,
@@ -223,7 +228,7 @@ impl PyAmazonS3Context {
     }
 }
 
-#[pyclass(name = "Http", module = "datafusion.store", subclass)]
+#[pyclass(frozen, name = "Http", module = "datafusion.store", subclass)]
 #[derive(Debug, Clone)]
 pub struct PyHttpContext {
     pub url: String,

src/substrait.rs

@@ -27,7 +27,7 @@ use datafusion_substrait::serializer;
 use datafusion_substrait::substrait::proto::Plan;
 use prost::Message;
 
-#[pyclass(name = "Plan", module = "datafusion.substrait", subclass)]
+#[pyclass(frozen, name = "Plan", module = "datafusion.substrait", subclass)]
 #[derive(Debug, Clone)]
 pub struct PyPlan {
     pub plan: Plan,
@@ -59,7 +59,7 @@ impl From<Plan> for PyPlan {
 /// A PySubstraitSerializer is a representation of a Serializer that is capable of both serializing
 /// a `LogicalPlan` instance to Substrait Protobuf bytes and also deserialize Substrait Protobuf bytes
 /// to a valid `LogicalPlan` instance.
-#[pyclass(name = "Serde", module = "datafusion.substrait", subclass)]
+#[pyclass(frozen, name = "Serde", module = "datafusion.substrait", subclass)]
 #[derive(Debug, Clone)]
 pub struct PySubstraitSerializer;
 
@@ -112,7 +112,7 @@ impl PySubstraitSerializer {
     }
 }
 
-#[pyclass(name = "Producer", module = "datafusion.substrait", subclass)]
+#[pyclass(frozen, name = "Producer", module = "datafusion.substrait", subclass)]
 #[derive(Debug, Clone)]
 pub struct PySubstraitProducer;
 
@@ -129,7 +129,7 @@ impl PySubstraitProducer {
     }
 }
 
-#[pyclass(name = "Consumer", module = "datafusion.substrait", subclass)]
+#[pyclass(frozen, name = "Consumer", module = "datafusion.substrait", subclass)]
 #[derive(Debug, Clone)]
 pub struct PySubstraitConsumer;
 

src/udaf.rs

@@ -155,7 +155,7 @@ pub fn to_rust_accumulator(accum: PyObject) -> AccumulatorFactoryFunction {
 }
 
 /// Represents an AggregateUDF
-#[pyclass(name = "AggregateUDF", module = "datafusion", subclass)]
+#[pyclass(frozen, name = "AggregateUDF", module = "datafusion", subclass)]
 #[derive(Debug, Clone)]
 pub struct PyAggregateUDF {
     pub(crate) function: AggregateUDF,

src/udf.rs

@@ -81,7 +81,7 @@ fn to_scalar_function_impl(func: PyObject) -> ScalarFunctionImplementation {
 }
 
 /// Represents a PyScalarUDF
-#[pyclass(name = "ScalarUDF", module = "datafusion", subclass)]
+#[pyclass(frozen, name = "ScalarUDF", module = "datafusion", subclass)]
 #[derive(Debug, Clone)]
 pub struct PyScalarUDF {
     pub(crate) function: ScalarUDF,