Use auto_doc_cfg

This makes it so one doesn't need to manually configure the
documentation cfg for feature flags.

rust-lang/rust#90502

Fixes pola-rs#6123

Author: n8henrie

.github/workflows/docs-deploy.yaml

@@ -38,7 +38,7 @@ jobs:
 
       - name: Build Rust documentation
         env:
-          RUSTFLAGS: --cfg docsrs
+          RUSTDOCFLAGS: --cfg docsrs
         run: cargo doc --features=docs-selection --package polars
 
       - name: Prepare deployment

polars/polars-algo/src/lib.rs

@@ -1,4 +1,4 @@
-#![cfg_attr(docsrs, feature(doc_cfg))]
+#![cfg_attr(docsrs, feature(doc_auto_cfg))]
 mod algo;
 pub use algo::*;
 pub mod prelude;

polars/polars-core/src/chunked_array/mod.rs

@@ -30,18 +30,15 @@ mod from;
 pub(crate) mod list;
 pub(crate) mod logical;
 #[cfg(feature = "object")]
-#[cfg_attr(docsrs, doc(cfg(feature = "object")))]
 pub mod object;
 #[cfg(feature = "random")]
-#[cfg_attr(docsrs, doc(cfg(feature = "random")))]
 mod random;
 pub mod strings;
 #[cfg(any(
     feature = "temporal",
     feature = "dtype-datetime",
     feature = "dtype-date"
 ))]
-#[cfg_attr(docsrs, doc(cfg(feature = "temporal")))]
 pub mod temporal;
 mod trusted_len;
 pub mod upstream_traits;

polars/polars-core/src/chunked_array/ndarray.rs

@@ -9,7 +9,6 @@ where
 {
     /// If data is aligned in a single chunk and has no Null values a zero copy view is returned
     /// as an `ndarray`
-    #[cfg_attr(docsrs, doc(cfg(feature = "ndarray")))]
     pub fn to_ndarray(&self) -> PolarsResult<ArrayView1<T::Native>> {
         let slice = self.cont_slice()?;
         Ok(aview1(slice))
@@ -18,7 +17,6 @@ where
 
 impl ListChunked {
     /// If all nested `Series` have the same length, a 2 dimensional `ndarray::Array` is returned.
-    #[cfg_attr(docsrs, doc(cfg(feature = "ndarray")))]
     pub fn to_ndarray<N>(&self) -> PolarsResult<Array2<N::Native>>
     where
         N: PolarsNumericType,
@@ -96,7 +94,6 @@ impl DataFrame {
     ///  [2.0, 8.0],
     ///  [3.0, 6.0]], shape=[3, 2], strides=[2, 1], layout=C (0x1), const ndim=2/
     /// ```
-    #[cfg_attr(docsrs, doc(cfg(feature = "ndarray")))]
     pub fn to_ndarray<N>(&self) -> PolarsResult<Array2<N::Native>>
     where
         N: PolarsNumericType,

polars/polars-core/src/chunked_array/ops/mod.rs

@@ -476,7 +476,6 @@ pub trait ChunkUnique<T: PolarsDataType> {
 
     /// The most occurring value(s). Can return multiple Values
     #[cfg(feature = "mode")]
-    #[cfg_attr(docsrs, doc(cfg(feature = "mode")))]
     fn mode(&self) -> PolarsResult<ChunkedArray<T>> {
         Err(PolarsError::InvalidOperation(
             "mode is not implemented for this dtype".into(),
@@ -728,7 +727,6 @@ pub trait ChunkPeaks {
 
 /// Check if element is member of list array
 #[cfg(feature = "is_in")]
-#[cfg_attr(docsrs, doc(cfg(feature = "is_in")))]
 pub trait IsIn {
     /// Check if elements of this array are in the right Series, or List values of the right Series.
     fn is_in(&self, _other: &Series) -> PolarsResult<BooleanChunked> {
@@ -750,7 +748,6 @@ pub trait ArgAgg {
 
 /// Repeat the values `n` times.
 #[cfg(feature = "repeat_by")]
-#[cfg_attr(docsrs, doc(cfg(feature = "repeat_by")))]
 pub trait RepeatBy {
     /// Repeat the values `n` times, where `n` is determined by the values in `by`.
     fn repeat_by(&self, _by: &IdxCa) -> ListChunked {
@@ -759,7 +756,6 @@ pub trait RepeatBy {
 }
 
