Search_Retriever
Retrieves articles with the standard WordPress search used by the knowledge base.
Each keyword is searched on its own and articles are ranked by the keywords they match, rarer keywords counting for more. This handles natural-language questions, which rarely match every word, and doesn’t depend on how the site’s search combines words: search plugins may match any word rather than all of them. Long questions keep their rarest keywords, which say most about the topic. Both passes use WP_Query, so language plugins and other query filters still apply.
Namespace: WebberZone\Knowledge_Base\Pro\AI
Since: 3.2.0
Source: includes/pro/ai/class-search-retriever.php line 26
Implements: Retriever_Interface
Methods
__construct()
Constructor.
Line: 89
public function __construct( int $max_chars = ... );| Type | Name | Description |
|---|---|---|
int |
$max_chars |
Maximum characters of plain text per article. 0 uses the setting. |
retrieve()
Retrieve articles relevant to a question.
Since: 3.2.0
Line: 102
public function retrieve( string $question, int $limit ): array;| Type | Name | Description |
|---|---|---|
string |
$question |
Visitor question. |
int |
$limit |
Maximum number of articles. |
Returns: array — Articles, most relevant first.
prepare_article()
Build the article data sent as context.
Since: 3.2.0
Line: 140
public function prepare_article( WP_Post $post ): array;| Type | Name | Description |
|---|---|---|
\WP_Post |
$post |
Article. |
Returns: array — Article data.
get_plain_text()
Convert an article to plain text trimmed to the character limit.
Blocks are rendered, shortcodes stripped, HTML removed and whitespace collapsed. Block-level elements become line breaks so the text keeps its structure.
Since: 3.2.0
Line: 160
public function get_plain_text( WP_Post $post ): string;| Type | Name | Description |
|---|---|---|
\WP_Post |
$post |
Article. |
Returns: string — Plain text.
skip_block()
Skip blocks that add navigation or listings rather than article content.
Since: 3.2.0
Line: 188
public function skip_block( $pre_render, $parsed_block ): string|null;| Type | Name | Description |
|---|---|---|
string|null |
$pre_render |
Short-circuit value. |
array |
$parsed_block |
Parsed block. |
Returns: string|null — Empty string to skip the block, otherwise the original value.
get_skipped_blocks() static
Blocks left out of the article text.
Since: 3.2.0
Line: 205
static public function get_skipped_blocks( ): array;Returns: array — Block names.
html_to_text() static
Convert HTML to plain text.
Since: 3.2.0
Line: 241
static public function html_to_text( string $html ): string;| Type | Name | Description |
|---|---|---|
string |
$html |
HTML. |
Returns: string — Plain text with collapsed whitespace.
trim_text() static
Trim text to a maximum length, preferring a word boundary.
Since: 3.2.0
Line: 261
static public function trim_text( string $text, int $max_chars ): string;| Type | Name | Description |
|---|---|---|
string |
$text |
Text. |
int |
$max_chars |
Maximum characters, including the ellipsis. |
Returns: string — Trimmed text.
extract_keywords() static
Extract search keywords from a question.
Since: 3.2.0
Line: 286
static public function extract_keywords( string $question, int $max = ... ): array;| Type | Name | Description |
|---|---|---|
string |
$question |
Question. |
int |
$max |
Maximum number of keywords, in question order. |
Returns: array — Keywords.
get_stop_words() static
Words dropped from questions before searching.
Since: 3.2.0
Line: 315
static public function get_stop_words( ): array;Returns: array — Stop words.