Exceptions

All exceptions of the library inherit the base class AnyTSError and a built-in Python class, and can be caught by either.

Exception	Built-in class	When raised
AnyTSError	Exception	Base class, never raised itself
SourceTypeError	TypeError	An argument is of an unsupported type: the source or the text, a text where a list is expected, an extractor, basic statistics of another class, a path, a hook that is not callable (a tokenizer, a lemmatizer, a measure), or a tokenizer, a sentence splitter or the parse, analyze, syllabify, transcribe or find of a class that returns what the core cannot use
SourceError	ValueError	The source has no words, sentences or texts to compute a statistic on, lacks an annotation the statistic needs, or holds values it cannot take: a negative or infinite frequency, a matrix of distances that is not one
ParameterError	ValueError	A parameter is out of range or not a finite number, a name (of a measure, a variant, a preset, a layer, a norm) is unknown, the labels or the axes of a plot are set incorrectly, or a layer of the highlighting asks for an annotation the source lacks
UnknownStatError	ParameterError, KeyError	An unknown statistic is requested by name
DatasetNotFoundError	OSError	The spaCy model is not installed or a dataset is not downloaded; the message gives the command that fixes it
DataFileError	ValueError	The archive of a dataset cannot be extracted or is unsafe, or a line of a dataset file cannot be read
DownloadError	RuntimeError	A file cannot be downloaded, is cut short or fails the checksum verification, its name is not a plain name, or its directory cannot be created

The classes are available from anyts and from anyts.exceptions. A language library re-exports them under its own names as aliases, not subclasses, so its base class catches the errors of the core too.

CodeResult

from anyts import AnyTSError, ParameterError, WordsExtractor

try:
    WordsExtractor(ngram_range=(2, 1))
except ParameterError as e:
    print(e)

try:
    WordsExtractor(tokenizer=42).extract("The cat sleeps.")
except AnyTSError as e:
    print(type(e).__name__)

The lower N-gram bound is greater than the upper
SourceTypeError

Messages

The errors of the classes of statistics (Stats and its subclasses), of StatsComponent, of the extractors SentsExtractor, WordsExtractor and CharNgramsExtractor, of ConnectorMatcher and of HighlightedText take their messages from the class attribute messages, a mapping of a key to a message, anyts.constants.MESSAGES in English by default, a read-only mapping shared by every class; the messages with fields in braces are format strings. A language library sets the messages of its language on its subclasses, messages = MappingProxyType({**MESSAGES, **own}), and they cover every error such a class raises itself; source_type_error of anyts.utils gives the error of a source of a wrong type by them. The functions check_params, calc_windowed, calc_reading_time, split_windows and sentence_rhythm take the messages as the keyword messages.

The shared checks of anyts.utils (check_integer, check_number, check_bool, check_words, check_sequence), which these classes call too, and the other functions of the core keep their messages in English.

CodeResult

from types import MappingProxyType

from anyts.constants import MESSAGES
from anyts.diversity_stats import DiversityStats


class Diversity(DiversityStats):
    messages = MappingProxyType({**MESSAGES, "no_words": "Il testo non contiene parole"})


try:
    Diversity([])
except ValueError as e:
    print(e)

Il testo non contiene parole