Benchmark windows and lots of doc clarity updates

benches/bedtools_comparison.rs

@@ -18,7 +18,7 @@ fn bench_range_adjustment(c: &mut Criterion) {
  let input_bedfile = random_bed3file(BED_LENGTH);
 
  // configure the sample size for the group
- // group.sample_size(10);
+ group.sample_size(10);
 
  // bedtools slop
  group.bench_function("bedtools_slop", |b| {
@@ -145,10 +145,54 @@ fn bench_flank(c: &mut Criterion) {
  });
 }
 
+fn bench_windows(c: &mut Criterion) {
+ let width = 100_000;
+ let step = 1_000;
+
+ // create the benchmark group
+ let mut group = c.benchmark_group("windows");
+
+ // configure the sample size for the group
+ // group.sample_size(10);
+ group.bench_function("bedtools_makewindows", |b| {
+ b.iter(|| {
+ let bedtools_output = Command::new("bedtools")
+ .arg("makewindows")
+ .arg("-g")
+ .arg("tests_data/hg38_seqlens.tsv")
+ .arg("-w")
+ .arg(width.to_string())
+ .arg("-s")
+ .arg(step.to_string())
+ .output()
+ .expect("bedtools makewindows failed");
+ assert!(bedtools_output.status.success());
+ });
+ });
+
+ group.bench_function("granges_windows", |b| {
+ b.iter(|| {
+ let granges_output = Command::new(granges_binary_path())
+ .arg("windows")
+ .arg("--genome")
+ .arg("tests_data/hg38_seqlens.tsv")
+ .arg("--width")
+ .arg(width.to_string())
+ .arg("--step")
+ .arg(step.to_string())
+ .output()
+ .expect("granges windows failed");
+ assert!(granges_output.status.success());
+ });
+ });
+}
+
+
 criterion_group!(
  benches,
  bench_filter_adjustment,
  bench_range_adjustment,
  bench_flank,
-);
+ bench_windows,
+ );
 criterion_main!(benches);

src/granges.rs

@@ -632,7 +632,7 @@ where
  /// [`GRanges<VecRangesIndexed, JoinData>`].
  ///
  /// The [`JoinData`] container contains the owned left data container and has
- /// a reference to the right data container, as as well as a [`Vec<LeftGroupedJoin`]
+ /// a reference to the right data container, as as well as a [`Vec<LeftGroupedJoin>`]
  /// that contains information about each overlap between a left and zero or more right
  /// ranges.
  fn left_overlaps(
@@ -674,7 +674,7 @@ where
  /// [`GRanges<VecRangesIndexed, JoinDataRightEmpty>`].
  ///
  /// The [`JoinData`] container contains the left data container and has
- /// a reference to the right data container, as as well as a [`Vec<LeftGroupedJoin`]
+ /// a reference to the right data container, as as well as a [`Vec<LeftGroupedJoin>`]
  /// that contains information about each overlap between a left and zero or more right
  /// ranges.
  fn left_overlaps(
@@ -726,7 +726,7 @@ where
  /// [`GRanges<VecRangesIndexed, JoinDataLeftEmpty>`].
  ///
  /// The [`JoinDataLeftEmpty`] contains no left data, and a reference to the
- /// right data container, as as well as a [`Vec<LeftGroupedJoin`]
+ /// right data container, as as well as a [`Vec<LeftGroupedJoin>`]
  /// that contains information about each overlap between a left and zero or more right
  /// ranges.
  fn left_overlaps(
@@ -778,7 +778,7 @@ where
  /// [`GRanges<VecRangesIndexed, JoinDataBothEmpty>`].
  ///
  /// The [`JoinDataBothEmpty`] contains no data, since neither left of right
- /// [`GRanges`] objects had data. However, it does contain a [`Vec<LeftGroupedJoin`],
+ /// [`GRanges`] objects had data. However, it does contain a [`Vec<LeftGroupedJoin>`],
  /// and each [`LeftGroupedJoin`] contains information about the number of overlapping
  /// ranges and their lengths. This can be used to summarize, e.g. the number
  /// of overlapping basepairs, the overlap fraction, etc.

src/io/parsers.rs

@@ -70,35 +70,6 @@
 //! iterator types defined in [`GenomicRangesParser`]. The best examples of this are the
 //! `granges` subcommands implementations.
 //!
-//! ## File Format Terminology
-//!
-//! - BED3
-//! - BED*
-//! - BED-like
-//!
-//!
-//! : it could be ranges-only (i.e. a BED3), or contain data (e.g. BED5). The lazy BED parser will
-//! output a [`GenomicRangeRecord<Option<String>>`], where the data would be `None` only in the
-//! case that three columns were encountered in the file (which must be a BED3).
-//!
-//! In GRanges, there are two types of ranges: ranges with an index to an element in the data
-//! container, and ranges without indices (i.e. what we would use in the case of processing a BED3
-//! file). Since a [`GRanges`] object needs to have a single, concrete range type in its range
-//! containers, it must be known *at compile time* how one should convert the
-//!
-//! Downstream pipelines must immediately determine how to handle whether there is additional data,
-//! or all of the [`GenomicRangeRecord`] entries are `None`, and
-//!
-//!
-//! # BED-like File Parser Design
-//!
-//! All BED formats (BED3, BED5, etc) are built upon a BED3. Often when working with these types
-//! of formats, many operations do not immediately require full parsing of the line, past the
-//! *range components*. This is because downstream operations may immediately filter away entries
-//! (e.g. based on width), or do an overlap operation, and then filter away entries based on some
-//! overlap criteria. Either way, it may be advantageous to work with ranges that just store some
-//! generic data.
-//!
 //! [`GRanges`]: crate::granges::GRanges
 //! [`GRanges<R, T>`]: crate::granges::GRanges
 //! [`GRangesEmpty`]: crate::granges::GRangesEmpty