Module similar 

Is a new feature’s description close to a requirement the project already has? A deterministic text similarity — normalised tokens, a fixed stop-word list, a suffix stemmer, TF-IDF weights over the project’s requirements and cosine — so the answer is the same on every run and explains itself by the terms it shares. A hint: it never blocks, and no model produces or suppresses a match.

Structs

Place

Where a requirement is written.

Requirement

One requirement of the corpus.

SimilarMatch

A requirement close to the description.

SimilarReport

The similar payload: what was compared and what is close.

Enums

SimilarLevel

How close a match is.

Constants

CLOSE

A score at or above this is close.

MAX_MATCHES

At most this many matches.

MAX_SHARED

At most this many shared terms per match.

MAX_UNITS

The corpus is limited to the requirements of this many units, the most recently changed first.

VERY_CLOSE

A score at or above this is very close.

Functions

matches

The requirements of corpus at or above CLOSE to query, at most MAX_MATCHES, by score, then unit number, then identifier. The same input gives the same output; an empty query gives none.

shared_terms

The terms two texts share, sorted (for a symmetric check).

tokenize

The terms of text: NFKC-normalised, lowercased runs of letters and digits of two characters or more, stop words dropped, stemmed.