mirror of
https://github.com/advplyr/audiobookshelf.git
synced 2025-06-23 01:18:59 +02:00
Merge 5017e7ce9e
into 108b2a60f5
This commit is contained in:
commit
017e51f2da
@ -16,6 +16,7 @@
|
||||
<p v-if="book.author" class="text-gray-300 text-xs md:text-sm">{{ $getString('LabelByAuthor', [book.author]) }}</p>
|
||||
<p v-if="book.narrator" class="text-gray-400 text-xs">{{ $strings.LabelNarrators }}: {{ book.narrator }}</p>
|
||||
<p v-if="book.duration" class="text-gray-400 text-xs">{{ $strings.LabelDuration }}: {{ $elapsedPrettyExtended(bookDuration, false) }} {{ bookDurationComparison }}</p>
|
||||
<p v-if="book.matchConfidence" class="text-gray-400 text-xs">{{ $strings.LabelMatchConfidence }}: {{ book.matchConfidence.toFixed(3) }}</p>
|
||||
<div v-if="book.series?.length" class="flex py-1 -mx-1">
|
||||
<div v-for="(series, index) in book.series" :key="index" class="bg-white/10 rounded-full px-1 py-0.5 mx-1">
|
||||
<p class="leading-3 text-xs text-gray-400">
|
||||
|
@ -425,6 +425,7 @@
|
||||
"LabelLogLevelWarn": "Warn",
|
||||
"LabelLookForNewEpisodesAfterDate": "Look for new episodes after this date",
|
||||
"LabelLowestPriority": "Lowest Priority",
|
||||
"LabelMatchConfidence": "Match Confidence",
|
||||
"LabelMatchExistingUsersBy": "Match existing users by",
|
||||
"LabelMatchExistingUsersByDescription": "Used for connecting existing users. Once connected, users will be matched by a unique id from your SSO provider",
|
||||
"LabelMaxEpisodesToDownload": "Max # of episodes to download. Use 0 for unlimited.",
|
||||
|
@ -7,7 +7,7 @@ const FantLab = require('../providers/FantLab')
|
||||
const AudiobookCovers = require('../providers/AudiobookCovers')
|
||||
const CustomProviderAdapter = require('../providers/CustomProviderAdapter')
|
||||
const Logger = require('../Logger')
|
||||
const { levenshteinDistance, escapeRegExp } = require('../utils/index')
|
||||
const { levenshteinDistance, levenshteinSimilarity, escapeRegExp, isValidASIN } = require('../utils/index')
|
||||
const htmlSanitizer = require('../utils/htmlSanitizer')
|
||||
|
||||
class BookFinder {
|
||||
@ -385,7 +385,11 @@ class BookFinder {
|
||||
|
||||
if (!title) return books
|
||||
|
||||
books = await this.runSearch(title, author, provider, asin, maxTitleDistance, maxAuthorDistance)
|
||||
const isTitleAsin = isValidASIN(title.toUpperCase())
|
||||
|
||||
let actualTitleQuery = title
|
||||
let actualAuthorQuery = author
|
||||
books = await this.runSearch(actualTitleQuery, actualAuthorQuery, provider, asin, maxTitleDistance, maxAuthorDistance)
|
||||
|
||||
if (!books.length && maxFuzzySearches > 0) {
|
||||
// Normalize title and author
|
||||
@ -408,19 +412,26 @@ class BookFinder {
|
||||
for (const titlePart of titleParts) titleCandidates.add(titlePart)
|
||||
titleCandidates = titleCandidates.getCandidates()
|
||||
for (const titleCandidate of titleCandidates) {
|
||||
if (titleCandidate == title && authorCandidate == author) continue // We already tried this
|
||||
if (titleCandidate == actualTitleQuery && authorCandidate == actualAuthorQuery) continue // We already tried this
|
||||
if (++numFuzzySearches > maxFuzzySearches) break loop_author
|
||||
