Koziev stress dictionary¶

StressDict(data_dir=DEFAULT_DATASET_DIR)

Ilya Koziev's stress dictionary (all_accents): 1,680,535 word forms with the position of the primary stress. Data from Wikipedia and Wiktionary is extended with forms from the inflection tables of the author's grammar dictionary under the assumption that stress does not move in inflection: forms with mobile stress (воды́ - во́ды) get one variant, homographs (за́мок - замо́к) share one line, ё is replaced with е.

License - CC0. The dictionary is downloaded from the author's repository at a pinned commit (a 10.6 MB archive, a 77 MB file) and read once per process: loading takes a fraction of a second, the dictionary occupies about 100 MB of memory.

The dictionary is used by VerseStats and the accentuation functions word_stress and accentuate: they correct known dictionary errors and restore stress from ё, and VerseStats also fits it to the meter.

Raises:

SourceTypeError - a data_dir that is neither a string nor a Path.

CodeResult

from ruts.datasets import StressDict

sd = StressDict()
sd.download()
sd.info["license"]

'CC0-1.0'

Parameters¶

Parameter	Type	Default	Description
`data_dir`	`str` / `Path`	`DEFAULT_DATASET_DIR`	Path to the dictionary directory; `dicts` in `DEFAULT_DATA_DIR` by default

Attributes¶

Attribute	Type	Description
`info`	`dict[str, str]`	Name of the dictionary and reference information about it

Methods¶

download(force=False) downloads the dictionary and extracts the file; if the archive exists but the dictionary file does not, the archive is extracted again. The archive is checked against a SHA-256 checksum, a corrupted or replaced file is deleted.

Parameter	Type	Default	Description
`force`	`bool`	`False`	Download the dictionary even if it is already downloaded

A file that cannot be downloaded or fails the checksum raises DownloadError, an archive that cannot be extracted DataFileError.

check_data() checks that the dictionary file is in place and returns True; a dictionary that is not downloaded raises DatasetNotFoundError with the commands to download it, and so does a lookup in it.

lookup(word) returns the zero-based number of the stressed syllable of a word form in any case, with or without ё (syllables are counted by vowels), None for a form outside the dictionary or without a stress mark.

Parameter	Type	Default	Description
`word`	`str`	-	Word form

A word form that is not a string raises SourceTypeError.

CodeResult

sd.lookup("корова"), sd.lookup("ёжик"), sd.lookup("котоведение")
"корова" in sd, len(sd)

(1, 0, None)
(True, 1680535)

get_records(limit=None) returns a generator of dictionary records in file (alphabetical) order: the word form and the number of the stressed syllable; get_texts with the same parameter returns a generator of word forms.

Parameter	Type	Default	Description
`limit`	`int`	`None`	Number of records

A limit that is not a non-negative integer raises ParameterError.

CodeResult

for record in sd.get_records(limit=2):
    print(record)
list(sd.get_texts(limit=2))

{'word': '-де', 'stress': 0}
{'word': '-ка', 'stress': 0}
['-де', '-ка']