Add a Relevance sort order and improve search

Here's what's in `IndexableText` right now:

1. Title
2. Collections
3. Authors
4. Tags
5. LocSubjects
6. TocEntries

Here is the proposed new ranking:

```
10 * Title +
8 * Authors +
3 * Collections +
IndexableText
```

New indices for existing DBs:

```
ALTER TABLE `Ebooks` ADD COLUMN `IndexableAuthors` text NOT NULL;
ALTER TABLE `Ebooks` ADD COLUMN `IndexableCollections` text NULL;
ALTER TABLE `Ebooks` ADD FULLTEXT `indexSearchTitle` (`Title`);
ALTER TABLE `Ebooks` ADD FULLTEXT `idxSearchAuthors` (`IndexableAuthors`);
ALTER TABLE `Ebooks` ADD FULLTEXT `idxSearchCollections` (`IndexableCollections`);
```

config/sql/se/Ebooks.sql

@@ -25,8 +25,13 @@ CREATE TABLE IF NOT EXISTS `Ebooks` (
   `EbookUpdated` datetime NULL,
   `TextSinglePageByteCount` bigint unsigned NULL,
   `IndexableText` text NOT NULL,
+  `IndexableAuthors` text NOT NULL,
+  `IndexableCollections` text NULL,
   PRIMARY KEY (`EbookId`),
   UNIQUE KEY `index1` (`Identifier`),
   KEY `index2` (`EbookCreated`),
-  FULLTEXT `idxSearch` (`IndexableText`)
+  FULLTEXT `idxSearch` (`IndexableText`),
+  FULLTEXT `idxSearchTitle` (`Title`),
+  FULLTEXT `idxSearchAuthors` (`IndexableAuthors`),
+  FULLTEXT `idxSearchCollections` (`IndexableCollections`)
 ) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4;

lib/Constants.php

@@ -44,6 +44,10 @@ const EBOOKS_MAX_STRING_LENGTH = 250;
 const EBOOKS_MAX_LONG_STRING_LENGTH = 500;
 const EBOOK_SINGLE_PAGE_SIZE_WARNING = 3 * 1024 * 1024; // 3145728 bytes.
 
+const EBOOK_SEARCH_WEIGHT_TITLE = 10;
+const EBOOK_SEARCH_WEIGHT_AUTHORS = 8;
+const EBOOK_SEARCH_WEIGHT_COLLECTIONS = 3;
+
 const ARTWORK_THUMBNAIL_HEIGHT = 350;
 const ARTWORK_THUMBNAIL_WIDTH = 350;
 const ARTWORK_PER_PAGE = 20;

lib/Ebook.php

@@ -44,6 +44,8 @@ use function Safe\shell_exec;
  * @property string $TextSinglePageUrl
  * @property string $TextSinglePageSizeFormatted
  * @property string $IndexableText
+ * @property string $IndexableAuthors
+ * @property ?string $IndexableCollections
  * @property ?EbookPlaceholder $EbookPlaceholder
  * @property array<Project> $Projects
  * @property array<Project> $PastProjects
@@ -128,6 +130,8 @@ final class Ebook{
 	protected string $_TextSinglePageUrl;
 	protected string $_TextSinglePageSizeFormatted;
 	protected string $_IndexableText;
+	protected string $_IndexableAuthors;
+	protected ?string $_IndexableCollections = null;
 	protected ?EbookPlaceholder $_EbookPlaceholder = null;
 	/** @var array<Project> $_Projects */
 	protected array $_Projects;
@@ -732,13 +736,40 @@ final class Ebook{
 				}
 			}
 