books = await this.runSearch(titleCandidate, authorCandidate, provider, asin, maxTitleDistance, maxAuthorDistance)
|
||||
actualTitleQuery = titleCandidate
|
||||
actualAuthorQuery = authorCandidate
|
||||
books = await this.runSearch(actualTitleQuery, actualAuthorQuery, provider, asin, maxTitleDistance, maxAuthorDistance)
|
||||
if (books.length) break loop_author
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if (books.length) {
|
||||
const resultsHaveDuration = provider.startsWith('audible')
|
||||
if (resultsHaveDuration && libraryItem?.media?.duration) {
|
||||
const libraryItemDurationMinutes = libraryItem.media.duration / 60
|
||||
// If provider results have duration, sort by ascendinge duration difference from libraryItem
|
||||
const isAudibleProvider = provider.startsWith('audible')
|
||||
const libraryItemDurationMinutes = libraryItem?.media?.duration ? libraryItem.media.duration / 60 : null
|
||||
|
||||
books.forEach((book) => {
|
||||
if (typeof book !== 'object' || !isAudibleProvider) return
|
||||
book.matchConfidence = this.calculateMatchConfidence(book, libraryItemDurationMinutes, actualTitleQuery, actualAuthorQuery, isTitleAsin)
|
||||
})
|
||||
|
||||
if (isAudibleProvider && libraryItemDurationMinutes) {
|
||||
books.sort((a, b) => {
|
||||
const aDuration = a.duration || Number.POSITIVE_INFINITY
|
||||
const bDuration = b.duration || Number.POSITIVE_INFINITY
|
||||
@ -433,6 +444,120 @@ class BookFinder {
|
||||
return books
|
||||
}
|
||||
|
||||
/**
|
||||
* Calculate match confidence score for a book
|
||||
* @param {Object} book - The book object to calculate confidence for
|
||||
* @param {number|null} libraryItemDurationMinutes - Duration of library item in minutes
|
||||
* @param {string} actualTitleQuery - Actual title query
|
||||
* @param {string} actualAuthorQuery - Actual author query
|
||||
* @param {boolean} isTitleAsin - Whether the title is an ASIN
|
||||
* @returns {number|null} - Match confidence score or null if not applicable
|
||||
*/
|
||||
calculateMatchConfidence(book, libraryItemDurationMinutes, actualTitleQuery, actualAuthorQuery, isTitleAsin) {
|
||||
// ASIN results are always a match
|
||||
if (isTitleAsin) return 1.0
|
||||
|
||||
let durationScore
|
||||
if (libraryItemDurationMinutes && typeof book.duration === 'number') {
|
||||
const durationDiff = Math.abs(book.duration - libraryItemDurationMinutes)
|
||||
// Duration scores:
|
||||
// diff | score
|
||||
// 0 | 1.0
|
||||
// 1 | 1.0
|
||||
// 2 | 0.9
|
||||
// 3 | 0.8
|
||||
// 4 | 0.7
|
||||
// 5 | 0.6
|
||||
// 6 | 0.48
|
||||
// 7 | 0.36
|
||||
// 8 | 0.24
|
||||
// 9 | 0.12
|
||||
// 10 | 0.0
|
||||
if (durationDiff <= 1) {
|
||||
// Covers durationDiff = 0 for score 1.0
|
||||
durationScore = 1.0
|
||||
} else if (durationDiff <= 5) {
|
||||
// (1, 5] - Score from 1.0 down to 0.6
|
||||
// Linearly interpolates between (1, 1.0) and (5, 0.6)
|
||||
// Equation: y = 1.0 - 0.08 * x
|
||||
durationScore = 1.1 - 0.1 * durationDiff
|
||||
} else if (durationDiff <= 10) {
|
||||
// (5, 10] - Score from 0.6 down to 0.0
|
||||
// Linearly interpolates between (5, 0.6) and (10, 0.0)
|
||||
// Equation: y = 1.2 - 0.12 * x
|
||||
durationScore = 1.2 - 0.12 * durationDiff
