Mirrored_Repos/yattee
1
0
Fork 0
mirror of https://github.com/yattee/yattee.git synced 2024-11-09 15:58:20 +00:00
Code Issues Packages Projects Releases Wiki Activity

more robust regex for chapters from description

Browse Source 
This should reduce the number of falsely matched chapters, e.g., when time-code-like numbers appear in the middle of the text, like 16:9 or sports results.

It also checks for chapters that have an end time and omits the end time code from the title.

Track lists in music videos are now also properly displayed as chapters.
This commit is contained in:
Toni Förster 2023-12-27 22:36:52 +01:00
parent 39f6319043
commit ecba91f35d
No known key found for this signature in database
GPG Key ID: 292F3E5086C83FC7
1 changed files with 50 additions and 39 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

89
Model/Applications/VideosAPI.swift
View File

@ -144,52 +144,63 @@ extension VideosAPI {
}
func extractChapters(from description: String) -> [Chapter] {
guard let chaptersRegularExpression = try? NSRegularExpression(
pattern: "(?<start>(?:[0-9]+:){1,}(?:[0-9]+))(?:\\s)+(?:- ?)?(?<title>.*)",
options: .caseInsensitive
) else { return [] }
/*
The following chapter patterns are covered:
let chapterLines = chaptersRegularExpression.matches(
in: description,
range: NSRange(description.startIndex..., in: description)
)
start - end - title / start - end: Title / start - end title
start - title / start: title / start title / [start] - title / [start]: title / [start] title
index. title - start / index. title start
title: (start)
return chapterLines.compactMap { line in
let titleRange = line.range(withName: "title")
let startRange = line.range(withName: "start")
The order is important!
*/
let patterns = [
"(?<=\\n|^)\\s*(?:►\\s*)?\\[?(?<start>(?:[0-9]+:){1,2}[0-9]+)\\]?(?:\\s*-\\s*)?(?<end>(?:[0-9]+:){1,2}[0-9]+)?(?:\\s*-\\s*|\\s*[:]\\s*)?(?<title>.*)(?=\\n|$)",
"(?<=\\n|^)\\s*(?:►\\s*)?\\[?(?<start>(?:[0-9]+:){1,2}[0-9]+)\\]?\\s*[-:]?\\s*(?<title>.+)(?=\\n|$)",
"(?<=\\n|^)(?<index>[0-9]+\\.\\s)(?<title>.+?)(?:\\s*-\\s*)?(?<start>(?:[0-9]+:){1,2}[0-9]+)(?=\\n|$)",
"(?<=\\n|^)(?<title>.+?):\\s*\\((?<start>(?:[0-9]+:){1,2}[0-9]+)\\)(?=\\n|$)"
]
guard let titleSubstringRange = Range(titleRange, in: description),
let startSubstringRange = Range(startRange, in: description) else { return nil }
for pattern in patterns {
guard let chaptersRegularExpression = try? NSRegularExpression(pattern: pattern, options: .caseInsensitive) else { continue }
let chapterLines = chaptersRegularExpression.matches(in: description, range: NSRange(description.startIndex..., in: description))
let titleCapture = String(description[titleSubstringRange])
let startCapture = String(description[startSubstringRange])
let startComponents = startCapture.components(separatedBy: ":")
guard startComponents.count <= 3 else { return nil }
if !chapterLines.isEmpty {
return chapterLines.compactMap { line in
let titleRange = line.range(withName: "title")
let startRange = line.range(withName: "start")
guard let titleSubstringRange = Range(titleRange, in: description),
let startSubstringRange = Range(startRange, in: description)
else {
return nil
}
let titleCapture = String(description[titleSubstringRange]).trimmingCharacters(in: .whitespaces)
let startCapture = String(description[startSubstringRange])
let startComponents = startCapture.components(separatedBy: ":")
guard startComponents.count <= 3 else { return nil }
var hours: Double?
var minutes: Double?
var seconds: Double?
var hours: Double?
var minutes: Double?
var seconds: Double?
if startComponents.count == 3 {
hours = Double(startComponents[0])
minutes = Double(startComponents[1])
seconds = Double(startComponents[2])
} else if startComponents.count == 2 {
minutes = Double(startComponents[0])
seconds = Double(startComponents[1])
if startComponents.count == 3 {
hours = Double(startComponents[0])
minutes = Double(startComponents[1])
seconds = Double(startComponents[2])
} else if startComponents.count == 2 {
minutes = Double(startComponents[0])
seconds = Double(startComponents[1])
}
guard var startSeconds = seconds else { return nil }
startSeconds += (minutes ?? 0) * 60
startSeconds += (hours ?? 0) * 60 * 60
return .init(title: titleCapture, start: startSeconds)
}
}
guard var startSeconds = seconds else { return nil }
if let minutes {
startSeconds += 60 * minutes
}
if let hours {
startSeconds += 60 * 60 * hours
}
return .init(title: titleCapture, start: startSeconds)
}
return []
}
}