Parse words to numbers (thirteen => 13) about parsedatetime HOT 3 CLOSED

always willing to consider new code - can you work up a pull request that implements it along with tests?

from parsedatetime.

This is supported for English language numbers up to and including twenty. After that it would get tricky since numbers may be expressed in multiple words. For larger numbers, it looks like parsedatetime has the text2num logic incorporated in the _convertUnitAsWords function, but that function is not actually used anywhere.

It would probably be possibly to parse those multi-word numbers but likely not worth the effort. Except in limited circumstances I would imagine that spelling out numbers higher than twenty is very rare.

from parsedatetime.

I'm sorry, I realize that my problem lies with an older version of parsedatetime that ships with Ubuntu 14.04. Using the latest from pypi does correctly parse words, like "three months";

Old version in Ubuntu 14.04:

>>> pdt.parse("3 months")
(time.struct_time(tm_year=2015, tm_mon=8, tm_mday=25, tm_hour=8, tm_min=3, tm_sec=37, tm_wday=1, tm_yday=237, tm_isdst=-1), 1)
>>> pdt.parse("three months")
(time.struct_time(tm_year=2015, tm_mon=6, tm_mday=4, tm_hour=8, tm_min=3, tm_sec=38, tm_wday=3, tm_yday=155, tm_isdst=-1), 1)
>>> 

Latest pypi:

>>> pdt.parse("3 months")
(time.struct_time(tm_year=2015, tm_mon=8, tm_mday=25, tm_hour=8, tm_min=2, tm_sec=29, tm_wday=1, tm_yday=237, tm_isdst=-1), 1)
>>> pdt.parse("three months")
(time.struct_time(tm_year=2015, tm_mon=8, tm_mday=25, tm_hour=8, tm_min=4, tm_sec=8, tm_wday=1, tm_yday=237, tm_isdst=-1), 1)
>>> 

from parsedatetime.