This project is temporary archived.
Convert number words from different languages with Python, CSharp or Java API eg. three hundred and forty two to numbers (342) or vingt-et-un (21) or две целых три десятых (2.3).Word2Number-i18n convert number words (eg. twenty one) to numeric digits (21). Below is the installation, usage and other details of this module.
- French: We can calculate! So 4*20+10 are 90 and this is normal (quatre-vingt-dix).
- German: We have our own rules. All under million is one word (but often false written) and also we say for 21 onetwenty (einunzwanzig) not twenty one.
- Spanish: Less is more, we need no word for 1.000.000.000 as word an says mil millón (for example tres mil millones).
- English
- French
- Portugues
- Russian
- Slovak
- Spanish
- Persian (just python implementation)
Do follow steps
1. check your request (directory data), textfile name ISO-639-1 code
2. if not found check ISO-639-3 code
3. if not found create new file with new ISO-639-1/3 code
You want to tranfer NLP CARD to numeric value for Lower Sorbian. German (de) isn't it.
You do not found an ISO-639-1 code, you do not found an ISO-639-1 file for dsb
extension.
You create a new file number_system_dsb.txt
with utf-8 encoding
null=0
jaden=1
dwa=2
tśi=3
styri=4
pěś=5
šesć=6
sedym=7
wósym=8
źewjeś=9
źaseś=10
Every config file need the values from zero to nine and every supported language need a config file.
And add all single words with values. In extensions you need also value for point and additional replacement words and measure words in the same configuration file. Its different to value names where the key is localized now the key is internal used and parts or full predefined. Feel free to copy a existing language config.
Example:
# 0-9
# much more
twenty=20
thousand=1000
# special
point=komma
replace:what=with
replace:две=два
replace:dozen=twelfe
replace:gran=thousand
measure:aLotOf=1000
measure:namesAreLikeIceOnMadeira=1000000
The point is elementary to work with decimal values and this value is the word in
your language. The measure are the multiplier for the numbers. It extends the numeric
value with an label or like NLP with an tag. So the name thousand
is the name for
the numeric value and a gran
are a synonym for thousand
and thousand
is
a measure word.
So giving twenty nice gran
do the follow:
word2number-i18n
register initial value1000
as measure. The nameaLotOf
is your comment.word2number-i18n
looking for replace and change text totwenty nice thousand
word2number-i18n
filter none number words to trash and your text istwenty thousand
word2number-i18n
find out1000
is a measure and the localized name isthousand
. In result of thisthousand
is here a multiplier for the numeric value before.word2number-i18n
take the numeric value20
instead oftwenty
and the multiplier after to give your the result20000
Internal much more especially checks inside working.
Please note that these Python implementation is similar to Java and CSharp.
Please ensure that you have updated pip to the latest version before installing word2number-i18n.
You can install the module using Python Package Index using the below command.
pip3 install word2number-i18n
On macOS
# git clone https://github.com/bastie/w2ni18n.git w2n
# python3 setup.py install
Make sure you install all requirements given in requirements.txt
pip3 install -r requirements.txt
Add the word2number-i18n to requirements.txt in your project. First you have to import the module using the below code.
from word2numberi18n import w2n
Then you can use the word_to_num method to convert a number-word to numeric digits, as shown below.
print(w2n.word_to_num("two million three thousand nine hundred and eighty four"))
2003984
print(w2n.word_to_num('two point three'))
2.3
print(w2n.word_to_num('one hundred thirty-five'))
135
print(w2n.word_to_num('million million'))
Error: Redundant number! Please enter a valid number word (eg. two million twenty three thousand and forty nine)
None
print(w2n.word_to_num('blah'))
Error: No valid number words found! Please enter a valid number word (eg. two million twenty three thousand and forty nine)
None
word2number looking for your specific language with
1. given parameter for language and if not found
2. defined environment variable w2n.lang with ISO lang code like en, hi, de and if not found
3. over locale.getdefaultlocale() and if not found
4. over environment variable "LANGUAGE" and if not found
5. fallback to english
In result use it classic Python like (example unit_testing_ru.py):
os.environ['w2n.lang'] = 'ru'
w2n.word_to_num('две целых три десятых')
Also work with object oriented Python like (example unit_testing_fr.py):
instance = w2n.W2N(lang_param="fr")
instance.word_to_num('trente-et-un')
Place in the data directory your language specific dictionary file with ISO lang code in the name.