 #[cfg(feature = "is_first")]
-#[cfg_attr(docsrs, doc(cfg(feature = "is_first")))]
 /// Mask the first unique values as `true`
 pub trait IsFirst<T: PolarsDataType> {
     fn is_first(&self) -> PolarsResult<BooleanChunked> {
@@ -770,7 +766,6 @@ pub trait IsFirst<T: PolarsDataType> {
 }
 
 #[cfg(feature = "is_first")]
-#[cfg_attr(docsrs, doc(cfg(feature = "is_first")))]
 /// Mask the last unique values as `true`
 pub trait IsLast<T: PolarsDataType> {
     fn is_last(&self) -> PolarsResult<BooleanChunked> {
@@ -781,7 +776,6 @@ pub trait IsLast<T: PolarsDataType> {
 }
 
 #[cfg(feature = "concat_str")]
-#[cfg_attr(docsrs, doc(cfg(feature = "concat_str")))]
 /// Concat the values into a string array.
 pub trait StrConcat {
     /// Concat the values into a string array.

polars/polars-core/src/datatypes/mod.rs

@@ -119,13 +119,11 @@ impl PolarsDataType for ListType {
 }
 
 #[cfg(feature = "object")]
-#[cfg_attr(docsrs, doc(cfg(feature = "object")))]
 pub struct ObjectType<T>(T);
 #[cfg(feature = "object")]
 pub type ObjectChunked<T> = ChunkedArray<ObjectType<T>>;
 
 #[cfg(feature = "object")]
-#[cfg_attr(docsrs, doc(cfg(feature = "object")))]
 impl<T: PolarsObject> PolarsDataType for ObjectType<T> {
     fn get_dtype() -> DataType {
         DataType::Object(T::type_name())

polars/polars-core/src/frame/asof_join/groups.rs

@@ -619,7 +619,6 @@ fn dispatch_join<T: PolarsNumericType>(
 }
 
 impl DataFrame {
-    #[cfg_attr(docsrs, doc(cfg(feature = "asof_join")))]
     #[allow(clippy::too_many_arguments)]
     #[doc(hidden)]
     pub fn _join_asof_by(
@@ -704,7 +703,6 @@ impl DataFrame {
     /// This is similar to a left-join except that we match on nearest key rather than equal keys.
     /// The keys must be sorted to perform an asof join. This is a special implementation of an asof join
     /// that searches for the nearest keys within a subgroup set by `by`.
-    #[cfg_attr(docsrs, doc(cfg(feature = "asof_join")))]
     #[allow(clippy::too_many_arguments)]
     pub fn join_asof_by<I, S>(
         &self,

polars/polars-core/src/frame/asof_join/mod.rs

@@ -196,7 +196,6 @@ impl DataFrame {
 
     /// This is similar to a left-join except that we match on nearest key rather than equal keys.
     /// The keys must be sorted to perform an asof join
-    #[cfg_attr(docsrs, doc(cfg(feature = "asof_join")))]
     pub fn join_asof(
         &self,
         other: &DataFrame,

polars/polars-core/src/frame/mod.rs

@@ -2778,7 +2778,6 @@ impl DataFrame {
 
     /// Aggregate the column horizontally to their min values.
     #[cfg(feature = "zip_with")]
-    #[cfg_attr(docsrs, doc(cfg(feature = "zip_with")))]
     pub fn hmin(&self) -> PolarsResult<Option<Series>> {
         let min_fn = |acc: &Series, s: &Series| {
             let mask = acc.lt(s)? & acc.is_not_null() | s.is_null();
@@ -2808,7 +2807,6 @@ impl DataFrame {
 
     /// Aggregate the column horizontally to their max values.
     #[cfg(feature = "zip_with")]
-    #[cfg_attr(docsrs, doc(cfg(feature = "zip_with")))]
     pub fn hmax(&self) -> PolarsResult<Option<Series>> {
         let max_fn = |acc: &Series, s: &Series| {
             let mask = acc.gt(s)? & acc.is_not_null() | s.is_null();
@@ -3305,7 +3303,6 @@ impl DataFrame {
 
