Add fingerprint cache

2026-07-27 15:39:22 +00:00
parent 246198f2eb
commit f64f4cea36
1 changed files with 118 additions and 0 deletions
@@ -0,0 +1,118 @@
+import Foundation
+import Vision
+
+/// 1ファイル分のキャッシュ。ファイルサイズと更新日時が一致するときだけ再利用する。
+struct FingerprintRecord: Codable, Sendable {
+    var fileSize: Int64
+    var modificationDate: Date
+    var basics: ImageBasics?
+    var featurePrint: FeaturePrintObservation?
+
+    func matches(_ file: ScannedFile) -> Bool {
+        fileSize == file.fileSize
+            && abs(modificationDate.timeIntervalSince(file.modificationDate)) < 1
+    }
+}
+
+/// dHash と feature print をディスクに保存し、同じフォルダの再スキャンを速くする。
+///
+/// feature print は1枚あたり数 KB あるので、JSON ではなくバイナリ plist で持つ。
+actor FingerprintCache {
+    private var records: [String: FingerprintRecord] = [:]
+    private var isDirty = false
+    private let fileURL: URL?
+
+    init(fileURL: URL? = FingerprintCache.defaultFileURL()) {
+        self.fileURL = fileURL
+        self.records = Self.loadRecords(from: fileURL)
+    }
+
+    static func defaultFileURL() -> URL? {
+        guard let caches = FileManager.default.urls(for: .cachesDirectory, in: .userDomainMask).first
+        else { return nil }
+        let directory = caches.appendingPathComponent("Photochk", isDirectory: true)
+        try? FileManager.default.createDirectory(at: directory, withIntermediateDirectories: true)
+        return directory.appendingPathComponent("fingerprints.plist")
+    }
+
+    // MARK: - 読み出し
+
+    func basics(for files: [ScannedFile]) -> [URL: ImageBasics] {
+        var found: [URL: ImageBasics] = [:]
+        for file in files {
+            if let record = records[file.url.path], record.matches(file), let basics = record.basics {
+                found[file.url] = basics
+            }
+        }
+        return found
+    }
+
+    func featurePrints(for files: [ScannedFile]) -> [URL: FeaturePrintObservation] {
+        var found: [URL: FeaturePrintObservation] = [:]
+        for file in files {
+            if let record = records[file.url.path], record.matches(file),
+               let print = record.featurePrint {
+                found[file.url] = print
+            }
+        }
+        return found
+    }
+
+    // MARK: - 書き込み
+
+    func store(basics: [URL: ImageBasics], for files: [ScannedFile]) {
+        for file in files {
+            guard let value = basics[file.url] else { continue }
+            update(file) { $0.basics = value }
+        }
+    }
+
+    func store(featurePrints: [URL: FeaturePrintObservation], for files: [ScannedFile]) {
+        for file in files {
+            guard let value = featurePrints[file.url] else { continue }
+            update(file) { $0.featurePrint = value }
+        }
+    }
+
+    private func update(_ file: ScannedFile, _ body: (inout FingerprintRecord) -> Void) {
+        var record = records[file.url.path] ?? FingerprintRecord(
+            fileSize: file.fileSize,
+            modificationDate: file.modificationDate
+        )
+        if !record.matches(file) {
+            // ファイルが変わっていたら作り直す
+            record = FingerprintRecord(
+                fileSize: file.fileSize,
+                modificationDate: file.modificationDate
+            )
+        }
+        body(&record)
+        records[file.url.path] = record
+        isDirty = true
+    }
+
+    /// 消えたファイルのエントリを捨てる。スキャンのたびに肥大化するのを防ぐ。
+    func prune(keeping files: [ScannedFile]) {
+        let alive = Set(files.map(\.url.path))
+        let before = records.count
+        records = records.filter { alive.contains($0.key) }
+        if records.count != before { isDirty = true }
+    }
+
+    // MARK: - 永続化
+
+    func save() {
+        guard isDirty, let fileURL else { return }
+        let encoder = PropertyListEncoder()
+        encoder.outputFormat = .binary
+        guard let data = try? encoder.encode(records) else { return }
+        try? data.write(to: fileURL, options: .atomic)
+        isDirty = false
+    }
+
+    private static func loadRecords(from fileURL: URL?) -> [String: FingerprintRecord] {
+        guard let fileURL, let data = try? Data(contentsOf: fileURL) else { return [:] }
+        // 形式が変わっていたら黙って捨てる。キャッシュなので作り直せばよい。
+        return (try? PropertyListDecoder().decode([String: FingerprintRecord].self, from: data)) ?? [:]
+    }
+}