For example, if one was transcribing a Hindi text with many English loanwords and some stray characters of Simplified Chinese, one might use the following code (Python 3): from epitran.backoff import Backoff > backoff = Backoff (, cedict_file = ‘ cedict_1_0_ts_utf - 8 _mdbg. lang_script_codes is a list of codes like eng-Latn or hin-Deva. It also does not support punctuation normalization. Note that the Backoff class does not currently support parameterized preprocessor and postprocessor application and does not support non-standard ligatures. This functionality is provided by the Backoff class:īackoff(lang_script_codes, cedict_file=None) If one language mode does not work, it can be useful to fall back to another, and so on. Sometimes, when parsing text in more than one script, it is useful to employ a graceful backoff. Here is an example of an interaction with word_to_tuples (Python 2): > import epitran > epi = epitran. The structure of segments is as follows: ( The above data structure is likely to change in subsequent versions of the library. For example, "L" corresponds to letters and "P" corresponds to production marks. The codes for character_category are from the initial characters of the two character sequences listed in the "General Category" codes found in Chapter 4 of the Unicode Standard. Note that word_to_tuples is not implemented for all language-script pairs. The tuples have the following structure: ( Takes a word (a Unicode string) in a supported orthography as input and returns a list of tuples with each tuple corresponding to an IPA segment of the word. transliterate ( u 'Düğün' )) dyɰynĮpitran. transliterate ( u 'Düğün' ) u 'dy \u0270 yn' > print ( epi. Usage is illustrated below (Python 2): > epi. normpunc enables punctuation normalization and ligatures enables non-standard IPA ligatures like "ʤ" and "ʨ".
Convert text (in Unicode-encoded orthography of the language specified in the constructor) to IPA, which is returned. transliterate(text, normpunc=False, ligatures=False). The most useful public method of the Epitran class is transliterate:Įpitran. Epitran ( 'cmn-Hans', cedict_file = 'cedict_1_0_ts_utf-8_mdbg.txt' ) For Chinese, it is necessary to point the constructor to a copy of the CC-CEDict dictionary: > import epitran > epi = epitran. It is now possible to use the Epitran class for English and Mandarin Chinese (Simplified and Traditional) G2P as well as the other langugages that use Epitran's "classic" model. Epitran ( 'uig-Arab' ) # Uyghur in Perso-Arabic script
ENGLISH PHONETIZER ISO
Its constructor takes one argument, code, the ISO 639-3 code of the language to be transliterated plus a hyphen plus a four letter code for the script (e.g. The most general functionality in the epitran module is encapsulated in the very simple Epitran class:Įpitran(code, preproc=True, postproc=True, ligatures=False, cedict_file=None). Using the epitran Module The Epitran class
ENGLISH PHONETIZER INSTALL
If you wish to use Epitran to convert English to IPA, you must install the Flite (including lex_lookup) as detailed below. The Python modules epitran and epitran.vector can be used to easily write more sophisticated Python programs for deploying the Epitran mapping tables, preprocessors, and postprocessors. A library and tool for transliterating orthographic text as IPA (International Phonetic Alphabet).
0 kommentar(er)
