English documentation

This section has four parts. Which one you want depends on what you are trying to do right now.

What you want	Where to go
I have never used this, walk me through it	Tutorial
I need to get a specific job done	How-to
I am looking up what a feature is	Reference
I want to understand why it works this way	Explanation
I want to start from a working script	Example scripts (ten scripts, each answering one question; comments in Turkish)

If you are not sure, start with the tutorial: it takes you from installation to your first table.

At a glance

import turkish_linguistic_features as tlf

oz = tlf.analyze("Colorless green ideas sleep furiously.", lang="en")
len(oz)               # 181
oz["word_len_mean"]   # 6.6

The library has eleven public names. That is all of them:

Name	What it does
analyze	Extracts every feature from one text
analyze_corpus	Runs analyze on every text in a folder
ngram_matches	Shows what a sequence of words or POS tags you give (such as ["ADJ", "NOUN"]) matched in a text, with counts
segment_text	Splits a text into fixed-size segments
save_csv	Writes the resulting rows to CSV
describe_feature	Gives a feature's definition, formula, scale, the minimum data it needs, and its citation
FeatureParams	The settings object holding thresholds and windows
LinguisticFeaturesError	Base class for the library's own errors
ModelNotFoundError	Raised when required language data is missing: a spaCy model, or NLTK's cmudict for English syllable counts
MissingDependencyWarning	Warned when an optional package is missing
ParagraphStructureWarning	Warned when a text over 1000 words has no paragraph boundary

Every feature, including the phrases you count with custom_ngrams, comes out of analyze.

Terms

The documentation uses fixed terminology. If a word is unclear: glossary.