chore: add docs, part of #37 (apache#6433)

* chore: add docs, part of #37

- add pragma `#![warn(missing_docs)]` to the following
  - `arrow-array`
  - `arrow-cast`
  - `arrow-csv`
  - `arrow-data`
  - `arrow-json`
  - `arrow-ord`
  - `arrow-pyarrow-integration-testing`
  - `arrow-row`
  - `arrow-schema`
  - `arrow-select`
  - `arrow-string`
  - `arrow`
  - `parquet_derive`

- add docs to those that generated lint warnings

- Remove `bitflags` workaround in `arrow-schema`
At some point, a change in `bitflags v2.3.0` had
started generating lint warnings in `arrow-schema`,

This was handled using a
[workaround](apache#4233)

[Issue](bitflags/bitflags#356)

`bitflags v2.3.1` fixed the issue hence the
workaround is no longer needed.

* fix: resolve comments on PR apache#6433

Author: ByteBaker

arrow-array/src/builder/generic_bytes_builder.rs

@@ -537,7 +537,7 @@ mod tests {
         write!(builder, "buz").unwrap();
         builder.append_value("");
         let a = builder.finish();
-        let r: Vec<_> = a.iter().map(|x| x.unwrap()).collect();
+        let r: Vec<_> = a.iter().flatten().collect();
         assert_eq!(r, &["foo", "bar\n", "fizbuz"])
     }
 }

arrow-array/src/builder/generic_bytes_view_builder.rs

@@ -555,6 +555,8 @@ pub fn make_view(data: &[u8], block_id: u32, offset: u32) -> u128 {
 
 #[cfg(test)]
 mod tests {
+    use core::str;
+
     use super::*;
     use crate::Array;
 
@@ -642,7 +644,7 @@ mod tests {
         let array = v.finish_cloned();
         array.to_data().validate_full().unwrap();
         assert_eq!(array.data_buffers().len(), 5);
-        let actual: Vec<_> = array.iter().map(Option::unwrap).collect();
+        let actual: Vec<_> = array.iter().flatten().collect();
         assert_eq!(
             actual,
             &[
@@ -692,13 +694,13 @@ mod tests {
         let mut exp_builder = StringViewBuilder::new();
         let mut fixed_builder = StringViewBuilder::new().with_fixed_block_size(STARTING_BLOCK_SIZE);
 
-        let long_string = String::from_utf8(vec![b'a'; STARTING_BLOCK_SIZE as usize]).unwrap();
+        let long_string = str::from_utf8(&[b'a'; STARTING_BLOCK_SIZE as usize]).unwrap();
 
         for i in 0..9 {
             // 8k, 16k, 32k, 64k, 128k, 256k, 512k, 1M, 2M
             for _ in 0..(2_u32.pow(i)) {
-                exp_builder.append_value(&long_string);
-                fixed_builder.append_value(&long_string);
+                exp_builder.append_value(long_string);
+                fixed_builder.append_value(long_string);
             }
             exp_builder.flush_in_progress();
             fixed_builder.flush_in_progress();
@@ -721,7 +723,7 @@ mod tests {
         }
 
         // Add one more value, and the buffer stop growing.
-        exp_builder.append_value(&long_string);
+        exp_builder.append_value(long_string);
         exp_builder.flush_in_progress();
         assert_eq!(
             exp_builder.completed.last().unwrap().capacity(),

arrow-cast/src/cast/mod.rs

@@ -9394,7 +9394,7 @@ mod tests {
             Some(vec![Some(0), None, Some(2)]),
         ]);
         let a = cast_with_options(&array, &DataType::Utf8, &options).unwrap();
-        let r: Vec<_> = a.as_string::<i32>().iter().map(|x| x.unwrap()).collect();
+        let r: Vec<_> = a.as_string::<i32>().iter().flatten().collect();
         assert_eq!(r, &["[0, 1, 2]", "[0, null, 2]"]);
     }
     #[test]

arrow-csv/src/writer.rs

@@ -442,6 +442,7 @@ mod tests {
     };
     use arrow_array::types::*;
     use arrow_buffer::i256;
+    use core::str;
     use std::io::{Cursor, Read, Seek};
     use std::sync::Arc;
 
