Implementation:Googleapis Python genai Sphinx EnglishStemmer

Knowledge Sources	Googleapis_Python_genai Snowball English Stemmer
Domains	Documentation, NLP, Text_Processing
Last Updated	2026-02-15 14:00 GMT

Overview

Concrete tool implementing the English Porter/Snowball stemming algorithm for the Sphinx documentation search engine.

Description

The EnglishStemmer class extends BaseStemmer with English-specific morphological rules. It implements the multi-step Porter stemming algorithm: apostrophe handling (r_prelude), region marking (r_mark_regions), suffix removal steps 1a through 5 (r_Step_1a through r_Step_5), exception handling (r_exception1), and Y-to-y postprocessing (r_postlude). Contains 12 suffix lookup tables and character group definitions for vowels and consonants. This is a Sphinx documentation asset generated from Snowball grammar, not part of the GenAI SDK.

Usage

This JavaScript file is loaded by the Sphinx documentation site to reduce English words to their root form during full-text search, improving recall by matching related word forms.

Code Reference

Source Location

• Repository: Googleapis_Python_genai
• File: docs/_static/english-stemmer.js
• Lines: 1-1066

Signature

var EnglishStemmer = function() {
    BaseStemmer.call(this);

    this.stem = function() { ... };       // Execute full stemming pipeline
    this.stemWord = function(word) { ... }; // Convenience: stem a single word
};

Import

<script src="_static/base-stemmer.js"></script>
<script src="_static/english-stemmer.js"></script>

I/O Contract

Inputs

Name	Type	Required	Description
word	string	Yes (stemWord)	English word to stem

Outputs

Name	Type	Description
stemWord() returns	string	Root form of the input word
stem() modifies	internal state	Applies stemming rules to current buffer

Usage Examples

var stemmer = new EnglishStemmer();

// Stem individual words
stemmer.stemWord("running");   // "run"
stemmer.stemWord("connections"); // "connect"
stemmer.stemWord("generalized"); // "general"

// Used by Search.performTermsSearch() to match query terms

Related Pages

(No SDK environment or heuristic dependencies - documentation asset only)