Mycroft's multilingual text parsing and formatting library
Lingua Franca (noun)
a framework that is adopted as the common language between speakers with different native tongues
Convert data into spoken equivalents
spoken versions of numbers
from lingua_franca.format import nice_number, pronounce_number
assert nice_number(25/6) == "4 and a sixth"
assert nice_number(201) == "201"
assert nice_number(3.14159269) == "3 and a seventh"
assert pronounce_number(3.14159269) == "three point one four"
assert pronounce_number(0) == "zero"
assert pronounce_number(10) == "ten"
assert pronounce_number(201) == "two hundred and one"
assert pronounce_number(102.3) == "one hundred and two point three"
assert pronounce_number(
4092949192) == "four billion, ninety two million, nine hundred and forty nine thousand, one hundred and ninety two"
assert pronounce_number(100034000000299792458, short_scale=True) == \
"one hundred quintillion, thirty four quadrillion, " \
"two hundred and ninety nine million, seven hundred and ninety " \
"two thousand, four hundred and fifty eight"
assert pronounce_number(100034000000299792458, short_scale=False) == \
"one hundred trillion, thirty four thousand billion, " \
"two hundred and ninety nine million, seven hundred and ninety " \
"two thousand, four hundred and fifty eight"spoken date for datetime.datetime objects
from lingua_franca.format import nice_date, nice_date_time, nice_time
import datetime
dt = datetime.datetime(2017, 1, 31, 13, 22, 3)
assert nice_date(dt) == "tuesday, january thirty-first, twenty seventeen"
assert nice_time(dt) == "one twenty two"
assert nice_time(dt, use_ampm=True) == "one twenty two p.m."
assert nice_time(dt, speech=False) == "1:22"
assert nice_time(dt, speech=False, use_ampm=True) == "1:22 PM"
assert nice_time(dt, speech=False, use_24hour=True) == "13:22"
assert nice_time(dt, speech=False, use_24hour=True, use_ampm=True) == "13:22"
assert nice_time(dt, use_24hour=True, use_ampm=True) == "thirteen twenty two"
assert nice_time(dt, use_24hour=True, use_ampm=False) == "thirteen twenty two"
assert nice_date_time(dt) == "tuesday, january thirty-first, twenty seventeen at one twenty two"spoken number of seconds or datetime.timedelta objects
from lingua_franca.format import nice_duration
assert nice_duration(1) == "one second"
assert nice_duration(3) == "three seconds"
assert nice_duration(1, speech=False) == "0:01"
assert nice_duration(61), "one minute one second"
assert nice_duration(61, speech=False) == "1:01"
assert nice_duration(5000) == "one hour twenty three minutes twenty seconds"
assert nice_duration(5000, speech=False), "1:23:20"
assert nice_duration(50000) == "thirteen hours fifty three minutes twenty seconds"
assert nice_duration(50000, speech=False) == "13:53:20"
assert nice_duration(500000) == "five days eighteen hours fifty three minutes twenty seconds"
assert nice_duration(500000, speech=False), "5d 18:53:20"
from datetime import timedelta
assert nice_duration(timedelta(seconds=500000), speech=False) == "5d 18:53:20"Extract data from natural language text
from lingua_franca.parse import extract_number, extract_numbers
# extract a number
assert extract_number("nothing") is False
assert extract_number("two million five hundred thousand tons of spinning "
"metal") == 2500000
assert extract_number("six trillion") == 6000000000000.0
assert extract_number("six trillion", short_scale=False) == 6e+18
assert extract_number("1 and 3/4 cups") == 1.75
assert extract_number("1 cup and a half") == 1.5
## extracts all numbers
assert extract_numbers("nothing") == []
assert extract_numbers("this is a one twenty one test") == [1.0, 21.0]
assert extract_numbers("1 dog, seven pigs, macdonald had a farm, "
"3 times 5 macarena") == [1, 7, 3, 5]extract datetime.timedelta objects
## extract durations
from lingua_franca.parse import extract_duration
from datetime import timedelta
assert extract_duration("nothing") == (None, 'nothing')
assert extract_duration("Nineteen minutes past the hour") == (
timedelta(minutes=19),
"past the hour")
assert extract_duration("wake me up in three weeks, four hundred ninety seven"
" days, and three hundred 91.6 seconds") == (
timedelta(weeks=3, days=497, seconds=391.6),
"wake me up in , , and")
assert extract_duration(
"The movie is one hour, fifty seven and a half minutes long") == (
timedelta(hours=1, minutes=57.5),
"the movie is , long")extract datetime.datetime objects
## extract date times
from datetime import datetime
from lingua_franca.parse import extract_datetime, normalize
def extractWithFormat(text):
date = datetime(2017, 6, 27, 13, 4) # Tue June 27, 2017 @ 1:04pm
[extractedDate, leftover] = extract_datetime(text, date)
extractedDate = extractedDate.strftime("%Y-%m-%d %H:%M:%S")
return [extractedDate, leftover]
def testExtract(text, expected_date, expected_leftover):
res = extractWithFormat(normalize(text))
assert res[0] == expected_date
assert res[1] == expected_leftover
testExtract("now is the time",
"2017-06-27 13:04:00", "is time")
testExtract("in a couple minutes",
"2017-06-27 13:06:00", "")
testExtract("What is the day after tomorrow's weather?",
"2017-06-29 00:00:00", "what is weather")
testExtract("Remind me at 10:45 pm",
"2017-06-27 22:45:00", "remind me")
testExtract("what is the weather on friday morning",
"2017-06-30 08:00:00", "what is weather")
testExtract("what is tomorrow's weather",
"2017-06-28 00:00:00", "what is weather")
testExtract("remind me to call mom next tuesday",
"2017-07-04 00:00:00", "remind me to call mom")
testExtract("remind me to call mom in 3 weeks",
"2017-07-18 00:00:00", "remind me to call mom")
testExtract("set an alarm for tonight 9:30",
"2017-06-27 21:30:00", "set alarm")
testExtract("on the evening of june 5th 2017 remind me to call my mother",
"2017-06-05 19:00:00", "remind me to call my mother")check other languages
If you wish to develop a corpus in a language other than English, then Mimic Recording Studio can be used to produce voice recordings for TTS voices in additional languages. If you are building a corpus in a language other than English, we encourage you to choose phrases which:
occur in natural, everyday speech in the target language
have a variety of string lengths
cover a wide variety of phonemes (basic sounds)
there is a file generator that generates any sentences from wikipedia. just call the command.
'python3 backend/corpus_file_gen.py' to run. you are always asked about the wiki language 'en'.
if you have only a simple text file without line length and tab you can only check the file.
python3 backend/corpus_file_gen.py --prepare_file 3 --file english_corpus.csv if you have already started a file, the generator will expand the file 35K.
python3 backend/corpus_file_gen.py --prepare_file 1 --file english_corpus.csv or --help for help.
This is a very simple generator. You should always check the file and delite false records. We are working on a solution to change numbers into words
IMPORTANT: For now, you must reset the sqlite database to use a new corpus. If you've recorded on another corpus and would like to save that data, you can simply rename your sqlite db found in backend/db/ to another name. The backend will detect that mimicstudio.db is not there and create a new one for you. You may continue recording data for your new corpus.