@@ -508,7 +509,7 @@ Lorem ipsum dolor sit amet,123.564532,3,true,,00:20:34,cupcakes
 consectetur adipiscing elit,,2,false,2019-04-18T10:54:47.378,06:51:20,cupcakes
 sed do eiusmod tempor,-556132.25,1,,2019-04-18T02:45:55.555,23:46:03,foo
 "#;
-        assert_eq!(expected.to_string(), String::from_utf8(buffer).unwrap());
+        assert_eq!(expected, str::from_utf8(&buffer).unwrap());
     }
 
     #[test]
@@ -558,7 +559,7 @@ sed do eiusmod tempor,-556132.25,1,,2019-04-18T02:45:55.555,23:46:03,foo
 ,
 0.290472,0.290472
 "#;
-        assert_eq!(expected.to_string(), String::from_utf8(buffer).unwrap());
+        assert_eq!(expected, str::from_utf8(&buffer).unwrap());
     }
 
     #[test]

arrow-data/src/byte_view.rs

@@ -40,6 +40,7 @@ pub struct ByteView {
 
 impl ByteView {
     #[inline(always)]
+    /// Convert `ByteView` to `u128` by concatenating the fields
     pub fn as_u128(self) -> u128 {
         (self.length as u128)
             | ((self.prefix as u128) << 32)

arrow-data/src/data.rs

@@ -231,6 +231,7 @@ pub struct ArrayData {
     nulls: Option<NullBuffer>,
 }
 
+/// A thread-safe, shared reference to the Arrow array data.
 pub type ArrayDataRef = Arc<ArrayData>;
 
 impl ArrayData {
@@ -1747,7 +1748,12 @@ pub enum BufferSpec {
     /// for array slicing and interoperability with `Vec`, which cannot be over-aligned.
     ///
     /// Note that these alignment requirements will vary between architectures
-    FixedWidth { byte_width: usize, alignment: usize },
+    FixedWidth {
+        /// The width of each element in bytes
+        byte_width: usize,
+        /// The alignment required by Rust for an array of the corresponding primitive
+        alignment: usize,
+    },
     /// Variable width, such as string data for utf8 data
     VariableWidth,
     /// Buffer holds a bitmap.
@@ -1783,6 +1789,7 @@ pub struct ArrayDataBuilder {
 
 impl ArrayDataBuilder {
     #[inline]
+    /// Creates a new array data builder
     pub const fn new(data_type: DataType) -> Self {
         Self {
             data_type,
@@ -1796,61 +1803,72 @@ impl ArrayDataBuilder {
         }
     }
 
+    /// Creates a new array data builder from an existing one, changing the data type
     pub fn data_type(self, data_type: DataType) -> Self {
         Self { data_type, ..self }
     }
 
     #[inline]
     #[allow(clippy::len_without_is_empty)]
+    /// Sets the length of the [ArrayData]
     pub const fn len(mut self, n: usize) -> Self {
         self.len = n;
         self
     }
 
+    /// Sets the null buffer of the [ArrayData]
     pub fn nulls(mut self, nulls: Option<NullBuffer>) -> Self {
         self.nulls = nulls;
         self.null_count = None;
         self.null_bit_buffer = None;
         self
     }
 
+    /// Sets the null count of the [ArrayData]
     pub fn null_count(mut self, null_count: usize) -> Self {
         self.null_count = Some(null_count);
         self
     }
 
+    /// Sets the `null_bit_buffer` of the [ArrayData]
     pub fn null_bit_buffer(mut self, buf: Option<Buffer>) -> Self {
         self.nulls = None;
         self.null_bit_buffer = buf;
         self
     }
 
