Skip to main content

figwig

PyPI Downloads Unit Tests Documentation Status

[docs][release notes]

A fast, multithreaded reader and writer of bigWig files, into and out of numpy. Built on work from Nezar Abdennur and Jack Huey.

Loading training data for a genomics model means reading the signal under tens or hundreds of thousands of windows: 1,000 bp around every peak and background region of an experiment, say. figwig reads all of them in one call, straight into a float32 array, from one bigWig or from several at once, and writes bigWigs from the same windows, such as a model's predictions, or from intervals. The common readers make one call per window, and each call holds Python's GIL, so adding threads does not help them. figwig decompresses and compresses with zlib or libdeflate and decodes with numba, all without the GIL, so its threads run in parallel. It depends only on numpy and numba.

Its command, figwig bam2bw, turns SAM/BAM files of reads, or BED/tsv files of fragments, into bigWigs of per-base counts. It is bam2bw with the same arguments, outputs and messages, with faster readers.

figwig also defines its own file format, the figwig file, which holds any number of per-base tracks over one genome with a JSON record of attributes for each, and reads windows from any subset of them in one call, as integers, into (windows, tracks, width). figwig bw2fw, fw2bw and bam2fw convert bigWigs and reads to and from it.

figwig fastest other tool
Read 167,750 windows of 1,000 bp from an ATAC-seq bigWig 0.135 s on 8 threads 1.25 s, pybigtools
Write 15.9 million per-base counts 0.18 s on 8 threads 4.20 s, pyBigWig
Write 21.2 million intervals, with zoom levels 0.55 s on 8 threads 5.47 s, pybigtools
Convert a 2.4 GB ATAC-seq BAM to two stranded bigWigs 4.65 s, 910 MB peak, at -p 2 58.48 s, 3,039 MB peak, bam2bw 0.5.1
The same BAM's per-base coverage, like deepTools 3.4 s at -p 8 14.9 s on 8 processes, bamCoverage --binSize 1, the same value at every base
Read 8,192 windows of 2,048 bp from 48 count tracks on chr1 0.085 s on 8 threads, 0.044 s on 32, from one figwig file 0.27 s and 0.11 s from 48 bigWigs, figwig; 5.9 s pybigtools; 4.0 s zarr
A DataLoader batch of 64 windows x 48 tracks as a CUDA tensor 4 ms from one figwig file 30 ms from the bigWigs, figwig; 120 ms pybigtools
10 pseudobulks of 10,246 cells over 8,192 windows 0.30 s on 8 threads 5.1 s from an in-memory scipy CSR; 95 s from one bigWig per cell
Convert 257.6 million 10x fragments to 10,246 per-cell tracks 9 s at -p 8, 5.5 s at -p 32, 18 GB peak 140 s, 24 GB, a pandas recipe; 217 s, 8 GB, SnapATAC2 (fragments, not counts)

Measured on a 2x AMD EPYC 9575F with Python 3.13 and numba 0.68. The benchmarks page has the full setup.

More numbers and plots

Reading

167,750 windows of 1,000 bp centred on the fold-0 training peaks and negatives of ENCODE ATAC-seq experiment ENCSR123WME, read from the ATAC-seq bigWig ENCFF830RWF (238 MB) and the DNase-seq bigWig ENCFF989SAK (645 MB) on tmpfs, sorted by position, file opening included; median of three runs. The other libraries' 8-thread runs split the windows into 8 chunks on a thread pool. Every library returned the same values, once pyBigWig's NaN for bases without an interval was set to 0.

library ATAC, 1 thread ATAC, 8 threads DNase, 1 thread DNase, 8 threads
pybigtools 0.2.5, values() per window 1.25 s 1.42 s 1.98 s 2.21 s
pyBigWig 0.3.26, values() per window 3.95 s 4.08 s 4.50 s 4.64 s
pybbi 0.4.2, stackup 9.61 s 9.48 s 10.75 s 10.63 s
figwig 0.1.0 0.81 s 0.135 s 1.85 s 0.277 s