-			// Remove diacritics and non-alphanumeric characters.
+			$this->_IndexableText = Formatter::RemoveDiacriticsAndNonalphanumerics($this->_IndexableText);
-			$this->_IndexableText = trim(preg_replace('|[^a-zA-Z0-9 ]|ius', ' ', Formatter::RemoveDiacritics($this->_IndexableText)));
 		}
 
 		return $this->_IndexableText;
 	}
 
+	protected function GetIndexableAuthors(): string{
+		if(!isset($this->_IndexableAuthors)){
+			$this->_IndexableAuthors = '';
+
+			foreach($this->Authors as $author){
+				$this->_IndexableAuthors .= ' ' . $author->Name;
+			}
+
+			$this->_IndexableAuthors = Formatter::RemoveDiacriticsAndNonalphanumerics($this->_IndexableAuthors);
+		}
+
+		return $this->_IndexableAuthors;
+	}
+
+	protected function GetIndexableCollections(): ?string{
+		if(!isset($this->_IndexableCollections)){
+			foreach($this->CollectionMemberships as $collectionMembership){
+				$this->_IndexableCollections .= ' ' . $collectionMembership->Collection->Name;
+			}
+
+			if(isset($this->_IndexableCollections)){
+				$this->_IndexableCollections = Formatter::RemoveDiacriticsAndNonalphanumerics($this->_IndexableCollections);
+			}
+		}
+
+		return $this->_IndexableCollections;
+	}
+
 	protected function GetEbookPlaceholder(): ?EbookPlaceholder{
 		if(!isset($this->_EbookPlaceholder)){
 			if(!isset($this->EbookId)){
@@ -1561,6 +1592,22 @@ final class Ebook{
 			$error->Add(new Exceptions\EbookIndexableTextRequiredException());
 		}
 
+		if(isset($this->IndexableAuthors)){
+			$this->IndexableAuthors = trim($this->IndexableAuthors ?? '');
+
+			if($this->IndexableAuthors == ''){
+				$error->Add(new Exceptions\EbookIndexableAuthorsRequiredException());
+			}
+		}
+		else{
+			$error->Add(new Exceptions\EbookIndexableAuthorsRequiredException());
+		}
+
+		$this->IndexableCollections = trim($this->IndexableCollections ?? '');
+		if($this->IndexableCollections == ''){
+			$this->IndexableCollections = null;
+		}
+
 		if(isset($this->EbookPlaceholder)){
 			try{
 				$this->EbookPlaceholder->Validate();
@@ -1897,7 +1944,8 @@ final class Ebook{
 			INSERT into Ebooks (Identifier, WwwFilesystemPath, RepoFilesystemPath, KindleCoverUrl, EpubUrl,
 				AdvancedEpubUrl, KepubUrl, Azw3Url, DistCoverUrl, Title, FullTitle, AlternateTitle,
 				Description, LongDescription, Language, WordCount, ReadingEase, GitHubUrl, WikipediaUrl,
-				EbookCreated, EbookUpdated, TextSinglePageByteCount, IndexableText)
+				EbookCreated, EbookUpdated, TextSinglePageByteCount, IndexableText, IndexableAuthors,
+				IndexableCollections)
 			values (?,
 				?,
 				?,
@@ -1920,12 +1968,15 @@ final class Ebook{
 				?,
 				?,
 				?,
+				?,
+				?,
 				?)
 		', [$this->Identifier, $this->WwwFilesystemPath, $this->RepoFilesystemPath, $this->KindleCoverUrl, $this->EpubUrl,
 				$this->AdvancedEpubUrl, $this->KepubUrl, $this->Azw3Url, $this->DistCoverUrl, $this->Title,
 				$this->FullTitle, $this->AlternateTitle, $this->Description, $this->LongDescription,
 				$this->Language, $this->WordCount, $this->ReadingEase, $this->GitHubUrl, $this->WikipediaUrl,
-				$this->EbookCreated, $this->EbookUpdated, $this->TextSinglePageByteCount, $this->IndexableText]);
+				$this->EbookCreated, $this->EbookUpdated, $this->TextSinglePageByteCount, $this->IndexableText,
+				$this->IndexableAuthors, $this->IndexableCollections]);
 