+    /// Sets the offset of the [ArrayData]
     #[inline]
     pub const fn offset(mut self, n: usize) -> Self {
         self.offset = n;
         self
     }
 
+    /// Sets the buffers of the [ArrayData]
     pub fn buffers(mut self, v: Vec<Buffer>) -> Self {
         self.buffers = v;
         self
     }
 
+    /// Adds a single buffer to the [ArrayData]'s buffers
     pub fn add_buffer(mut self, b: Buffer) -> Self {
         self.buffers.push(b);
         self
     }
 
-    pub fn add_buffers(mut self, bs: Vec<Buffer>) -> Self {
+    /// Adds multiple buffers to the [ArrayData]'s buffers
+    pub fn add_buffers<I: IntoIterator<Item = Buffer>>(mut self, bs: I) -> Self {
         self.buffers.extend(bs);
         self
     }
 
+    /// Sets the child data of the [ArrayData]
     pub fn child_data(mut self, v: Vec<ArrayData>) -> Self {
         self.child_data = v;
         self
     }
 
+    /// Adds a single child data to the [ArrayData]'s child data
     pub fn add_child_data(mut self, r: ArrayData) -> Self {
         self.child_data.push(r);
         self
@@ -1873,22 +1891,25 @@ impl ArrayDataBuilder {
 
     /// Same as [`Self::build_unchecked`] but ignoring `force_validate` feature flag
     unsafe fn build_impl(self) -> ArrayData {
-        let nulls = self.nulls.or_else(|| {
-            let buffer = self.null_bit_buffer?;
-            let buffer = BooleanBuffer::new(buffer, self.offset, self.len);
-            Some(match self.null_count {
-                Some(n) => NullBuffer::new_unchecked(buffer, n),
-                None => NullBuffer::new(buffer),
+        let nulls = self
+            .nulls
+            .or_else(|| {
+                let buffer = self.null_bit_buffer?;
+                let buffer = BooleanBuffer::new(buffer, self.offset, self.len);
+                Some(match self.null_count {
+                    Some(n) => NullBuffer::new_unchecked(buffer, n),
+                    None => NullBuffer::new(buffer),
+                })
             })
-        });
+            .filter(|b| b.null_count() != 0);
 
         ArrayData {
             data_type: self.data_type,
             len: self.len,
             offset: self.offset,
             buffers: self.buffers,
             child_data: self.child_data,
-            nulls: nulls.filter(|b| b.null_count() != 0),
+            nulls,
         }
     }
 

arrow-data/src/decimal.rs

@@ -15,6 +15,10 @@
 // specific language governing permissions and limitations
 // under the License.
 
+//! Defines maximum and minimum values for `decimal256` and `decimal128` types for varying precisions.
+//!
+//! Also provides functions to validate if a given decimal value is within the valid range of the decimal type.
+
 use arrow_buffer::i256;
 use arrow_schema::ArrowError;
 

arrow-data/src/lib.rs

@@ -19,6 +19,7 @@
 //!
 //! For a higher-level, strongly-typed interface see [arrow_array](https://docs.rs/arrow_array)
 
+#![warn(missing_docs)]
 mod data;
 pub use data::*;
 

arrow-data/src/transform/mod.rs

@@ -15,6 +15,11 @@
 // specific language governing permissions and limitations
 // under the License.
 
+//! Low-level array data abstractions.
+//!
+//! Provides utilities for creating, manipulating, and converting Arrow arrays
+//! made of primitive types, strings, and nested types.
+
 use super::{data::new_buffers, ArrayData, ArrayDataBuilder, ByteView};
 use crate::bit_mask::set_bits;
 use arrow_buffer::buffer::{BooleanBuffer, NullBuffer};

arrow-json/src/reader/mod.rs

@@ -1007,7 +1007,7 @@ mod tests {
         let map_values = map.values().as_list::<i32>();
         assert_eq!(map.value_offsets(), &[0, 1, 3, 5]);
 
-        let k: Vec<_> = map_keys.iter().map(|x| x.unwrap()).collect();
+        let k: Vec<_> = map_keys.iter().flatten().collect();
         assert_eq!(&k, &["a", "a", "b", "c", "a"]);
 
         let list_values = map_values.values().as_string::<i32>();

arrow-ord/src/lib.rs

@@ -43,6 +43,7 @@
 //! ```
 //!
 