# python3 -m reuse lint
# python3 -m flake8 | grep -v ":80: E501"
#
# python3 setup.py sdist bdist_wheel
# python3 -m twine check dist/*
# python3 -m twine upload dist/*
Make sure you install all requirements given in development.txt
pip3 install -r development.txt
Please note that these Java implementation is similar to Python and CSharp.
Download the latest version of word2number from GitHub.
On macOS
# git clone https://github.com/bastie/w2ni18n.git w2n
# ./w2n/java/src/build.sh
word2number looking for your specific language with
1. construct your object instance with ISO lang code as parameter like fr, de and if not
2. defined environment variable (not property) w2n.lang with ISO lang code like en, hi, de and if not found
3. over java.util.Locale.getDefault() and if null
4. over environment variable "LANGUAGE" and if not found
5. fallback to english
Place in the data directory your language specific dictionary file with ISO lang code in the name.
Add the word2number-i18n to module-info.java in your project.
requires word2number;
Then import the class using the below code.
import word2number.W2N;
Then you can use the wordToNum method to convert a number-word to numeric digits, as shown below.
Locale.setDefault(Locale.CANADA);
W2N english = new W2N();
System.out.println(english.wordToNum("three hundred fifty"));
Locale.setDefault(Locale.CANADA_FRENCH);
W2N french = new W2N();
System.out.println(french.wordToNum("vingt et un"));
System.out.println(english.wordToNum("three point one four"));
In result it prints
350
21
3.14
Call the build.sh
script and use the new w2ni18n-VERSION-.jar
file
Please note that these CSharp implementation is similar to Java and Python.
Download the latest version of word2number from nuget called w2ni18n.
On macOS
# git clone https://github.com/bastie/w2ni18n.git w2n
# cd ./w2n/csharp
# ./build.sh
word2number looking for your specific language with
1. Construct your W2N instance with ISO lang code parameter like es, fr, pt and if not
2. defined environment variable (not property) w2n.lang with ISO lang code like en, hi, de and if not found
3. over java.util.Locale.getDefault() and if null
4. over environment variable "LANGUAGE" and if not found
5. fallback to english
Place in the data directory your language specific dictionary file with ISO lang code in the name.
dotnet add MyNextProject.csproj package w2ni18n
Then import the namespace using the below code.
using word2number;
Then you can use the wordToNum method to convert a number-word to numeric digits, as shown below.
Environment.SetEnvironmentVariable ("w2n.lang","en");
W2N english = new W2N();
Console.WriteLine(english.wordToNum("one billion two million twenty three thousand and forty nine point two three six nine"));
In result it prints
1002023049.2369
Call the build.sh
script and use the new W2N.dll
file
- german language need more specific algorithm
- french language need update the property
point=virgule
. This is included in source but not in releases.
Add regex to fix comma bug
fixed by jnelson16fixed floating point conversation bug
accept also number values
because it is more understandable to handle str:112 as same as int:112
Thanks to word2number i18n coder and contributors
- Akshay Nagpal akshaynagpal
- Ben Batorsky bpben
- Alex ledovsky
- Tal Yarkoni tyarkoni
- ButteredGroove ButteredGroove
- Jonathan Nelson jnelsen16
- Daniel Junior danieljunior
- Kristian Toldy KToldy
The MIT License (MIT)
Copyright (c) 2016 Akshay Nagpal
Copyright (c) 2020-2021 Sebastian Ritter
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.