|
||||
} else {
|
||||
// durationDiff > 10 - Score is 0.0
|
||||
durationScore = 0.0
|
||||
}
|
||||
Logger.debug(`[BookFinder] Duration diff: ${durationDiff}, durationScore: ${durationScore}`)
|
||||
} else {
|
||||
// Default score if library item duration or book duration is not available
|
||||
durationScore = 0.1
|
||||
}
|
||||
|
||||
const calculateTitleScore = (titleQuery, book, keepSubtitle = false) => {
|
||||
const cleanTitle = cleanTitleForCompares(book.title || '', keepSubtitle)
|
||||
const cleanSubtitle = keepSubtitle && book.subtitle ? `: ${book.subtitle}` : ''
|
||||
const normBookTitle = `${cleanTitle}${cleanSubtitle}`
|
||||
const normTitleQuery = cleanTitleForCompares(titleQuery, keepSubtitle)
|
||||
const titleSimilarity = levenshteinSimilarity(normTitleQuery, normBookTitle)
|
||||
Logger.debug(`[BookFinder] keepSubtitle: ${keepSubtitle}, normBookTitle: ${normBookTitle}, normTitleQuery: ${normTitleQuery}, titleSimilarity: ${titleSimilarity}`)
|
||||
return titleSimilarity
|
||||
}
|
||||
const titleQueryHasSubtitle = hasSubtitle(actualTitleQuery)
|
||||
const titleScore = calculateTitleScore(actualTitleQuery, book, titleQueryHasSubtitle)
|
||||
|
||||
let authorScore
|
||||
const normAuthorQuery = cleanAuthorForCompares(actualAuthorQuery)
|
||||
const normBookAuthor = cleanAuthorForCompares(book.author || '')
|
||||
if (!normAuthorQuery) {
|
||||
// Original query had no author
|
||||
authorScore = 1.0 // Neutral score
|
||||
} else {
|
||||
// Original query HAS an author (cleanedQueryAuthorForScore is not empty)
|
||||
if (normBookAuthor) {
|
||||
const bookAuthorParts = normBookAuthor.split(',').map((name) => name.trim().toLowerCase())
|
||||
// Filter out empty parts that might result from ", ," or trailing/leading commas
|
||||
const validBookAuthorParts = bookAuthorParts.filter((p) => p.length > 0)
|
||||
|
||||
if (validBookAuthorParts.length === 0) {
|
||||
// Book author string was present but effectively empty (e.g. ",,")
|
||||
// Since cleanedQueryAuthorForScore is non-empty here, this is a mismatch.
|
||||
authorScore = 0.0
|
||||
} else {
|
||||
let maxPartScore = levenshteinSimilarity(normAuthorQuery, normBookAuthor)
|
||||
Logger.debug(`[BookFinder] normAuthorQuery: ${normAuthorQuery}, normBookAuthor: ${normBookAuthor}, similarity: ${maxPartScore}`)
|
||||
if (validBookAuthorParts.length > 1 || normBookAuthor.includes(',')) {
|
||||
validBookAuthorParts.forEach((part) => {
|
||||
// part is guaranteed to be non-empty here
|
||||
// cleanedQueryAuthorForScore is also guaranteed non-empty here.
|
||||
// levenshteinDistance lowercases by default, but part is already lowercased.
|
||||
const similarity = levenshteinSimilarity(normAuthorQuery, part)
|
||||
Logger.debug(`[BookFinder] normAuthorQuery: ${normAuthorQuery}, bookAuthorPart: ${part}, similarity: ${similarity}`)
|
||||
const currentPartScore = similarity
|
||||
maxPartScore = Math.max(maxPartScore, currentPartScore)
|
||||
})
|
||||
}
|
||||
authorScore = maxPartScore
|
||||
}
|
||||
} else {
|
||||
// Book has NO author (or not a string, or empty string)
|
||||
// Query has an author (cleanedQueryAuthorForScore is non-empty), book does not.