Writing

The plus-strand 5' ends of ENCODE ATAC-seq BAM ENCFF877LRY, as counts at 15,875,960 single bases, and the fold-change signal ENCFF932UQM, as 21,212,477 intervals, written from memory onto tmpfs at compression level 6, opening and closing included; median of three runs. pybigtools always writes zoom levels.

writer counts counts, zoom levels signal signal, zoom levels
pyBigWig 0.3.26 4.20 s 12.80 s 7.41 s 10.48 s
pybigtools 0.2.5 4.45 s 5.47 s
figwig, zlib, 1 thread 4.25 s 12.09 s 6.21 s 8.10 s
figwig, zlib, 8 threads 0.53 s 1.63 s 0.78 s 1.09 s
figwig, libdeflate, 1 thread 1.07 s 4.85 s 1.80 s 3.27 s
figwig, libdeflate, 8 threads 0.18 s 0.75 s 0.30 s 0.55 s

figwig's files were the size of pyBigWig's. For the counts, its peak memory was above pyBigWig's without zoom levels, 554 MB against 354 MB on one thread, and below it with them, 739 MB on 8 threads against 1,089 MB.

figwig bam2bw

ENCODE ATAC-seq BAM ENCFF877LRY (2.4 GB) at default flags, and 18.9 million scATAC-seq fragments with -f -u, from the page cache; figwig's median of five runs at -p 2 and one at -p 8, and one run of bam2bw, which reads one file in one process whatever -p is.

BAM fragments peak memory, BAM peak memory, fragments
bam2bw 0.5.1, -p 2 58.48 s 28.45 s 3,039 MB 2,866 MB
figwig bam2bw, -p 2 4.65 s 1.73 s 910 MB 1,061 MB
figwig bam2bw, -p 8 1.82 s 1.31 s 949 MB 1,180 MB

Both also ran on every one of 591 public BAM, SAM, BED and TSV test files that bam2bw takes, from 3 records to a 10 GB ATAC-seq BAM, under 5 to 9 sets of flags each, timed once, 16 at a time. In all 3,781 runs that bam2bw completed, figwig bam2bw gave the same exit code, messages and decoded entries. Where bam2bw took a second or more, figwig bam2bw was a median 3.2 times faster; where bam2bw took under half a second, figwig bam2bw took a median 0.12 s longer, mostly importing numba.

Wall time of figwig bam2bw against bam2bw 0.5.1, log scales

Peak and mean memory of figwig bam2bw against bam2bw 0.5.1, log scales

The memory runs were made before figwig bam2bw dropped the keys it had counted while still reading a BAM. On the 10 GB BAM at -p 1 its peak is now 2.9 GB, against bam2bw's 11.6 GB, and 4.9 GB with -u -f, against 20.2 GB.

figwig files

8,192 random windows of 2,048 bp over chr1 (2,048 windows at 512 tracks), read from N count tracks on tmpfs into float32, seconds at 1 / 8 / 32 threads; the tracks are the six count tracks of the examples and copies of them rolled by a seeded offset. Every reader returned the same bytes. The benchmarks page has writes, sizes, real ENCODE signal tracks, single cells, reads as tracks and the time to a CUDA tensor.

The format is specified byte by byte, with a Kaitai Struct definition and a reference reader, on the format page.

The format is built for count, pileup, single-cell and other sparse or integer tracks read many at a time, where it is a fifth of the bigWigs' size and reads faster. Dense float tracks, such as conservation or fold change, gain less in size, and a 2 kb window of them decodes a whole tile of every track in its group; three phyloP and phastCons tracks over chr22 still read 1.6 to 3.2 times faster than their bigWigs on 1 to 32 threads, and the converters page says when a smaller tile pays.

format 6 tracks 48 tracks 512 tracks
figwig file, position order 0.065 / 0.017 / 0.021 s 0.43 / 0.085 / 0.044 s 1.16 / 0.22 / 0.11 s
figwig file, track order 0.046 / 0.029 / 0.030 s 0.36 / 0.17 / 0.18 s 2.77 / 0.66 / 0.56 s
bigWigs, figwig 0.24 / 0.048 / 0.036 s 1.85 / 0.27 / 0.11 s 13.1 / 1.76 / 0.70 s
bigWigs, pybigtools 0.77 / 1.41 / 1.74 s 5.9 / 11.0 / 13.6 s 21 / 36 / 42 s
bigWigs, pyBigWig 2.23 / 2.56 / 2.75 s 17.5 / 20.0 / 21.7 s not run
zarr, zstd 1.95 / 2.25 / 2.61 s 8.1 / 4.0 / 3.7 s 46 / 15 / 12 s
HDF5, gzip 4 3.8 / 9.5 / 9.5 s 31 / 78 / 75 s 229 / 230 / 230 s

