Timestamp library and utility
tslib is a timestamp processing toolkit and library, written in Python. It is very helpful when dealing with the millisecond-precision timestamps that we encounter all the time in computer systems, and to work with human-readable time deltas away from “now”.
Installation
You can install tslib directly from PyPI with the following command:
$ pip install tslib
If you’re installing from a clone of this repository, use:
$ git clone https://github.com/mpetazzoni/tslib $ cd tslib/ $ pip install -e .
tslib requires Python 3.11 or newer. Named timezones use Python’s zoneinfo module and the system IANA timezone database, with the automatically installed tzdata package as a fallback. Neither pytz nor six is required.
You’re now ready to use ts, which should be directly available in your $PATH.
Usage
$ ts -h
usage: ts [-h] [-t TZ] [-n] [-i] [-a {minute,hour,day}]
Human readable timestamps
options:
-h, --help show this help message and exit
-t, --timezone TZ Use IANA timezone TZ (default: OS local timezone)
-n, --timestamp-only Only output the resulting millisecond timestamp(s)
-i, --inline Apply timestamp replacements inline of the incoming
text
-a, --align {minute,hour,day}
Align the timestamp
ts transforms milliseconds timestamps and human-readable deltas into fully qualified date and time printouts, containing:
the UTC timestamp (since Epoch)
the UTC time represented by the timestamp
the local time represented by the timestamp
the human-readable delta of this timestamp against current time
The output is tab-delimited and can easily be re-segmented by piping into column -t -s "\t".
Arguments can be passed as command-line arguments, or piped into ts via STDIN. If no input comes from either of these sources, ts will simply output the current time:
$ ts -t America/Los_Angeles 1423599084206 2015-02-10 20:11:24.206 UTC+0000 2015-02-10 12:11:24.206 PST-0800 0
Inline mode
ts can operate in inline-replacement mode, for timestamps only. In this mode, you can pipe in any text and ts will output it back to you with all timestamps replaced with a human-readable date and time and human-readable delta representation.
Given the inherent difficulty of accurately matching timestamps, the matching is limited to 13-digits millisecond precision timestamps (that may be immediately followed by a L). Numbers smaller, or larger than 13 digits are not matched, and are left alone; meaning no 13-digit section of them is matched and replaced.
$ echo B_b-3PAAYAA B_cJXlvAYAA B_cJXl6AcAA \
| xargs sfc mb g -p TSVH -f sf_checkpointTimestampMs -f sf_updatedOnMs \
| ts -i -t America/Los_Angeles \
| column -t -s " "
sf_id sf_checkpointTimestampMs sf_updatedOnMs
B_b-3PAAYAA 2015-03-06 13:38:58.851 PST-0800 (-2w5h18m6s210) 2015-03-06 13:38:58.852 PST-0800 (-2w5h18m6s209)
B_cJXlvAYAA 2015-03-08 09:04:36.218 PDT-0700 (-1w5d10h52m28s843) 2015-03-08 09:04:36.218 PDT-0700 (-1w5d10h52m28s844)
B_cJXl6AcAA 2015-03-08 09:04:34.252 PDT-0700 (-1w5d10h52m30s810) 2015-03-08 09:04:34.342 PDT-0700 (-1w5d10h52m30s720)
More examples
Piping:
$ cat << EOF | ts -t America/Los_Angeles pipe heredoc> 1404424797009L pipe heredoc> 1415917836779L pipe heredoc> EOF 1404424797009 2014-07-03 21:59:57.009 UTC+0000 2014-07-03 14:59:57.009 PDT-0700 -31w5d1h14m26s38 1415917836779 2014-11-13 22:30:36.779 UTC+0000 2014-11-13 14:30:36.779 PST-0800 -12w5d43m46s268
Using the column output (assuming the same input in a /tmp/ts.txt file):
$ cat /tmp/ts.txt | ts | cut -f4 -31w5d1h16m17s703 -12w5d45m37s933
Sorting on a column (here, by descending timestamps in first column):
$ cat /tmp/ts.txt | ts -t America/Los_Angeles | sort -k1 -t $'\t' -r 1415917836779 2014-11-13 22:30:36.779 UTC+0000 2014-11-13 14:30:36.779 PST-0800 -12w5d46m6s924 1404424797009 2014-07-03 21:59:57.009 UTC+0000 2014-07-03 14:59:57.009 PDT-0700 -31w5d1h16m46s694
Deltas
Human-readable deltas can be expressed in weeks (w), days (d), hours (h), minutes (m), and seconds (s). The remainder, without a unit, is assumed to be milliseconds. Any “segment” can be omitted, the only requirement is that the segments that are specified are written in descending order of span (days before hours, hours before minutes, etc.).