|
||||
authorScore = 0.0
|
||||
}
|
||||
}
|
||||
|
||||
const W_DURATION = 0.7
|
||||
const W_TITLE = 0.2
|
||||
const W_AUTHOR = 0.1
|
||||
|
||||
Logger.debug(`[BookFinder] Duration score: ${durationScore}, Title score: ${titleScore}, Author score: ${authorScore}`)
|
||||
const confidence = W_DURATION * durationScore + W_TITLE * titleScore + W_AUTHOR * authorScore
|
||||
Logger.debug(`[BookFinder] Confidence: ${confidence}`)
|
||||
return Math.max(0, Math.min(1, confidence))
|
||||
}
|
||||
|
||||
/**
|
||||
* Search for books
|
||||
*
|
||||
@ -464,6 +589,7 @@ class BookFinder {
|
||||
} else {
|
||||
books = await this.getGoogleBooksResults(title, author)
|
||||
}
|
||||
|
||||
books.forEach((book) => {
|
||||
if (book.description) {
|
||||
book.description = htmlSanitizer.sanitize(book.description)
|
||||
@ -505,6 +631,9 @@ class BookFinder {
|
||||
}
|
||||
module.exports = new BookFinder()
|
||||
|
||||
function hasSubtitle(title) {
|
||||
return title.includes(':') || title.includes(' - ')
|
||||
}
|
||||
function stripSubtitle(title) {
|
||||
if (title.includes(':')) {
|
||||
return title.split(':')[0].trim()
|
||||
@ -523,12 +652,12 @@ function replaceAccentedChars(str) {
|
||||
}
|
||||
}
|
||||
|
||||
function cleanTitleForCompares(title) {
|
||||
function cleanTitleForCompares(title, keepSubtitle = false) {
|
||||
if (!title) return ''
|
||||
title = stripRedundantSpaces(title)
|
||||
|
||||
// Remove subtitle if there (i.e. "Cool Book: Coolest Ever" becomes "Cool Book")
|
||||
let stripped = stripSubtitle(title)
|
||||
let stripped = keepSubtitle ? title : stripSubtitle(title)
|
||||
|
||||
// Remove text in paranthesis (i.e. "Ender's Game (Ender's Saga)" becomes "Ender's Game")
|
||||
let cleaned = stripped.replace(/ *\([^)]*\) */g, '')
|
||||
|
@ -34,6 +34,14 @@ const levenshteinDistance = (str1, str2, caseSensitive = false) => {
|
||||
}
|
||||
module.exports.levenshteinDistance = levenshteinDistance
|
||||
|
||||
const levenshteinSimilarity = (str1, str2, caseSensitive = false) => {
|
||||
const distance = levenshteinDistance(str1, str2, caseSensitive)
|
||||
const maxLength = Math.max(str1.length, str2.length)
|
||||
if (maxLength === 0) return 1
|
||||
return 1 - distance / maxLength
|
||||
}
|
||||
module.exports.levenshteinSimilarity = levenshteinSimilarity
|
||||
|
||||
module.exports.isObject = (val) => {
|
||||
return val !== null && typeof val === 'object'
|
||||
}
|
||||
|
@ -5,6 +5,12 @@ const bookFinder = require('../../../server/finders/BookFinder')
|
||||
const { LogLevel } = require('../../../server/utils/constants')
|
||||
const Logger = require('../../../server/Logger')
|
||||
Logger.setLogLevel(LogLevel.INFO)
|
||||
const { levenshteinDistance } = require('../../../server/utils/index')
|
||||
|
||||
// levenshteinDistance is needed for manual calculation of expected scores in tests.
|
||||
// Assuming it's accessible for testing purposes or we mock/replicate its basic behavior if needed.
|
||||
// For now, we'll assume bookFinder.search uses it internally correctly.
|
||||
// const { levenshteinDistance } = require('../../../server/utils/index') // Not used directly in test logic, but for reasoning.
