quickadd is a natural language date & time parser written in python. It builds on top of ctparse and is an actively maintained fork.
With pip install -e git+https://github.com/Acreom/quickadd.git#egg=quickadd
or run python setup.py install in the root directory after forking.
Recurring events
r = ctparse("beer daily 4pm")
r.resolution
Recurring[5-14]{daily 1 2021-05-09 16:00 (X/X) 2021-05-09 16:00 (X/X)}
r = ctparse("beer every thursday 4")
r.resolution
Recurring[5-21]{weekly 1 2021-04-15 16:00 (X/X) 2021-04-15 16:00 (X/X)}
r = ctparse("beer every friday 9-5")
r.resolution
Recurring[5-21]{weekly 1 2021-05-14 09:00 (X/X) 2021-05-14 17:00 (X/X)}
r = ctparse("beer september 24 / beer every 24.9")
r.resolution
Recurring[5-21]{YEARLY 1 2021-09-24 (X/X) 2021-09-24 (X/X)}
r = ctparse("beer thursdays 3pm and wednesdays 4pm")
r.resolution
RecurringArray[5-37]{
Recurring instance: weekly 1 2021-05-13 15:00 (X/X) 2021-05-13 15:00 (X/X)
Recurring instance: weekly 1 2021-05-12 16:00 (X/X) 2021-05-12 16:00 (X/X)
}
r = ctparse("beer 9pm weekdays")
r.resolution
RecurringArray[5-17]{
Recurring instance: weekly 1 2021-05-10 21:00 (X/X) 2021-05-10 21:00 (X/X)
Recurring instance: weekly 1 2021-05-11 21:00 (X/X) 2021-05-11 21:00 (X/X)
Recurring instance: weekly 1 2021-05-12 21:00 (X/X) 2021-05-12 21:00 (X/X)
Recurring instance: weekly 1 2021-05-13 21:00 (X/X) 2021-05-13 21:00 (X/X)
Recurring instance: weekly 1 2021-05-14 21:00 (X/X) 2021-05-14 21:00 (X/X)}More rules
ruleNextFrequency
#reference date = Dec 13th 2022
r = ctparse("code next week 4pm")
r.resolution
Time[5-18]{2022-12-20 16:00
r = ctparse("code next month")
r.resolution
Time[5-15]{2023-01-13 X:X (X/X)}ruleLastDOM
#reference date = Dec 13th 2022
r = ctparse("code last monday of the month")
r.resolution
Time[5-17]{2022-12-26 X:X (X/X)}rrule support
r.resolution.to_rrule()
Out[4]: 'RRULE:FREQ=DAILY;COUNT=1'Subject extraction
r = ctparse("beers and burgers friday 8pm-9pm")
r.subject
Out[2]: 'beers and burgers'PM bias
r = ctparse("fix the issue tmrw 2")
r.resolution
Time[14-20]{2022-11-23 14:00 (X/X)}
r = ctparse("fix the issue tmrw 2", pm_bias=False)
r.resolution
Time[14-20]{2022-11-23 02:00 (X/X)}Rules for ambigious natural language expressions
r = ctparse("code 9-5")
r.resolution
Interval[0-0]{2022-11-23 09:00 (X/X) - 2022-11-23 17:00 (X/X)}US/EU date format
r = ctparse("fix the issue 5.3")
r.resolution
Time[14-17]{2023-03-05 X:X (X/X)}
r = ctparse("fix the issue 5.3", date_format="US")
r.resolution
Time[14-17]{2023-05-03 X:X (X/X)}Rule combinations
r = ctparse("beer in 3 days 4pm")
r.resolution
Time[5-18]{2021-05-12 16:00 (X/X)}
r = ctparse("beer in 3 days 4pm every week")
r.resolution
Recurring[5-29]{weekly 1 2021-05-12 16:00 (X/X) 2021-05-12 16:00 (X/X)}
r = ctparse("beer every friday 4-6:30pm")
r.resolution
Recurring[5-26]{WEEKLY 1 2022-11-25 16:00 (X/X) 2022-11-25 18:30 (X/X)}+ performance improvements
"beer thursday 4"
Time[5-15]{2021-05-13 16:00 (X/X)}"beer 4-6"
Interval[0-0]{2021-05-09 16:00 (X/X) - 2021-05-09 18:00 (X/X)}"beer in 4 hours"
Duration[5-15]{4 hours}The package ctparse is a pure python package to parse time
expressions from natural language (i.e. strings). In many ways it builds
on similar concepts as Facebook’s duckling package
(https://github.com/facebook/duckling). However, for the time being it
only targets times and only German and English text.