The 48 tracks are 148 MB as a figwig file, 1.3 GB as bigWigs figwig writes, 267 MB as zarr and 309 MB as HDF5; genome-wide, the six real tracks are 971 MB as ENCODE's bigWigs and 227 MB as a figwig file.

Threads

The same reads, writes and conversions at 1 to 32 threads, one run at a time, median of three. Reading gained up to 32 threads, writing with libdeflate up to 8 or 16, and figwig bam2bw up to about 8.

Wall time against threads for reading, writing and figwig bam2bw, log scales

Installation

pip install figwig

Or, with uv:

uv add figwig

It needs Python 3.10 or later, numpy 1.23 or later, and numba 0.58 or later. Optional extras:

  • figwig[fast] adds the deflate package, which writes bigWigs several times faster and reads figwig files about twice as fast.
  • figwig[isal] adds the isal package, whose ISA-L reads bigWigs 1.3 to 1.5 times faster, with the same values.
  • figwig[bam2bw] adds pysam, pyfaidx, biopython, tqdm, isal and deflate, which figwig bam2bw needs. figwig bam2bw runs on Linux and macOS, since pysam does not support Windows.
  • figwig[cuda] adds cupy (for CUDA 12; install cupy-cuda13x or cupy-cuda11x otherwise), with which FigWigReader.read(..., device="cuda") lands windows in GPU memory, for torch.from_dlpack and its JAX and TensorFlow equivalents.
pip install "figwig[fast]"
pip install "figwig[isal]"
pip install "figwig[bam2bw]"
pip install "figwig[cuda]"

Development install

git clone https://github.com/jmschrei/figwig.git
cd figwig
uv sync --extra dev
uv run pytest -n 8    # the tests on 8 cores; the first run compiles the kernels

Claude Code Skill

figwig ships a Claude Code skill that teaches a coding agent to use figwig in any project: reading windows, writing predictions, intervals and counts, converting reads with figwig bam2bw, figwig files and their converters, the rules a write follows, what each error means, and how much threads gain. Install it into ~/.claude/skills/figwig with:

figwig install-skill

After upgrading figwig, run figwig install-skill --force to replace the installed copy, which otherwise raises FileExistsError. -d DIRECTORY installs into another skills directory, and --symlink links to the copy inside the installed package instead of copying it.

Python API

The examples read two public ENCODE bigWigs: the ATAC-seq signal of HG02943 (ENCFF830RWF, 238 MB) and the DNase-seq signal of dorsolateral prefrontal cortex (ENCFF989SAK, 645 MB).

wget https://www.encodeproject.org/files/ENCFF830RWF/@@download/ENCFF830RWF.bigWig
wget https://www.encodeproject.org/files/ENCFF989SAK/@@download/ENCFF989SAK.bigWig

Reading windows

import numpy
from figwig import BigWigReader

bw = BigWigReader("ENCFF830RWF.bigWig")
chroms = numpy.array(["chr1", "chr1", "chr2"])
starts = numpy.array([1_000_000, 2_500_000, 300_000])
y = bw.read(chroms, starts, width=1000, n_jobs=8)
print(y.shape, y.dtype)
# (3, 1000) float32
print(y.sum(axis=1))
# [116.  43.  10.]

Window j covers [starts[j], starts[j] + width) on chroms[j], 0-based as in BED files, and is written into row j, whatever order the windows are given in. A base that no interval covers is missing, 0 unless given, and a base past the end of its chromosome is NaN. A BigWigReader reads the file's index once and keeps it, and can be pickled to DataLoader workers.