|
||||
|
||||
describe('TitleCandidates', () => {
|
||||
describe('cleanAuthor non-empty', () => {
|
||||
@ -326,31 +332,262 @@ describe('search', () => {
|
||||
const sorted = [{ duration: 1000 }, { duration: 500 }, { duration: 2000 }, { duration: 3000 }]
|
||||
|
||||
beforeEach(() => {
|
||||
runSearchStub.withArgs(t, a, provider).resolves(unsorted)
|
||||
runSearchStub.withArgs(t, a, provider).resolves(structuredClone(unsorted))
|
||||
})
|
||||
|
||||
afterEach(() => {
|
||||
sinon.restore()
|
||||
})
|
||||
|
||||
it('returns results sorted by library item duration diff', async () => {
|
||||
expect(await bookFinder.search(libraryItem, provider, t, a)).to.deep.equal(sorted)
|
||||
const result = (await bookFinder.search(libraryItem, provider, t, a)).map((r) => (r.duration ? { duration: r.duration } : {}))
|
||||
expect(result).to.deep.equal(sorted)
|
||||
})
|
||||
|
||||
it('returns unsorted results if library item is null', async () => {
|
||||
expect(await bookFinder.search(null, provider, t, a)).to.deep.equal(unsorted)
|
||||
const result = (await bookFinder.search(null, provider, t, a)).map((r) => (r.duration ? { duration: r.duration } : {}))
|
||||
expect(result).to.deep.equal(unsorted)
|
||||
})
|
||||
|
||||
it('returns unsorted results if library item duration is undefined', async () => {
|
||||
expect(await bookFinder.search({ media: {} }, provider, t, a)).to.deep.equal(unsorted)
|
||||
const result = (await bookFinder.search({ media: {} }, provider, t, a)).map((r) => (r.duration ? { duration: r.duration } : {}))
|
||||
expect(result).to.deep.equal(unsorted)
|
||||
})
|
||||
|
||||
it('returns unsorted results if library item media is undefined', async () => {
|
||||
expect(await bookFinder.search({}, provider, t, a)).to.deep.equal(unsorted)
|
||||
const result = (await bookFinder.search({}, provider, t, a)).map((r) => (r.duration ? { duration: r.duration } : {}))
|
||||
expect(result).to.deep.equal(unsorted)
|
||||
})
|
||||
|
||||
it('should return a result last if it has no duration', async () => {
|
||||
const unsorted = [{}, { duration: 3000 }, { duration: 2000 }, { duration: 1000 }, { duration: 500 }]
|
||||
const sorted = [{ duration: 1000 }, { duration: 500 }, { duration: 2000 }, { duration: 3000 }, {}]
|
||||
runSearchStub.withArgs(t, a, provider).resolves(unsorted)
|
||||
runSearchStub.withArgs(t, a, provider).resolves(structuredClone(unsorted))
|
||||
const result = (await bookFinder.search(libraryItem, provider, t, a)).map((r) => (r.duration ? { duration: r.duration } : {}))
|
||||
expect(result).to.deep.equal(sorted)
|
||||
})
|
||||
})
|
||||
|
||||
expect(await bookFinder.search(libraryItem, provider, t, a)).to.deep.equal(sorted)
|
||||
describe('matchConfidence score', () => {
|
||||
const W_DURATION = 0.7
|
||||
const W_TITLE = 0.2
|
||||
const W_AUTHOR = 0.1
|
||||
const DEFAULT_DURATION_SCORE_MISSING_INFO = 0.1
|
||||
|
||||
const libraryItemPerfectDuration = { media: { duration: 600 } } // 10 minutes
|
||||
|
||||
// Helper to calculate expected title/author score based on Levenshtein
|
||||
// Assumes queryPart and bookPart are already "cleaned" for length calculation consistency with BookFinder.js
|
||||
const calculateStringMatchScore = (cleanedQueryPart, cleanedBookPart) => {
|
||||
if (!cleanedQueryPart) return cleanedBookPart ? 0 : 1 // query empty: 1 if book empty, else 0
|
||||
if (!cleanedBookPart) return 0 // query non-empty, book empty: 0
|
||||
|
||||
// Use the imported levenshteinDistance. It defaults to case-insensitive, which is what we want.
