Open compressed files transparently
This Python module provides an xopen function that works like the built-in open function but also transparently deals with compressed files. Supported compression formats are currently gzip, bzip2, xz and optionally Zstandard.
xopen selects the most efficient method for reading or writing a compressed file. This often means opening a pipe to an external tool, such as pigz, which is a parallel version of gzip, or igzip, which is a highly optimized version of gzip.
If threads=0 is passed to xopen(), no external process is used. For gzip files, this will then use python-isal (which binds isa-l) if it is installed (since python-isal is a dependency of xopen, this should always be the case). Neither igzip nor python-isal support compression levels greater 3, so if no external tool is available or threads has been set to 0, Python’s built-in gzip.open is used.
For xz files, a pipe to the xz program is used because it has built-in support for multithreaded compression.
For bz2 files, pbzip2 (parallel bzip2) is used.
xopen falls back to Python’s built-in functions (gzip.open, lzma.open, bz2.open) if none of the other methods can be used.
The file format to use is determined from the file name if the extension is recognized (.gz, .bz2, .xz or .zst). When reading a file without a recognized file extension, xopen attempts to detect the format by reading the first couple of bytes from the file.
xopen is compatible with Python versions 3.7 and later.
Open a file for reading:
from xopen import xopen with xopen("file.txt.gz") as f: content = f.read()
Write to a file in binary mode, set the compression level and avoid using an external process:
from xopen import xopen with xopen("file.txt.xz", mode="wb", threads=0, compresslevel=3) f.write(b"Hello")
xopen writes gzip files in a reproducible manner.
Normally, gzip files contain a timestamp in the file header, which means that compressing the same data at different times results in different output files. xopen disables this for all of the supported gzip compression backends. For example, when using an external process, it sets the command-line option --no-name (same as -n).
Note that different gzip compression backends typically do not produce identical output, so reproducibility may no longer be given when the execution environment changes from one xopen() invocation to the next. This includes the CPU architecture as igzip adjusts its algorithm depending on it.
bzip2 and xz compression methods do not store timestamps in the file headers, so output from them is also reproducible.
Optional Zstandard support
For reading and writing Zstandard (.zst) files, either the zstd command-line program or the Python zstandard package needs to be installed.
If the threads parameter to xopen() is None (the default) or any value greater than 0, xopen uses an external zstd process.
If the above fails (because no zstd program is available) or if threads is 0, the zstandard package is used.
To ensure that you get the correct zstandard version, you can specify the zstd extra for xopen, that is, install it using pip install xopen[zstd].
#91: Added optional support for Zstandard (.zst) files. This requires that the Python zstandard package is installed or that the zstd command-line program is available.
#94: When writing gzip files, the timestamp and name of the original file is omitted (equivalent to using gzip --no-name (or -n) on the command line). This allows files to be written in a reproducible manner.
#100: Dropped Python 3.6 support
#101: Added support for piping into and from an external xz process. Contributed by @fanninpm.
#102: Support setting the xz compression level. Contributed by @tsibley.
Add seek() and tell() to the PipedCompressionReader classes (for Windows compatibility)
xopen is now available on Windows (in addition to Linux and macOS).
For greater compatibility with the built-in open() function, xopen() has gained the parameters encoding, errors and newlines with the same meaning as in open(). Unlike built-in open(), though, encoding is UTF-8 by default.
A parameter format has been added that allows to force the compression file format.
pbzip2 is now used to open .bz2 files if threads is greater than zero (contributed by @DriesSchaumont).
Python 3.5 support is dropped.
On Linux systems, python-isal is now added as a requirement. This will speed up the reading of gzip files significantly when no external processes are used.
If installed, the igzip program (part of Intel ISA-L) is now used for reading and writing gzip-compressed files at compression levels 1-3, which results in a significant speedup.
#80: When the file name extension of a file to be opened for reading is not available, the content is inspected (if possible) and used to determine which compression format applies (contributed by @bvaisvil).
This release drops Python 2.7 and 3.4 support. Python 3.5 or later is now required.
When reading gzipped files, force pigz to use only a single process. pigz cannot use multiple cores anyway when decompressing. By default, it would use extra I/O processes, which slightly reduces wall-clock time, but increases CPU time. Single-core decompression with pigz is still about twice as fast as regular gzip.
Allow threads=0 for specifying that no external pigz/gzip process should be used (then regular gzip.open() is used instead).
#20: When reading gzipped files, let pigz use at most four threads by default. This limit previously only applied when writing to a file. Contributed by @bernt-matthias.
Support Python 3.8
#14: Speed improvements when iterating over gzipped files.
For reading from gzipped files, xopen will now use a pigz subprocess. This is faster than using gzip.open.
Python 2 support will be dropped in one of the next releases.
By default, pigz is now only allowed to use at most four threads. This hopefully reduces problems some users had with too many threads when opening many files at the same time.
xopen now accepts pathlib.Path objects.
Drop Python 3.3 support
Add a threads parameter (passed on to pigz)
#6: Make multi-block bz2 work on Python 2 by using external bz2file library.
Drop Python 2.6 support
#5: Fix PipedGzipReader.read() not returning anything
Add gzip compression parameter
#3: Allow appending to bz2 and lzma files where possible
Fix a deadlock
The name xopen was taken from the C function of the same name in the utils.h file which is part of BWA.
Some ideas were taken from the canopener project. If you also want to open S3 files, you may want to use that module instead.
@kyleabeauchamp contributed support for appending to files before this repository was created.
For a list of contributors, see <https://github.com/pycompression/xopen/graphs/contributors>
Release history Release notifications | RSS feed
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.