 		$this->EbookId = Db::GetLastInsertedId();
 
@@ -1990,7 +2041,9 @@ final class Ebook{
 				EbookCreated = ?,
 				EbookUpdated = ?,
 				TextSinglePageByteCount = ?,
-				IndexableText = ?
+				IndexableText = ?,
+				IndexableAuthors = ?,
+				IndexableCollections = ?
 				where
 				EbookId = ?
 			', [$this->Identifier, $this->WwwFilesystemPath, $this->RepoFilesystemPath, $this->KindleCoverUrl, $this->EpubUrl,
@@ -1998,6 +2051,7 @@ final class Ebook{
 					$this->FullTitle, $this->AlternateTitle, $this->Description, $this->LongDescription,
 					$this->Language, $this->WordCount, $this->ReadingEase, $this->GitHubUrl, $this->WikipediaUrl,
 					$this->EbookCreated, $this->EbookUpdated, $this->TextSinglePageByteCount, $this->IndexableText,
+					$this->IndexableAuthors, $this->IndexableCollections,
 					$this->EbookId]);
 		}
 		catch(Exceptions\DuplicateDatabaseKeyException){
@@ -2380,6 +2434,7 @@ final class Ebook{
 	public static function GetAllByFilter(string $query = null, array $tags = [], Enums\EbookSortType $sort = null, int $page = 1, int $perPage = EBOOKS_PER_PAGE, Enums\EbookReleaseStatusFilter $releaseStatusFilter = Enums\EbookReleaseStatusFilter::All): array{
 		$limit = $perPage;
 		$offset = (($page - 1) * $perPage);
+		$relevanceScoreField = '';
 		$joinContributors = '';
 		$joinTags = '';
 		$params = [];
@@ -2424,10 +2479,22 @@ final class Ebook{
 		}
 
 		if($query !== null && $query != ''){
-			$query = trim(preg_replace('|[^a-zA-Z0-9 ]|ius', ' ', Formatter::RemoveDiacritics($query)));
+			// Preserve quotes in the query so the user can enter, e.g., "war and peace" for an exact match.
-			$query = sprintf('"%s"', $query);  // Require an exact match via double quotes.
+			$query = trim(preg_replace('|[^a-zA-Z0-9" ]|ius', ' ', Formatter::RemoveDiacritics($query)));
-			$whereCondition .= ' and match(e.IndexableText) against(? in boolean mode) ';
+
+			$relevanceScoreField = ', (
+				match(e.Title) against (?) * ' . EBOOK_SEARCH_WEIGHT_TITLE . ' +
+				match(e.IndexableAuthors) against (?) * ' . EBOOK_SEARCH_WEIGHT_AUTHORS . ' +
+				match(e.IndexableCollections) against (?) * ' . EBOOK_SEARCH_WEIGHT_COLLECTIONS . ' +
+				match(e.IndexableText) against (?)
+			) as relevance_score ';
+
+			$whereCondition .= ' and match(e.IndexableText) against(?) ';
 			$params[] = $query;
+
+			if($sort == null || $sort == Enums\EbookSortType::Relevance || $sort == Enums\EbookSortType::Newest){
+				$orderBy = 'relevance_score desc, e.EbookCreated desc';
+			}
 		}
 
 		try{
@@ -2439,11 +2506,17 @@ final class Ebook{
 					' . $whereCondition . '
 					', $params);
 
+			if($relevanceScoreField != ''){
+				// `relevance_score` is at the beginning of the query, so these params must go at the start of the array.
+				array_unshift($params, $query, $query, $query, $query);
+			}
+
 			$params[] = $limit;
 			$params[] = $offset;
 