Reading several bigWigs into one array

import numpy
from figwig import read_bigwig

chroms = numpy.array(["chr1", "chr1", "chr2"])
starts = numpy.array([1_000_000, 2_500_000, 300_000])
y = read_bigwig(["ENCFF830RWF.bigWig", "ENCFF989SAK.bigWig"], chroms, starts,
	width=1000)
print(y.shape)
# (3, 2, 1000)

Channel i holds the values from the i-th file, in the (batch, channels, length) layout that sequence models take, and every file's work shares one pool of threads.

Writing windows

import numpy
from figwig import read_bigwig
from figwig import write_bigwig

chrom_sizes = {"chr1": 248_956_422, "chr2": 242_193_529}
chroms = numpy.array(["chr2", "chr1", "chr1"])
starts = numpy.array([300_000, 2_500_000, 1_000_000])
y_hat = numpy.random.default_rng(0).random((3, 2, 1000), dtype=numpy.float32)

write_bigwig(["plus.bw", "minus.bw"], chrom_sizes, chroms, starts, y_hat)
y = read_bigwig(["plus.bw", "minus.bw"], chroms, starts, width=1000)
print(numpy.array_equal(y, y_hat))
# True

write_bigwig takes what read_bigwig gives, such as a model's predictions for the same windows. A base whose value is missing, 0.0 unless given, or NaN is left out of the file, so that reading with the same missing gives the values back.

Writing a batch at a time

from figwig import BigWigReader
from figwig import BigWigWriter

chrom_sizes = {"chr1": 248_956_422, "chr2": 242_193_529}
with BigWigWriter("example.bw", chrom_sizes) as writer:
	writer.write("chr1", [1000, 2000], [0.5, 2.0], ends=[1500, 2100])
	writer.write("chr1", [5000, 5003, 5004], [[3], [1], [1]])
	writer.write("chr2", [10_000], [[0.0, 1.5, 1.5, 0.0, 2.5]])

y = BigWigReader("example.bw").read(["chr1", "chr2"], [5000, 10_000], width=6)
print(y)
# [[3.  0.  0.  1.  1.  0. ]
#  [0.  1.5 1.5 0.  2.5 0. ]]

BigWigWriter writes intervals (with ends), single bases (windows of width

  1. and windows as they come, so a track larger than memory can be written a batch at a time. Each call's items come after the last call's, in the order of chrom_sizes. The header, index and zoom levels are written when the writer is closed.

Reading and writing figwig files

import numpy
from figwig import FigWigReader
from figwig import FigWigWriter

chrom_sizes = {"chr1": 248_956_422, "chr2": 242_193_529}
tracks = [
    {"name": "atac.+", "assay": "ATAC-seq", "strand": "+"},
    {"name": "atac.-", "assay": "ATAC-seq", "strand": "-"},
]
counts = numpy.random.default_rng(0).poisson(0.1, (3, 2, 2114))
chroms = numpy.array(["chr2", "chr1", "chr1"])
starts = numpy.array([300_000, 2_500_000, 1_000_000])

with FigWigWriter("example.fw", chrom_sizes, tracks, genome="hg38") as fw:
    fw.write(chroms, starts, counts)

fw = FigWigReader("example.fw")
y = fw.read(chroms, starts, width=2114, tracks=fw.select(strand="+"))
print(y.shape, y.dtype, numpy.array_equal(y[:, 0], counts[:, 0]))
# (3, 1, 2114) uint8 True

A figwig file holds many tracks over one genome, each with attributes that select matches, and is read like a bigWig: many windows of one width in one call, from any subset of the tracks, into (windows, tracks, width) in the tracks' own dtype. The same read gives the nonzero bases as a sparse list with sparse=True, the tracks summed per label with labels=, or each window summed per bin with bins=; sums gives each track's total over intervals of any lengths, such as peaks or genes, or over thousands of cells at once; tile_sums gives each track's sum per tile, or per bin of a zoom level, without decoding a chunk; and device="cuda" lands any of these in GPU memory, where torch.from_dlpack takes them without a copy. A base no track wrote is 0, as is a bigWig's no-data unless bw2fw --missing nan keeps them apart.