|
||||
const distance = levenshteinDistance(cleanedQueryPart, cleanedBookPart)
|
||||
return Math.max(0, 1 - distance / Math.max(cleanedQueryPart.length, cleanedBookPart.length))
|
||||
}
|
||||
|
||||
beforeEach(() => {
|
||||
runSearchStub.resolves([])
|
||||
})
|
||||
|
||||
afterEach(() => {
|
||||
sinon.restore()
|
||||
})
|
||||
|
||||
describe('for audible provider', () => {
|
||||
const provider = 'audible'
|
||||
|
||||
it('should be 1.0 for perfect duration, title, and author match', async () => {
|
||||
const bookResults = [{ duration: 10, title: 'The Great Novel', author: 'John Doe' }]
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', 'John Doe')
|
||||
// durationScore = 1.0 (diff 0 <= 1 min)
|
||||
// titleScore = 1.0 (exact match)
|
||||
// authorScore = 1.0 (exact match)
|
||||
const expectedConfidence = W_DURATION * 1.0 + W_TITLE * 1.0 + W_AUTHOR * 1.0
|
||||
expect(results[0].matchConfidence).to.be.closeTo(expectedConfidence, 0.001)
|
||||
})
|
||||
|
||||
it('should correctly score a large duration mismatch', async () => {
|
||||
const bookResults = [{ duration: 21, title: 'The Great Novel', author: 'John Doe' }] // 21 min, diff = 11 min
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', 'John Doe')
|
||||
// durationScore = 0.0
|
||||
// titleScore = 1.0
|
||||
// authorScore = 1.0
|
||||
const expectedConfidence = W_DURATION * 0.0 + W_TITLE * 1.0 + W_AUTHOR * 1.0
|
||||
expect(results[0].matchConfidence).to.be.closeTo(expectedConfidence, 0.001)
|
||||
})
|
||||
|
||||
it('should correctly score a medium duration mismatch', async () => {
|
||||
const bookResults = [{ duration: 16, title: 'The Great Novel', author: 'John Doe' }] // 16 min, diff = 6 min
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', 'John Doe')
|
||||
// durationScore = 1.2 - 6 * 0.12 = 0.48
|
||||
// titleScore = 1.0
|
||||
// authorScore = 1.0
|
||||
const expectedConfidence = W_DURATION * 0.48 + W_TITLE * 1.0 + W_AUTHOR * 1.0
|
||||
expect(results[0].matchConfidence).to.be.closeTo(expectedConfidence, 0.001)
|
||||
})
|
||||
|
||||
it('should correctly score a minor duration mismatch', async () => {
|
||||
const bookResults = [{ duration: 14, title: 'The Great Novel', author: 'John Doe' }] // 14 min, diff = 4 min
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', 'John Doe')
|
||||
// durationScore = 1.1 - 4 * 0.1 = 0.7
|
||||
// titleScore = 1.0
|
||||
// authorScore = 1.0
|
||||
const expectedConfidence = W_DURATION * 0.7 + W_TITLE * 1.0 + W_AUTHOR * 1.0
|
||||
expect(results[0].matchConfidence).to.be.closeTo(expectedConfidence, 0.001)
|
||||
})
|
||||
|
||||
it('should correctly score a tiny duration mismatch', async () => {
|
||||
const bookResults = [{ duration: 11, title: 'The Great Novel', author: 'John Doe' }] // 11 min, diff = 1 min
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', 'John Doe')
|
||||
// durationScore = 1.0
|
||||
// titleScore = 1.0
|
||||
// authorScore = 1.0
|
||||
const expectedConfidence = W_DURATION * 1.0 + W_TITLE * 1.0 + W_AUTHOR * 1.0
|
||||
expect(results[0].matchConfidence).to.be.closeTo(expectedConfidence, 0.001)
|
||||
})
|
||||
|
||||
it('should use default duration score if libraryItem duration is missing', async () => {
|
||||
const bookResults = [{ duration: 10, title: 'The Great Novel', author: 'John Doe' }]
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search({ media: {} }, provider, 'The Great Novel', 'John Doe')
|
||||
// durationScore = DEFAULT_DURATION_SCORE_MISSING_INFO (0.2)
|
||||
const expectedConfidence = W_DURATION * DEFAULT_DURATION_SCORE_MISSING_INFO + W_TITLE * 1.0 + W_AUTHOR * 1.0