     /// Split into multiple DataFrames partitioned by groups
     #[cfg(feature = "partition_by")]
-    #[cfg_attr(docsrs, doc(cfg(feature = "partition_by")))]
     pub fn partition_by(&self, cols: impl IntoVec<String>) -> PolarsResult<Vec<DataFrame>> {
         let cols = cols.into_vec();
         self._partition_by_impl(&cols, false)
@@ -3314,7 +3311,6 @@ impl DataFrame {
     /// Split into multiple DataFrames partitioned by groups
     /// Order of the groups are maintained.
     #[cfg(feature = "partition_by")]
-    #[cfg_attr(docsrs, doc(cfg(feature = "partition_by")))]
     pub fn partition_by_stable(&self, cols: impl IntoVec<String>) -> PolarsResult<Vec<DataFrame>> {
         let cols = cols.into_vec();
         self._partition_by_impl(&cols, true)
@@ -3323,7 +3319,6 @@ impl DataFrame {
     /// Unnest the given `Struct` columns. This means that the fields of the `Struct` type will be
     /// inserted as columns.
     #[cfg(feature = "dtype-struct")]
-    #[cfg_attr(docsrs, doc(cfg(feature = "dtype-struct")))]
     pub fn unnest<I: IntoVec<String>>(&self, cols: I) -> PolarsResult<DataFrame> {
         let cols = cols.into_vec();
         self.unnest_impl(cols.into_iter().collect())

polars/polars-core/src/frame/row.rs

@@ -18,7 +18,6 @@ impl<'a> Row<'a> {
 
 impl DataFrame {
     /// Get a row from a DataFrame. Use of this is discouraged as it will likely be slow.
-    #[cfg_attr(docsrs, doc(cfg(feature = "rows")))]
     pub fn get_row(&self, idx: usize) -> PolarsResult<Row> {
         let values = self
             .columns
@@ -31,7 +30,6 @@ impl DataFrame {
     /// Amortize allocations by reusing a row.
     /// The caller is responsible to make sure that the row has at least the capacity for the number
     /// of columns in the DataFrame
-    #[cfg_attr(docsrs, doc(cfg(feature = "rows")))]
     pub fn get_row_amortized<'a>(&'a self, idx: usize, row: &mut Row<'a>) -> PolarsResult<()> {
         for (s, any_val) in self.columns.iter().zip(&mut row.0) {
             *any_val = s.get(idx)?;
@@ -46,7 +44,6 @@ impl DataFrame {
     /// # Safety
     /// Does not do any bounds checking.
     #[inline]
-    #[cfg_attr(docsrs, doc(cfg(feature = "rows")))]
     pub unsafe fn get_row_amortized_unchecked<'a>(&'a self, idx: usize, row: &mut Row<'a>) {
         self.columns
             .iter()
@@ -58,14 +55,12 @@ impl DataFrame {
 
     /// Create a new DataFrame from rows. This should only be used when you have row wise data,
     /// as this is a lot slower than creating the `Series` in a columnar fashion
-    #[cfg_attr(docsrs, doc(cfg(feature = "rows")))]
     pub fn from_rows_and_schema(rows: &[Row], schema: &Schema) -> PolarsResult<Self> {
         Self::from_rows_iter_and_schema(rows.iter(), schema)
     }
 
     /// Create a new DataFrame from an iterator over rows. This should only be used when you have row wise data,
     /// as this is a lot slower than creating the `Series` in a columnar fashion
-    #[cfg_attr(docsrs, doc(cfg(feature = "rows")))]
     pub fn from_rows_iter_and_schema<'a, I>(mut rows: I, schema: &Schema) -> PolarsResult<Self>
     where
         I: Iterator<Item = &'a Row<'a>>,
@@ -108,7 +103,6 @@ impl DataFrame {
 
     /// Create a new DataFrame from rows. This should only be used when you have row wise data,
     /// as this is a lot slower than creating the `Series` in a columnar fashion
-    #[cfg_attr(docsrs, doc(cfg(feature = "rows")))]
     pub fn from_rows(rows: &[Row]) -> PolarsResult<Self> {
         let schema = rows_to_schema_first_non_null(rows, Some(50));
         let has_nulls = schema
@@ -177,7 +171,6 @@ impl DataFrame {
         }
     }
 
-    #[cfg_attr(docsrs, doc(cfg(feature = "rows")))]
     /// Transpose a DataFrame. This is a very expensive operation.
     pub fn transpose(&self) -> PolarsResult<DataFrame> {
         let height = self.height();