Make --compression-level=1 the default

Closes #808

CHANGES.rst

@@ -5,6 +5,17 @@ Changelog
 development version
 -------------------
 
+* :issue:`808`: Made gzip compression level 1 the default, which improves
+  runtime significantly in many cases. (Compressing the output is often a
+  bottleneck when using multiple threads.) Output files will be larger, but because
+  Cutadapt is typically used in intermediate step where the output would be
+  deleted, this may often not have an impact on final disk usage.
+  Suggested by @rhpvorderman (many times). Use ``--compression-level=N`` with
+  ``N`` greater than ``N`` to get higher compression at the cost of speed.
+  (``N=5`` is the old default.)
+  Option ``-Z`` (equivalent to ``--compression-level=1``) is now deprecated.
+* The previously hidden option ``--compression-level`` is now shown in the
+  ``--help`` output.
 * :issue:`820`: On Bioconda, Cutadapt is now also available for ARM64 Macs (M1/M2).
 * Dropped support for Python 3.8.
 * Added support for Python 3.13.

doc/recipes.rst

@@ -191,19 +191,12 @@ Speeding up Cutadapt
 
 There are several tricks for limiting wall-clock time while using Cutadapt.
 
-Option ``-Z`` (equivalent to ``--compression-level=1``) can be used to limit the
-amount of CPU time which is spent on the compression of output files.
-Alternatively, choosing filenames not ending with ``.gz``, ``.bz2`` or ``.xz``
-will make sure no CPU time is spent on compression at all. On systems
-with slow I/O, it can actually be faster to set a higher compression-level
-than 1.
-
 Increasing the number of cores with ``-j`` will increase the number of reads per
 minute at near-linear rate.
 
-It is also possible to use pipes in order to bypass the filesystem and pipe
+It is possible to use pipes in order to bypass the filesystem and pipe
 Cutadapt's output into an aligner such as BWA. The ``mkfifo`` command allows
-you to create named pipes in bash.
+you to create named pipes.
 
 .. code-block::bash
 
@@ -214,6 +207,10 @@ you to create named pipes in bash.
 This command will run cutadapt and BWA simultaneously, using Cutadapt’s output as
 BWA’s input, and capturing Cutadapt’s report in ``cutadapt.log``.
 
+.. versionadded:: 4.10
+   Option ``-Z`` (equivalent to ``--compression-level=1``), which was earlier
+   recommended for speeding up processing, is now the default.
+
 
 Check whether a FASTQ file is properly formatted
 ------------------------------------------------

doc/reference.rst

@@ -348,6 +348,8 @@ Output
     Use this option to force FASTA even in such a case.
 
 ``-Z``
+    **Deprecated**: This option has become the default in Cutadapt 4.10.
+
     Use compression level 1 for gzipped output files.
     This is a shorthand for ``--compression-level=1``.
 

src/cutadapt/cli.py

@@ -186,9 +186,6 @@ def get_argument_parser() -> ArgumentParser:
     # Buffer size for the reader process when running in parallel
     group.add_argument("--buffer-size", type=int, default=4000000,
         help=SUPPRESS)
-    # Compression level for gzipped output files. Not exposed since we have -Z
-    group.add_argument("--compression-level", type=int, default=5,
-        help=SUPPRESS)
     # Disable adapter index creation
     group.add_argument("--no-index", dest="index", default=True, action="store_false", help=SUPPRESS)
 
@@ -344,8 +341,10 @@ def get_argument_parser() -> ArgumentParser:
             "Default: write to standard output")
     group.add_argument("--fasta", default=False, action='store_true',
         help="Output FASTA to standard output even on FASTQ input.")
+    group.add_argument("--compression-level", type=int, default=1, metavar="N",
+        help="Compression level for compressed output files. Default: %(default)s")
     group.add_argument("-Z", action="store_const", const=1, dest="compression_level",
-        help="Use compression level 1 for gzipped output files (faster, but uses more space)")
+        help="DEPRECATED because compression level 1 is now the default.")
     group.add_argument("--info-file", metavar="FILE",
         help="Write information about each read and its adapter matches into FILE. "
             "See the documentation for the file format.")