 			$ebooks = Db::Query('
 					SELECT distinct e.*
+					' . $relevanceScoreField . '
 					from Ebooks e
 					' . $joinContributors . '
 					' . $joinTags . '

lib/Enums/EbookSortType.php

@@ -6,4 +6,5 @@ enum EbookSortType: string{
 	case AuthorAlpha = 'author-alpha';
 	case ReadingEase = 'reading-ease';
 	case Length = 'length';
+	case Relevance = 'relevance';
 }

lib/Exceptions/EbookIndexableAuthorsRequiredException.php

@@ -0,0 +1,7 @@
+<?
+namespace Exceptions;
+
+class EbookIndexableAuthorsRequiredException extends AppException{
+	/** @var string $message */
+	protected $message = 'Ebook IndexableAuthors required.';
+}

lib/Formatter.php

@@ -31,6 +31,13 @@ class Formatter{
 		}
 	}
 
+	/**
+	 * Remove diacritics and non-alphanumeric characters.
+	 */
+	public static function RemoveDiacriticsAndNonalphanumerics(string $text): string{
+		return trim(preg_replace('|[^a-zA-Z0-9 ]|ius', ' ', Formatter::RemoveDiacritics($text)));
+	}
+
 	/**
 	 * Escape a string so that it's appropriate to use in a URL slug.
 	 *

templates/SearchForm.php

@@ -18,13 +18,17 @@ $isAllSelected = sizeof($tags) == 0 || in_array('all', $tags);
 		</select>
 	</label>
 	<label>Keywords
-		<input type="search" name="query" value="<?= Formatter::EscapeHtml($query ?? '') ?>"/>
+		<input type="search" name="query" placeholder="Quotes allowed, e.g., &quot;war and peace&quot;" value="<?= Formatter::EscapeHtml($query ?? '') ?>"/>
 	</label>
 	<label class="sort">
 		<span>Sort</span>
 		<span>
 			<select name="sort">
-				<option value="<?= Enums\EbookSortType::Newest->value ?>"<? if($sort == Enums\EbookSortType::Newest){ ?> selected="selected"<? } ?>>S.E. release date (new &#x2192; old)</option>
+				<? if(isset($query) && $query != ''){ ?>
+					<option value="<?= Enums\EbookSortType::Relevance->value ?>"<? if($sort == Enums\EbookSortType::Relevance){ ?> selected="selected"<? } ?>>Relevance</option>
+				<? }else{ ?>
+					<option value="<?= Enums\EbookSortType::Newest->value ?>"<? if($sort == Enums\EbookSortType::Newest){ ?> selected="selected"<? } ?>>S.E. release date (new &#x2192; old)</option>
+				<? } ?>
 				<option value="<?= Enums\EbookSortType::AuthorAlpha->value ?>"<? if($sort == Enums\EbookSortType::AuthorAlpha){ ?> selected="selected"<? } ?>>Author name  (a &#x2192; z)</option>
 				<option value="<?= Enums\EbookSortType::ReadingEase->value ?>"<? if($sort == Enums\EbookSortType::ReadingEase){ ?> selected="selected"<? } ?>>Reading ease (easy &#x2192; hard)</option>
 				<option value="<?= Enums\EbookSortType::Length->value ?>"<? if($sort == Enums\EbookSortType::Length){ ?> selected="selected"<? } ?>>Length (short &#x2192; long)</option>

www/ebooks/index.php

@@ -28,6 +28,14 @@ try{
 		$view = null;
 	}
 
+	if($query != ''){
+		$queryStringParams['query'] = $query;
+		// If the user entered a query with the default sort order, change it to relevance sort.
+		if($sort == Enums\EbookSortType::Newest){
+			$sort = Enums\EbookSortType::Relevance;
+		}
+	}
+
 	if($sort == Enums\EbookSortType::Newest){
 		$sort = null;
 	}
@@ -38,10 +46,6 @@ try{
 
 	$pageDescription = 'Page ' . $page . ' of the Standard Ebooks free ebook library';
 
-	if($query != ''){
-		$queryStringParams['query'] = $query;
-	}
-
 	if(sizeof($tags) > 0){
 		$queryStringParams['tags'] = $tags;
 	}