Source code for cltk.phonology.enm.stress

"""Middle English stress module
"""

import re
from typing import List

from cltk.phonology.enm.syllabifier import (
    CONSONANTS,
    DIPHTHONGS,
    SHORT_VOWELS,
    TRIPHTHONGS,
)
from cltk.stem.enm import stem

__author__ = [
    "Eleftheria Chatziargyriou <ele.hatzy@gmail.com>",
    "Clément Besnier <clem@clementbesnier.fr>",
]

# Soundex replacement rules
dict_SE = {
    "p": "1",
    "b": "1",
    "f": "1",
    "v": "1",
    "gh": "1",
    "sh": "2",
    "t": "2",
    "d": "2",
    "s": "2",
    "z": "2",
    "r": "2",  # TODO
    "k": "2",
    "g": "2",
    "w": "2",
    "l": "3",
    "m": "4",
    "n": "4",
    "r": "5",  # TODO
}


[docs]class MiddleEnglishStresser: """ Middle English stresser """ def __init__(self, syllabifier=None): """ :param syllabifier: Syllabifier instance """ self.syllabifier = syllabifier
[docs] def stress(self, word, stress_rule="FSR") -> List: """ :param word: word to stress :param stress_rule: Stress Rule, valid options: 'FSR': French Stress Rule, stress falls on the ultima, unless it contains schwa (ends with e), in which case the penult is stressed. 'GSR': Germanic Stress Rule, stress falls on the first syllable of the stemm. Note that the accuracy of the function directly depends on that of the stemmer. 'LSR': Latin Stress Rule, stress falls on the penult if its heavy, else, if it has more than two syllables on the antepenult, else on the ultima. :return: A list containing the separate syllable, where the stressed syllable is prefixed by ' . Monosyllabic words are left unchanged, since stress indicates relative emphasis. Examples: >>> from cltk.phonology.syllabify import Syllabifier >>> from cltk.phonology.enm.syllabifier import DIPHTHONGS, TRIPHTHONGS, SHORT_VOWELS, LONG_VOWELS >>> enm_syllabifier = Syllabifier() >>> enm_syllabifier.set_short_vowels(SHORT_VOWELS) >>> enm_syllabifier.set_vowels(SHORT_VOWELS+LONG_VOWELS) >>> enm_syllabifier.set_diphthongs(DIPHTHONGS) >>> enm_syllabifier.set_triphthongs(TRIPHTHONGS) >>> stresser = MiddleEnglishStresser(enm_syllabifier) >>> stresser.stress('beren', stress_rule="FSR") ['ber', "'en"] >>> stresser.stress('prendre', stress_rule="FSR") ["'pren", 'dre'] >>> stresser.stress('yisterday', stress_rule="GSR") ['yi', 'ster', "'day"] >>> stresser.stress('day', stress_rule="GSR") ['day'] >>> stresser.stress('mervelus', stress_rule="LSR") ["'mer", 'vel', 'us'] >>> stresser.stress('verbum', stress_rule="LSR") ['ver', "'bum"] """ assert self.syllabifier is not None # Syllabify word syllabified = self.syllabifier.syllabify(word, mode="MOP") # Check whether word is monosyllabic if len(syllabified) == 1: return syllabified if stress_rule == "FSR": # Check whether ultima ends in e if syllabified[-1][-1] == "e": return ( syllabified[:-2] + ["'{0}".format(syllabified[-2])] + syllabified[-1:] ) else: return syllabified[:-1] + ["'{0}".format(syllabified[-1])] elif stress_rule == "GSR": # The word striped of suffixes st_word = stem(word, strip_suf=False) affix = word[: len(word) - len(st_word)] # Syllabify stripped word and affix syl_word = self.syllabifier.syllabify(st_word, mode="MOP") # Add stress syl_word = ["'{0}".format(syl_word[0])] + syl_word[1:] if affix: affix = self.syllabifier.syllabify(affix, mode="MOP") syl_word = affix + syl_word return syl_word elif stress_rule == "LSR": # Check whether penult is heavy (contains more than one mora) if sum(map(lambda x: x in SHORT_VOWELS, syllabified[-1])) > 1: return ( syllabified[:-2] + ["'{0}".format(syllabified[-2])] + syllabified[-1:] ) elif len(syllabified) > 2: return ( syllabified[:-3] + ["'{0}".format(syllabified[-3])] + syllabified[-2:] ) else: return syllabified[:-1] + ["'{0}".format(syllabified[-1])]
[docs] def phonetic_indexing(self, word, p="SE") -> str: """ :param word: word :param p: Specifies the phonetic indexing method SE: Soundex variant for MHG :return: Encoded string corresponding to the word's phonetic representation """ if p == "SE": return self._soundex(word)
[docs] def _soundex(self, word): """ The Soundex phonetic indexing algorithm adapted to ME phonology. Algorithm: Let w the original word and W the resulting one 1) Capitalize the first letter of w and append it to W 2) Apply the following replacement rules p, b, f, v, gh (non-nasal fricatives) -> 1 t, d, s, sh, z, r, k, g, w (non-nasal alveolars and velars) -> 2 l (alveolar lateral) -> 3 m, n (nasals) -> 4 r (alveolar approximant) -> 5 3) Concetate multiple occurrences of numbers into one 4) Remove non-numerical characters Notes: /h/ was thought to be either a voiceless or velar fricative when occurring in the coda with its most used grapheme being <gh>. Those phonemes either disappeared, resulting in the lengthening of preceding vowel clusters, or were developed into /f/ as evident by modern spelling (e.g. 'enough': /ɪˈnʌf/ and 'though': /ðəʊ/) Examples: >>> MiddleEnglishStresser().phonetic_indexing("midel", "SE") 'M230' >>> MiddleEnglishStresser().phonetic_indexing("myddle", "SE") 'M230' >>> MiddleEnglishStresser().phonetic_indexing("might", "SE") 'M120' >>> MiddleEnglishStresser().phonetic_indexing("myghtely", "SE") 'M123' """ self.word = word word = self.word[1:] for w, val in zip(dict_SE.keys(), dict_SE.values()): word = word.replace(w, val) # Remove multiple adjacent occurences of digit word = re.sub(r"(\d)\1+", r"\1", word) # Strip remaining letters word = re.sub(r"[a-zðþƿ]+", "", word) # Add trailing zeroes and return return (self.word[0].upper() + word + "0" * 3)[:4]