A file is written in position order, tiles across tracks, for reading windows across many tracks, or in track order for reading a few tracks over the genome, and figwig reorder rewrites it in the other; write_sparse takes nonzero bases as entries, such as the cut sites of single cells. Each chunk of a group's tiles is stored densely, as nonzero bases, as runs of one value or as a matrix of the tracks present, whichever is smallest, in the narrowest type that holds it exactly: six genome-wide count tracks that take 971 MB as bigWigs take 227 MB as one figwig file. fw.tracks, fw.stats and fw.chrom_sizes say what a file holds, as figwig info does from the shell. The figwig files guide covers the format, its layout, the API, the converters and the first read's compile time: the first read of each kind in a new environment compiles its kernels, 15 to 30 s each with nothing printed meanwhile, and numba caches them beside the package, so on a cluster whose install is read-only set NUMBA_CACHE_DIR to a writable directory.

The documentation's reading and writing guides cover the rest: values and coordinates, chromosomes a file does not have, PyTorch data loaders, binned values, the files figwig refuses and falling back to another reader, compression, zoom levels, and files identical to pyBigWig's.

Command line

figwig bam2bw

figwig bam2bw my.bam -s hg38.chrom.sizes -n test-run -p 8                  # test-run.+.bw, test-run.-.bw
figwig bam2bw fragments.tsv.gz -s hg38.chrom.sizes -n test-run -f -u -p 8  # test-run.bw

figwig bam2bw is bam2bw 0.5.1 with the same arguments, output files and messages, except that -p is a number of cores, 1 unless given. By default it counts the 5' end of every mapped read at each base, and writes the counts of the two strands to two bigWigs; -u writes one, -f counts both ends of each fragment, -3p the 3' ends, -c every aligned base of each read (the coverage deepTools' bamCoverage gives), -ps, -ns, -sf and -r shift and scale the counts, and -z writes zoom levels. A BAM file is inflated by libdeflate and walked by a numba kernel on -p threads, and a BED/tsv file is scanned by a numba kernel; SAM files, --mate_pairs, remote files and malformed records go through bam2bw's own pysam or line loop, so that their result, or their error, is bam2bw's. The bigWigs hold bam2bw's entries but not its bytes, since figwig compresses them with libdeflate at level 1. figwig bam2bw -h lists the arguments, bam2bw's README has an example of each, and the documentation says how files are read and shared across cores.

All arguments, from figwig bam2bw -h
usage: figwig bam2bw [-h] -s SIZES [-u] [-f | -3p | -c] [-ps POS_SHIFT]
                     [-ns NEG_SHIFT] [-mp] [--rna5 {read1,read2}]
                     [--opposite_strand] [-sf SCALE_FACTOR] [-r] [-p PARALLEL]
                     -n NAME [-z ZOOMS] [-v]
                     filename [filename ...]

This tool will convert BAM files to bigwig files without an intermediate.

positional arguments:
  filename              The SAM/BAM or tsv/tsv.gz file to be processed.

options:
  -h, --help            show this help message and exit
  -s, --sizes SIZES     A chrom_sizes, .fai, or FASTA file. Only the first two
                        columns of a chrom_sizes/.fai file are read. A
                        compressed FASTA must be BGZF, not gzip.
  -u, --unstranded      Have only one, unstranded, output.
  -f, --fragments       The data is fragments and so both ends should be
                        recorded.
  -3p, --three_prime    Record the 3' end of each read instead of the 5' end.
  -c, --coverage        Count every aligned base of each read (its CIGAR's
                        M, = and X runs, so a deletion or a skipped intron
                        is not covered), or every base of each interval of
                        a BED/tsv file, instead of one end: the coverage
                        deepTools' bamCoverage gives at 1 bp bins. The
                        shifts move the first base (-ps) and the end (-ns);
                        a shift that empties a read leaves it uncovered.
                        Not with --mate_pairs.
  -ps, --pos_shift POS_SHIFT
                        A shift to apply to positive strand reads.
  -ns, --neg_shift NEG_SHIFT
                        A shift to apply to negative strand reads.
  -mp, --mate_pairs     Treat paired-end reads as a single RNA/fragment tag
                        instead of counting each mate independently: buffer
                        reads by name, and once both mates of a pair are seen,
                        record one jointly-determined position (see --rna5 and
                        --opposite_strand). BAM/SAM input only.
  --rna5 {read1,read2}  Which mate carries the 5' end of the RNA/fragment. The
                        other mate's own 5' end is used as the RNA's 3' end.
                        Only used with --mate_pairs.
  --opposite_strand     Report the strand of the mate opposite the one chosen
                        by --rna5, instead of that mate's own strand. Only
                        used with --mate_pairs.
  -sf, --scale_factor SCALE_FACTOR
                        A scaling factor to multiply each position by.
  -r, --read_depth      Whether to divide through by total (pre-scaled) read
                        depth.
  -p, --parallel PARALLEL
                        The number of cores to use, or a negative number to
                        count back from the number of CPUs: -1 is all of them.
  -n, --name NAME
  -z, --zooms ZOOMS     The number of zooms to store in the bigwig.
  -v, --verbose