Here’s an example: -12w4d6m57s257. Note that hours are missing, which simply means 12 weeks, 4 days, 6 minutes, 57 seconds and 257 milliseconds.
As you might have guessed, deltas can be both negative and positive. For positive deltas, the leading + may be omitted if units are used, otherwise the number is assumed to be an absolute timestamp:
$ ts -t America/Los_Angeles -1
1423599850752 2015-02-10 20:24:10.752 UTC+0000 2015-02-10 12:24:10.752 PST-0800 -1
$ ts -t America/Los_Angeles 0 1
0 1970-01-01 00:00:00.000 UTC+0000 1969-12-31 16:00:00.000 PST-0800 -2353w5d20h24m14s145
1 1970-01-01 00:00:00.001 UTC+0000 1969-12-31 16:00:00.001 PST-0800 -2353w5d20h24m14s144
$ ts -t America/Los_Angeles +1
1423599855941 2015-02-10 20:24:15.941 UTC+0000 2015-02-10 12:24:15.941 PST-0800 1
Using a different local timezone
The third column shows the timestamp’s representation in local time. By default it uses the operating system’s configured local timezone, including the daylight-saving offset appropriate to the timestamp being converted. It does not depend on the legacy US/Pacific timezone alias. On POSIX systems, the TZ environment variable can override the local timezone.
Use -t to select an explicit IANA timezone name, such as UTC, America/Los_Angeles, or Europe/Paris. Invalid names produce a command-line usage error. The Pacific-time examples above explicitly select America/Los_Angeles so their output does not depend on the host timezone:
$ ts -t Europe/Paris 1423600015955 2015-02-10 20:26:55.955 UTC+0000 2015-02-10 21:26:55.955 CET+0100 0
Absolute, human-readable offsets from Epoch
By prefixing a human-readable delta with an equal sign (=), you obtain an absolute offset from the Epoch. The side-effect of this is that it allows for converting a human-readable delta into its corresponding millisecond duration.
$ ts -n '=1h'
3600000
$ ts -t America/Los_Angeles '=1d'
86400000 1970-01-02 00:00:00.000 UTC+0000 1970-01-01 16:00:00.000 PST-0800 -2365w6d21h56m20s98
Alignment
-a minute, -a hour, and -a day round forward to the next UTC boundary, even if the timestamp is already aligned. The chosen display timezone does not change the alignment:
$ ts -n -a hour 1
3600000
Library compatibility
UTC datetime helpers now return standard-library datetime objects with datetime.timezone.utc rather than pytz.utc. date_from_utc still requires a naive datetime and rejects an already timezone-aware value. Replace pytz-specific operations such as localize and normalize with standard-library datetime operations when upgrading.
render_date(date) keeps its UTC display default for library callers. Pass None to use the OS-local timezone, or a zoneinfo.ZoneInfo object to select a named timezone:
from zoneinfo import ZoneInfo
from tslib import date_from_utc_ts, render_date
print(render_date(date_from_utc_ts(1404424797009), ZoneInfo("Europe/Paris")))
Development
Install the development dependencies in a virtual environment, then run:
$ python -m pip install -e '.[dev]' $ python -m pytest $ python -m ruff check . $ python -m build $ python -m twine check dist/*
CI runs these checks on Python 3.11 through 3.14 on Linux, macOS, and Windows.
Metadata
Release files for tslib 2.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| tslib-2.0.tar.gz | 16.5 kB | With provenance | Details for tslib-2.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| tslib-2.0-py3-none-any.whl | Python 3 | none | any | Details for tslib-2.0-py3-none-any.whl |
Total release size: 28.4 kB
Release files / tslib-2.0.tar.gz
| Download URL | tslib-2.0.tar.gz |
|---|---|
| Size | 16.5 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
b6c269ccc8a78d1292b38585044144753350e8af6685a0ad21ec37cd248ad8eb
|
|
BLAKE2b-256 checksum How to use checksums |
2f69da2e7bd5c8e8597fcf146792d43cfe4f2cae9fccc9f9784a9f80a2d362f1
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
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.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Nov 15, 2026.
Transparency logRelease files / tslib-2.0-py3-none-any.whl
| Download URL | tslib-2.0-py3-none-any.whl |
|---|---|
| Size | 11.9 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
6c207b702d0b42b5204bc60f149eb3ff554bd04ea777d4c8394e064ed4aca07e
|
|
BLAKE2b-256 checksum How to use checksums |
89e8ae6dcbacba692169dd901fd8173491aab2525615724328217b488283ea34
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
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.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Nov 15, 2026.
Transparency logSecurity for tslib 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.
Full provenance
Both files in this release have been uploaded with provenance metadata.
Attestation for tslib-2.0.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Nov 15, 2026.
Transparency logA 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.