Phonostatistics¶

ruts.phon_stats.PhonStats

Description¶

A module for computing phonostatistics of a text: shares of sound classes, consonant clusters and hiatuses, CV pattern entropy, "hardness", alliteration and assonance indices, and syllable statistics. The data source can be either a text or a Doc object of the spaCy library.

The statistics are computed over letters without accounting for devoicing, reduction and stress. Letter classes are set by the ruts.constants constants:

Class	Letters	Constant
Vowels	а, е, и, о, у, ы, э, ю, я, ё	`RU_VOWELS`
Sonorant consonants	л, м, н, р, й	`RU_CONSONANTS_SONOR`, `RU_CONSONANTS_YET`
Voiced obstruents	б, в, г, д, ж, з	`RU_CONSONANTS_HIGH`
Voiceless obstruents	к, п, с, т, ф, х, ц, ч, ш, щ	`RU_CONSONANTS_LOW`
Signs	ь, ъ - not sounds, not counted in the shares	`RU_MARKS`

Words are divided into syllables by the rising sonority rule (Avanesov), see the syllabify function.

Syllables, CV patterns, clusters and hiatuses are computed once per word form and cached, the rest is aggregated over a counter of word forms, the alliteration and assonance indices - over a single word × letter matrix with numpy, so the time grows linearly with text length and on a corpus is dominated by tokenization.

Note

The statistics are computed when the PhonStats object is initialized.

Parameters¶

Parameter	Type	Default	Description
`source`	str/Doc	`-`	Data source (a string or a Doc object)
`words_extractor`	WordsExtractor	`None`	Word extraction tool
`window_len`	int	`3`	Window size in words for alliteration and assonance

Attributes¶

Attribute	Type	Description
`words`	tuple[str]	Tuple of extracted lowercased words
`syllables`	tuple[tuple[str, ...], ...]	Tuple of the syllables of each word
`n_vowels`	int	Number of vowels
`n_consonants`	int	Number of consonants
`n_sonorants`	int	Number of sonorant consonants
`n_voiced`	int	Number of voiced obstruents
`n_voiceless`	int	Number of voiceless obstruents
`n_marks`	int	Number of soft and hard signs
`c_clusters`	dict[int, int]	Distribution of consonant clusters by length
`c_syllable_patterns`	dict[str, int]	Distribution of syllables by CV pattern
`p_vowels`	float	Share of vowels among sounds
`p_sonorants`	float	Share of sonorant consonants among sounds
`p_voiced`	float	Share of voiced obstruents among sounds
`p_voiceless`	float	Share of voiceless obstruents among sounds
`consonant_vowel_ratio`	float	Consonant-to-vowel ratio
`p_heavy_clusters`	float	Share of clusters of 3 or more consonants
`p_hiatus`	float	Vowel hiatuses per word
`cv_entropy`	float	Entropy of word CV patterns in bits
`hardness`	float	Hardness - the ratio of voiceless obstruents to vowels and sonorants
`alliteration`	float	Alliteration index
`assonance`	float	Assonance index
`p_open_syllables`	float	Share of open syllables
`mean_syllable_len`	float	Mean syllable length in letters

Note

Every statistic can be computed separately by calling the corresponding function. Detailed information on the phonostatistics and the functions used to compute them is available in the corresponding section.

Methods¶

get_stats¶

Returns a dictionary with the computed phonostatistics.

Example

Code:

# Import the library
from ruts import PhonStats

# Prepare the data
text = "Ног нет, а хожу, рта нет, а скажу: когда спать, когда вставать, когда работу начинать"

# Compute the statistics
ps = PhonStats(text)
ps.get_stats()

Result:

{'p_vowels': 0.4032258064516129,
'p_sonorants': 0.11290322580645161,
'p_voiced': 0.1935483870967742,
'p_voiceless': 0.2903225806451613,
'consonant_vowel_ratio': 1.48,
'p_heavy_clusters': 0.034482758620689655,
'p_hiatus': 0.0,
'cv_entropy': 3.1395722619867223,
'hardness': 0.5625,
'alliteration': 0.9149440867502556,
'assonance': 0.802520508857449,
'p_open_syllables': 0.76,
'mean_syllable_len': 2.6}

Syllables and distributions are available as attributes:

Example

ps.syllables[:4]
# (('ног',), ('нет',), ('а',), ('хо', 'жу'))
ps.c_clusters
# {1: 22, 2: 6, 3: 1}
ps.c_syllable_patterns
# {'CCCV': 1, 'CCV': 5, 'CCVC': 1, 'CV': 11, 'CVC': 5, 'V': 2}

print_stats¶

Prints a table with the computed phonostatistics.

Example

Code:

...

# Print the table of computed statistics
ps.print_stats()

Result:

               Статистика               | Значение
--------------------------------------------------
Доля гласных                            |   0.40
Доля сонорных согласных                 |   0.11
Доля звонких шумных согласных           |   0.19
Доля глухих шумных согласных            |   0.29
Отношение согласных к гласным           |   1.48
Доля кластеров из 3 и более согласных   |   0.03
Зияний гласных на слово                 |   0.00
Энтропия CV-шаблонов слов (бит)         |   3.14
Жёсткость                               |   0.56
Индекс аллитерации                      |   0.91
Индекс ассонанса                        |   0.80
Доля открытых слогов                    |   0.76
Средняя длина слога (букв)              |   2.60