|
||||
expect(results[0].matchConfidence).to.be.closeTo(expectedConfidence, 0.001)
|
||||
})
|
||||
|
||||
it('should use default duration score if book duration is missing', async () => {
|
||||
const bookResults = [{ title: 'The Great Novel', author: 'John Doe' }] // No duration in book
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', 'John Doe')
|
||||
// durationScore = DEFAULT_DURATION_SCORE_MISSING_INFO (0.2)
|
||||
const expectedConfidence = W_DURATION * DEFAULT_DURATION_SCORE_MISSING_INFO + W_TITLE * 1.0 + W_AUTHOR * 1.0
|
||||
expect(results[0].matchConfidence).to.be.closeTo(expectedConfidence, 0.001)
|
||||
})
|
||||
|
||||
it('should correctly score a partial title match', async () => {
|
||||
const bookResults = [{ duration: 10, title: 'Novel', author: 'John Doe' }]
|
||||
runSearchStub.resolves(bookResults)
|
||||
// Query: 'Novel Ex', Book: 'Novel'
|
||||
// cleanTitleForCompares('Novel Ex') -> 'novel ex' (length 8)
|
||||
// cleanTitleForCompares('Novel') -> 'novel' (length 5)
|
||||
// levenshteinDistance('novel ex', 'novel') = 3
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'Novel Ex', 'John Doe')
|
||||
const expectedTitleScore = calculateStringMatchScore('novel ex', 'novel') // 1 - (3/8) = 0.625
|
||||
const expectedConfidence = W_DURATION * 1.0 + W_TITLE * expectedTitleScore + W_AUTHOR * 1.0
|
||||
expect(results[0].matchConfidence).to.be.closeTo(expectedConfidence, 0.001)
|
||||
})
|
||||
|
||||
it('should correctly score a partial author match (comma-separated)', async () => {
|
||||
const bookResults = [{ duration: 10, title: 'The Great Novel', author: 'Jane Smith, Jon Doee' }]
|
||||
runSearchStub.resolves(bookResults)
|
||||
// Query: 'Jon Doe', Book part: 'Jon Doee'
|
||||
// cleanAuthorForCompares('Jon Doe') -> 'jon doe' (length 7)
|
||||
// book author part (already lowercased) -> 'jon doee' (length 8)
|
||||
// levenshteinDistance('jon doe', 'jon doee') = 1
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', 'Jon Doe')
|
||||
// For the author part 'jon doee':
|
||||
const expectedAuthorPartScore = calculateStringMatchScore('jon doe', 'jon doee') // 1 - (1/7)
|
||||
// Assuming 'jane smith' gives a lower or 0 score, max score will be from 'jon doee'
|
||||
const expectedConfidence = W_DURATION * 1.0 + W_TITLE * 1.0 + W_AUTHOR * expectedAuthorPartScore
|
||||
expect(results[0].matchConfidence).to.be.closeTo(expectedConfidence, 0.001)
|
||||
})
|
||||
|
||||
it('should give authorScore 0 if query has author but book does not', async () => {
|
||||
const bookResults = [{ duration: 10, title: 'The Great Novel', author: null }]
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', 'John Doe')
|
||||
// authorScore = 0.0
|
||||
const expectedConfidence = W_DURATION * 1.0 + W_TITLE * 1.0 + W_AUTHOR * 0.0
|
||||
expect(results[0].matchConfidence).to.be.closeTo(expectedConfidence, 0.001)
|
||||
})
|
||||
|
||||
it('should give authorScore 1.0 if query has no author', async () => {
|
||||
const bookResults = [{ duration: 10, title: 'The Great Novel', author: 'John Doe' }]
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', '') // Empty author
|
||||
expect(results[0].matchConfidence).to.be.closeTo(1.0, 0.001)
|
||||
})
|
||||
|
||||
it('handles book author string that is only commas correctly (score 0)', async () => {
|
||||
const bookResults = [{ duration: 10, title: 'The Great Novel', author: ',, ,, ,' }]
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', 'John Doe')
|
||||
// cleanedQueryAuthorForScore = "john doe"
|
||||
// book.author leads to validBookAuthorParts being empty.