In principle ctparse can be used to detect time expressions in a
text, however its main use case is the semantic interpretation of such
expressions. Detecting time expressions in the first place can - to our
experience - be done more efficiently (and precisely) using e.g. CRFs or
other models targeted at this specific task.
ctparse is designed with the use case in mind where interpretation
of time expressions is done under the following assumptions:
- All expressions are relative to some pre-defined reference times
- Unless explicitly specified in the time expression, valid resolutions
are in the future relative to the reference time (i.e.
12.5.will be the next 12th of May, but12.5.2012should correctly resolve to the 12th of May 2012). - If in doubt, resolutions in the near future are more likely than resolutions in the far future (not implemented yet, but any resolution more than i.e. 3 month in the future is extremely unlikely).
The specific comtravo use-case is resolving time expressions in booking requests which almost always refer to some point in time within the next 4-8 weeks.
ctparse currently is language agnostic and supports German and
English expressions. This might get an extension in the future. The main
reason is that in real world communication more often than not people
write in one language (their business language) but use constructs to
express times that are based on their mother tongue and/or what they
believe to be the way to express dates in the target language. This
leads to text in German with English time expressions and vice-versa.
Using a language detection upfront on the complete original text is for
obvious no solution - rather it would make the problem worse.
from ctparse import ctparse
from datetime import datetime
# Set reference time
ts = datetime(2018, 3, 12, 14, 30)
ctparse('May 5th 2:30 in the afternoon', ts=ts)This should return a Time object represented as
Time[0-29]{2018-05-05 14:30 (X/X)}, indicating that characters
0-29 were used in the resolution, that the resolved date time is the
5th of May 2018 at 14:30 and that this resolution is neither based on a
day of week (first X) nor a part of day (second X).
Normally, ctparse will anchor time expressions to the reference time.
For example, when parsing the time expression 8:00 pm, ctparse will
resolve the expression to 8 pm after the reference time as follows
parse = ctparse("8:00 pm", ts=datetime(2020, 1, 1, 7, 0), latent_time=True) # default
# parse.resolution -> Time(2020, 1, 1, 20, 00)This behavior can be customized using the option latent_time=False, which will
return a time resolution not anchored to a particular date
parse = ctparse("8:00 pm", ts=datetime(2020, 1, 1, 7, 0), latent_time=False)
# parse.resolution -> Time(None, None, None, 20, 00)ctparse - as duckling - is a mixture of a rule and regular
expression based system + some probabilistic modeling. In this sense it
resembles a PCFG.
At the core ctparse is a collection of production rules over
sequences of regular expressions and (intermediate) productions.
Productions are either of type Time, Interval, Duration or Recurring and can
have certain predicates (e.g. whether a Time is a part of day like
'afternoon').
A typical rule than looks like this:
@rule(predicate('isDate'), dimension(Interval))I.e. this rule is applicable when the intermediate production resulted
in something that has a date, followed by something that is in interval
(like e.g. in 'May 5th 9-10').
The actual production is a python function with the following signature:
@rule(predicate('isDate'), dimension(Interval))
def ruleDateInterval(ts, d, i):
"""
param ts: datetime - the current refenrence time
d: Time - a time that contains at least a full date
i: Interval - some Interval
"""
if not (i.t_from.isTOD and i.t_to.isTOD):
return None
return Interval(
t_from=Time(year=d.year, month=d.month, day=d.day,
hour=i.t_from.hour, minute=i.t_from.minute),
t_to=Time(year=d.year, month=d.month, day=d.day,
hour=i.t_to.hour, minute=i.t_to.minute))This production will return a new interval at the date of
predicate('isDate') spanning the time coded in
dimension(Interval). If the latter does code for something else than
a time of day (TOD), no production is returned, e.g. the rule matched
but failed.
Some observations on the problem:
- Each rule is a combination of regular expressions and productions.
- Consequently, each production must originate in a sequence of regular expressions that must have matched (parts of) the text.
- Hence, only subsequence of all regular expressions in all rules can lead to a successful production.
To this end the algorithm proceeds as follows:
- Input a string and a reference time
- Find all matches of all regular expressions from all rules in the input strings. Each regular expression is assigned an identifier.
- Find all distinct sequences of these matches where two matches do not overlap nor have a gap inbetween
- To each such subsequence apply all rules at all possible positions until no further rules can be applied - in which case one solution is produced
Obviously, not all sequences of matching expressions and not all sequences of rules applied on top lead to meaningful results. Here the PCFG kicks in:
- Based on example data (
corpus.py) a model is calibrated to predict how likely a production is to lead to a/the correct result. Instead of doing a breadth first search, the most promising productions are applied first. - Resolutions are produced until there are no more resolutions or a timeout is hit.
- Based on the same model from all resolutions the highest scoring is returned.
This package was created with Cookiecutter and the audreyr/cookiecutter-pypackage project template.