+#![warn(missing_docs)]
 pub mod cmp;
 #[doc(hidden)]
 pub mod comparison;

arrow-ord/src/rank.rs

@@ -15,6 +15,8 @@
 // specific language governing permissions and limitations
 // under the License.
 
+//! Provides `rank` function to assign a rank to each value in an array
+
 use arrow_array::cast::AsArray;
 use arrow_array::types::*;
 use arrow_array::{downcast_primitive_array, Array, ArrowNativeTypeOp, GenericByteArray};

arrow-ord/src/sort.rs

@@ -635,7 +635,9 @@ where
 /// One column to be used in lexicographical sort
 #[derive(Clone, Debug)]
 pub struct SortColumn {
+    /// The column to sort
     pub values: ArrayRef,
+    /// Sort options for this column
     pub options: Option<SortOptions>,
 }
 

arrow-pyarrow-integration-testing/src/lib.rs

@@ -18,6 +18,7 @@
 //! This library demonstrates a minimal usage of Rust's C data interface to pass
 //! arrays from and to Python.
 
+#![warn(missing_docs)]
 use std::sync::Arc;
 
 use arrow::array::new_empty_array;

arrow-row/src/lib.rs

@@ -125,6 +125,7 @@
 //! [compared]: PartialOrd
 //! [compare]: PartialOrd
 
+#![warn(missing_docs)]
 use std::cmp::Ordering;
 use std::hash::{Hash, Hasher};
 use std::sync::Arc;

arrow-schema/src/datatype.rs

@@ -420,11 +420,13 @@ pub enum IntervalUnit {
     MonthDayNano,
 }
 
-// Sparse or Dense union layouts
+/// Sparse or Dense union layouts
 #[derive(Debug, Clone, PartialEq, Eq, Hash, PartialOrd, Ord, Copy)]
 #[cfg_attr(feature = "serde", derive(serde::Serialize, serde::Deserialize))]
 pub enum UnionMode {
+    /// Sparse union layout
     Sparse,
+    /// Dense union layout
     Dense,
 }
 

arrow-schema/src/error.rs

@@ -26,23 +26,39 @@ use std::error::Error;
 pub enum ArrowError {
     /// Returned when functionality is not yet available.
     NotYetImplemented(String),
+    /// Wraps an external error.
     ExternalError(Box<dyn Error + Send + Sync>),
+    /// Error during casting from one type to another.
     CastError(String),
+    /// Memory or buffer error.
     MemoryError(String),
+    /// Error during parsing from a string.
     ParseError(String),
+    /// Error during schema-related operations.
     SchemaError(String),
+    /// Error during computation.
     ComputeError(String),
+    /// Error during division by zero.
     DivideByZero,
+    /// Error when an arithmetic operation overflows.
     ArithmeticOverflow(String),
+    /// Error during CSV-related operations.
     CsvError(String),
+    /// Error during JSON-related operations.
     JsonError(String),
+    /// Error during IO operations.
     IoError(String, std::io::Error),
+    /// Error during IPC operations in `arrow-ipc` or `arrow-flight`.
     IpcError(String),
+    /// Error indicating that an unexpected or bad argument was passed to a function.
     InvalidArgumentError(String),
+    /// Error during Parquet operations.
     ParquetError(String),
     /// Error during import or export to/from the C Data Interface
     CDataInterface(String),
+    /// Error when a dictionary key is bigger than the key type
     DictionaryKeyOverflowError,
+    /// Error when the run end index in a REE array is bigger than the array length
     RunEndIndexOverflowError,
 }