Instead of trying to detect notebook vs console, collect one time when we have any kind if ipython environment.

Author: timsaucer

src/dataframe.rs

@@ -49,7 +49,7 @@ use crate::physical_plan::PyExecutionPlan;
 use crate::record_batch::PyRecordBatchStream;
 use crate::sql::logical::PyLogicalPlan;
 use crate::utils::{
-    get_tokio_runtime, py_obj_to_scalar_value, validate_pycapsule, wait_for_future,
+    get_tokio_runtime, is_ipython_env, py_obj_to_scalar_value, validate_pycapsule, wait_for_future,
 };
 use crate::{
     errors::PyDataFusionResult,
@@ -192,12 +192,18 @@ fn build_formatter_config_from_python(formatter: &Bound<'_, PyAny>) -> PyResult<
 #[derive(Clone)]
 pub struct PyDataFrame {
     df: Arc<DataFrame>,
+
+    // In IPython environment cache batches between __repr__ and _repr_html_ calls.
+    batches: Option<(Vec<RecordBatch>, bool)>,
 }
 
 impl PyDataFrame {
     /// creates a new PyDataFrame
     pub fn new(df: DataFrame) -> Self {
-        Self { df: Arc::new(df) }
+        Self {
+            df: Arc::new(df),
+            batches: None,
+        }
     }
 }
 
@@ -224,16 +230,22 @@ impl PyDataFrame {
         }
     }
 
-    fn __repr__(&self, py: Python) -> PyDataFusionResult<String> {
+    fn __repr__(&mut self, py: Python) -> PyDataFusionResult<String> {
         // Get the Python formatter config
         let PythonFormatter {
             formatter: _,
             config,
         } = get_python_formatter_with_config(py)?;
-        let (batches, has_more) = wait_for_future(
-            py,
-            collect_record_batches_to_display(self.df.as_ref().clone(), config),
-        )??;
+
+        let should_cache = *is_ipython_env(py) && self.batches.is_none();
+        let (batches, has_more) = match self.batches.take() {
+            Some(b) => b,
+            None => wait_for_future(
+                py,
+                collect_record_batches_to_display(self.df.as_ref().clone(), config),
+            )??,
+        };
+
         if batches.is_empty() {
             // This should not be reached, but do it for safety since we index into the vector below
             return Ok("No data to display".to_string());
@@ -247,16 +259,27 @@ impl PyDataFrame {
             false => "",
         };
 
+        if should_cache {
+            self.batches = Some((batches, has_more));
+        }
+
         Ok(format!("DataFrame()\n{batches_as_displ}{additional_str}"))
     }
 
-    fn _repr_html_(&self, py: Python) -> PyDataFusionResult<String> {
+    fn _repr_html_(&mut self, py: Python) -> PyDataFusionResult<String> {
         // Get the Python formatter and config
         let PythonFormatter { formatter, config } = get_python_formatter_with_config(py)?;
-        let (batches, has_more) = wait_for_future(
-            py,
-            collect_record_batches_to_display(self.df.as_ref().clone(), config),
-        )??;
+
+        let should_cache = *is_ipython_env(py) && self.batches.is_none();
+
+        let (batches, has_more) = match self.batches.take() {
+            Some(b) => b,
+            None => wait_for_future(
+                py,
+                collect_record_batches_to_display(self.df.as_ref().clone(), config),
+            )??,
+        };
+
         if batches.is_empty() {
             // This should not be reached, but do it for safety since we index into the vector below
             return Ok("No data to display".to_string());
@@ -266,7 +289,7 @@ impl PyDataFrame {
 
         // Convert record batches to PyObject list
         let py_batches = batches
-            .into_iter()
+            .iter()
             .map(|rb| rb.to_pyarrow(py))
             .collect::<PyResult<Vec<PyObject>>>()?;
 
@@ -282,6 +305,10 @@ impl PyDataFrame {
         let html_result = formatter.call_method("format_html", (), Some(&kwargs))?;
         let html_str: String = html_result.extract()?;
 
+        if should_cache {
+            self.batches = Some((batches, has_more));
+        }
+
         Ok(html_str)
     }
 

src/utils.rs

@@ -39,6 +39,17 @@ pub(crate) fn get_tokio_runtime() -> &'static TokioRuntime {
     RUNTIME.get_or_init(|| TokioRuntime(tokio::runtime::Runtime::new().unwrap()))
 }
 
+#[inline]
+pub(crate) fn is_ipython_env(py: Python) -> &'static bool {
+    static IS_IPYTHON_ENV: OnceLock<bool> = OnceLock::new();
+    IS_IPYTHON_ENV.get_or_init(|| {
+        py.import("IPython")
+            .and_then(|ipython| ipython.call_method0("get_ipython"))
+            .map(|ipython| !ipython.is_none())
+            .unwrap_or(false)
+    })
+}
+
 /// Utility to get the Global Datafussion CTX
 #[inline]
 pub(crate) fn get_global_ctx() -> &'static SessionContext {