|
||||
// authorScore = 0.0
|
||||
const expectedConfidence = W_DURATION * 1.0 + W_TITLE * 1.0 + W_AUTHOR * 0.0
|
||||
expect(results[0].matchConfidence).to.be.closeTo(expectedConfidence, 0.001)
|
||||
})
|
||||
|
||||
it('should return 1.0 for ASIN results', async () => {
|
||||
const bookResults = [{ duration: 10, title: 'The Great Novel', author: 'John Doe' }]
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'B000F28ZJ4', null)
|
||||
expect(results[0].matchConfidence).to.be.closeTo(1.0, 0.001)
|
||||
})
|
||||
|
||||
it('should return 1.0 when author matches one of the book authors', async () => {
|
||||
const bookResults = [{ duration: 10, title: 'The Great Novel', author: 'John Doe, Jane Smith' }]
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', 'John Doe')
|
||||
expect(results[0].matchConfidence).to.be.closeTo(1.0, 0.001)
|
||||
})
|
||||
|
||||
it('should return 1.0 when author query and multiple book authors are the same', async () => {
|
||||
const bookResults = [{ duration: 10, title: 'The Great Novel', author: 'John Doe, Jane Smith' }]
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', 'John Doe, Jane Smith')
|
||||
expect(results[0].matchConfidence).to.be.closeTo(1.0, 0.001)
|
||||
})
|
||||
|
||||
it('should correctly score against a book with a subtitle when the query has a subtitle', async () => {
|
||||
const bookResults = [{ duration: 10, title: 'The Great Novel', subtitle: 'A Novel', author: 'John Doe' }]
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel: A Novel', 'John Doe')
|
||||
expect(results[0].matchConfidence).to.be.closeTo(1.0, 0.001)
|
||||
})
|
||||
|
||||
it('should correctly score against a book with a subtitle when the query does not have a subtitle', async () => {
|
||||
const bookResults = [{ duration: 10, title: 'The Great Novel', subtitle: 'A Novel', author: 'John Doe' }]
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', 'John Doe')
|
||||
expect(results[0].matchConfidence).to.be.closeTo(1.0, 0.001)
|
||||
})
|
||||
|
||||
describe('after fuzzy searches', () => {
|
||||
it('should return 1.0 for a title candidate match', async () => {
|
||||
const bookResults = [{ duration: 10, title: 'The Great Novel', author: 'John Doe' }]
|
||||
runSearchStub.resolves([])
|
||||
runSearchStub.withArgs('the great novel', 'john doe').resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel - A Novel', 'John Doe')
|
||||
expect(results[0].matchConfidence).to.be.closeTo(1.0, 0.001)
|
||||
})
|
||||
|
||||
it('should return 1.0 for an author candidate match', async () => {
|
||||
const bookResults = [{ duration: 10, title: 'The Great Novel', author: 'John Doe' }]
|
||||
runSearchStub.resolves([])
|
||||
runSearchStub.withArgs('the great novel', 'john doe').resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', 'John Doe, Jane Smith')
|
||||
expect(results[0].matchConfidence).to.be.closeTo(1.0, 0.001)
|
||||
})
|
||||
})
|
||||
})
|
||||
|
||||
describe('for non-audible provider (e.g., google)', () => {
|
||||
const provider = 'google'
|
||||
it('should have not have matchConfidence', async () => {
|
||||
const bookResults = [{ title: 'The Great Novel', author: 'John Doe' }]
|
||||
runSearchStub.resolves(bookResults)
|
||||
const results = await bookFinder.search(libraryItemPerfectDuration, provider, 'The Great Novel', 'John Doe')
|
||||
expect(results[0]).to.not.have.property('matchConfidence')
|
||||
})
|
||||
})
|
||||
})
|
||||
})
|
||||
|
Loading…
Reference in New Issue
Block a user