timingdiagram

Work with discrete state changes over time.

• Reduce data to state changes
• Compare diagrams over time
• Query state by time
• Use any ordered index

Install

From PyPI with pip:

pip install timingdiagram-alkasm

From source:

git clone https://github.com/alkasm/timingdiagram
cd timingdiagram
pip install .

Note that you must have pip >= 19.0 installed in your environment to install from source, since this project uses the pyproject.toml file defined in PEP-0517 instead of setup.py. More from PyPA here.

Try it out

>>> from timingdiagram import TimingDiagram
>>> d1 = TimingDiagram(enumerate([False, False, False, True, True, False, True]))
>>> d2 = TimingDiagram(enumerate([False, True, False, False, True, False, False]))
>>> d1 | d2
TimingDiagram([(0, False), (1, True), (2, False), (3, True), (5, False), (6, True)])

Example

Suppose you had a log of users signing in and out of a service, and the log included the time, user id, and action the user took. We can view each user's login/logout history as a timing diagram, and simply & them all together to see when all users were logged in at the same time:

log = """2019-08-27T19:38:50 001768bf-af44-46a6-890d-048f2c50aa29 login
2019-08-27T19:51:11 084c07f0-dd0d-46a3-8eb5-1d4cb13756a4 logout
2019-08-27T19:55:25 001768bf-af44-46a6-890d-048f2c50aa29 logout
2019-08-27T19:58:37 001768bf-af44-46a6-890d-048f2c50aa29 login
2019-08-27T20:17:21 a8118353-eb81-4ce0-8d10-6f3f9de6d7ca login
2019-08-27T20:45:19 001768bf-af44-46a6-890d-048f2c50aa29 logout
2019-08-27T21:01:45 001768bf-af44-46a6-890d-048f2c50aa29 login
2019-08-27T21:18:09 001768bf-af44-46a6-890d-048f2c50aa29 logout
2019-08-27T22:02:37 084c07f0-dd0d-46a3-8eb5-1d4cb13756a4 login
2019-08-27T22:55:54 001768bf-af44-46a6-890d-048f2c50aa29 login
2019-08-27T23:08:07 001768bf-af44-46a6-890d-048f2c50aa29 logout
2019-08-27T23:23:04 a8118353-eb81-4ce0-8d10-6f3f9de6d7ca logout
2019-08-27T23:47:50 001768bf-af44-46a6-890d-048f2c50aa29 login
2019-08-27T23:55:10 084c07f0-dd0d-46a3-8eb5-1d4cb13756a4 logout
2019-08-27T23:56:33 001768bf-af44-46a6-890d-048f2c50aa29 logout""".split("\n")


from collections import defaultdict
from functools import reduce
from timingdiagram import TimingDiagram


sessions = defaultdict(list)
for row in log:
    ts, userid, action = row.split()
    sessions[userid].append((ts, action == "login"))

all_logged_in = reduce(lambda d1, d2: d1 & d2, map(TimingDiagram, sessions.values()))

From just a few lines of code, we get a timing diagram corresponding to when all the users were logged in:

TimingDiagram([
  ('2019-08-27T19:38:50', False), 
  ('2019-08-27T22:55:54', True), 
  ('2019-08-27T23:08:07', False), 
  ('2019-08-27T23:56:33', False)
 ])

So all users were logged in between 22:55:54 and 23:08:07 on 2019-08-27. The additional states at the beginning and end signify the start and end times of the logs.