figwig bw2fw, fw2bw, bam2fw, reorder, merge and info

figwig bw2fw -o k562.fw -m tracks.tsv -g hg38 -p 8 atac.+.bw atac.-.bw dnase.bw   # bigWigs -> k562.fw
figwig fw2bw k562.fw -o bigwigs -p 8 --select assay=ATAC-seq                      # tracks -> bigwigs/<name>.bw
figwig bam2fw a.bam b.bam -s hg38.chrom.sizes -n k562 -m files.tsv -p 8          # reads -> k562.fw
figwig bam2fw fragments.tsv.gz -s hg38.chrom.sizes -n pbmc --barcode-column 4 --cells cells.tsv -f -p 8   # a track per cell
figwig reorder k562.fw by_track.fw --order tracks                                # the other order
figwig merge -o all.fw k562.fw hepg2.fw                                          # files -> one file
figwig info k562.fw                                                              # what the file holds

figwig bw2fw writes bigWigs as the tracks of one figwig file, named after their files or by a metadata table whose other columns become the tracks' attributes, or adds them to a file with --append; figwig fw2bw writes a figwig file's tracks back as bigWigs with such a table; figwig bam2fw takes figwig bam2bw's arguments and writes each input file as its own pair of stranded tracks, or a fragments file as one track per cell with --barcode-column; figwig reorder rewrites a file in the other order; figwig merge copies several files' tracks into one; and figwig info prints a file's tracks, attributes and statistics. The converters page has the options.

figwig install-skill

Installs the Claude Code skill; see above.

Origin

figwig began as the bigWig reader inside tangermeme's extract_loci (PR #107). It was checked there against pybigtools, bit for bit, on the training and validation sets of 92 Cherimoya models and on 3,131 synthetic cases.

Metadata

Release files for figwig 0.2.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for figwig 0.2.0
File Size Uploaded Provenance Actions
figwig-0.2.0.tar.gz 346.2 kB Without provenance Details for figwig-0.2.0.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for figwig 0.2.0
File Interpreter ABI Platform Actions
figwig-0.2.0-py3-none-any.whl Python 3 none any Details for figwig-0.2.0-py3-none-any.whl

Total release size: 568.1 kB

Release files / figwig-0.2.0.tar.gz

Download URL figwig-0.2.0.tar.gz
Size 346.2 kB
Tags Source
SHA-256 checksum
How to use checksums
e2708fc6a6410f9338f3df6ee857f603bad238226dcd7d45903c31a8a87b7c2e
BLAKE2b-256 checksum
How to use checksums
ec8a3d1e23a83604b6190d4d854093681975cde3d1853918cb0ffe43e3a8ebc1
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.13.5

Release files / figwig-0.2.0-py3-none-any.whl

Download URL figwig-0.2.0-py3-none-any.whl
Size 221.9 kB
Tags Python 3
SHA-256 checksum
How to use checksums
aca77f12a82101336d12a5ec44a7a0e1bbad1bcd307966d6dc3600c6bc86a0af
BLAKE2b-256 checksum
How to use checksums
6c1156155bddf2653388c6f5694aadcee776f867e1a328f98d7ca5fe7ac7609f
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.13.5

Security for figwig 0.2.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page