Semantic is a Python library for extracting semantic information from text, such as dates and numbers. Full documentation is available on PyPI, with a list of primary features and uses-cases below.
Installing semantic is simple:
$ pip install semantic
Semantic consists of four main modules, each of which corresponds to a different semantic extractor. The test suite (test.py) contains tons of examples for each of the four modules, but some sample use-cases are described below.
Useful for:
- Extracting relative (e.g., "a week from today") and absolute (e.g., "December 11, 2013") dates from text snippets.
- Converting date objects to human-ready phrasing.
Useful for:
- Extracting numbers (integers or floats) from text snippets.
- Converting numbers to human-readable strings.
Example usage:
#!/usr/bin/env python
from semantic.numbers import NumberService
service = NumberService()
print service.parse("Two hundred and six")
# 206
print service.parse("Five point one five")
# 5.15
print service.parse("Eleven and two thirds")
# 11.666666666666666
print service.parseMagnitude("7e-05")
# "seven to the negative five"
Useful for performing mathematical operations expressed as words.
Example usage:
#!/usr/bin/env python
from semantic.solver import MathService
service = MathService()
print service.parseEquation("Log one hundred and ten")
# 4.70048
Useful for converting between units expressed as words.
Example usage:
#!/usr/bin/env python
from semantic.solver import ConversionService
service = ConversionService()
print service.convert("Seven and a half kilograms to pounds")
# (16.534, 'lbs')
print service.convert("Seven and a half pounds per square foot to kilograms per meter squared")
# (36.618, 'kg/m**2')
The Dates, Numbers, and Math modules can run in isolation (i.e., without any dependencies), while the Units module requires quantities and Numpy.
MIT © Charles Marsh