To avoid certain spelled-out words, alternate representations are used. For example:

15: 'ט״ו'
16: 'ט״ז'
#...
270: 'ע״ר'
272: 'ער״ב'
298: 'רח״צ'
#...

If you add in a YAML file with these alternate representations, you could first do a lookup on the YAML. If it's not in the list of specials, you can generate the letters as before.

reproduction steps

$ python --version
Python 2.7.12+  -- (it reproduces for Python 3.5.3 as well)
$ pip install --upgrade https://github.com/OriHoch/python-hebrew-numbers/archive/master.zip
$ python
>>> from hebrew_numbers import int_to_gematria

expected

• should work

actual

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/ori/tmp/venv/local/lib/python2.7/site-packages/hebrew_numbers/__init__.py", line 8, in <module>
    specialnumbers = yaml.safe_load(io.open(DATA_PATH, encoding='utf8'))
IOError: [Errno 2] No such file or directory: '/home/ori/tmp/venv/local/lib/python2.7/site-packages/hebrew-special-numbers/styles/default.yml'

Hi,

nice but few things:

1. you don't support the ה' in the beginning
2. we map instead of dict?

Tal

Hey!
The package does not support numbers over 1,000?
Can't convert numeric year to year in letters?

>>> from hebrew_numbers import int_to_gematria
>>> int_to_gematria(298)
רח"צ

implementation notes

• should use hebrew-special-numbers to implement it

TODO

• currently we have travis run multiple python versions, need to determine if the versions it's running are ok (I think we need to add python 3.6)
• should update to use tox
• should remove the tests from the published package
• should run the tests using standard test runner (e.g. pytest or python -m unittest)
• should have automatic publishing to pypi when github release is published. something like this - https://github.com/hasadna/knesset-data-python/blob/master/.travis/after_success.sh

it's a bit more complicated but not much..

see #1 for a list of special cases