jake/vOOice
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix OCR capturing VoiceInk status overlay instead of frontmost app window

Browse Source 
The screen capture service was selecting the first layer-0 window, which
during recording was VoiceInk's own status indicator overlay. This caused
OCR to always return 'No text detected' since the overlay has no readable
content.

Changes:
- Filter out windows owned by VoiceInk's process
- Prioritize windows belonging to NSWorkspace.frontmostApplication
- Refactor WindowCandidate struct to class scope
This commit is contained in:
Nico Duldhardt 2025-12-07 00:52:52 +01:00
parent df2a8d9901
commit d25ae52333
1 changed files with 23 additions and 38 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

61
VoiceInk/Services/ScreenCaptureService.swift
View File

@ -20,47 +20,32 @@ class ScreenCaptureService: ObservableObject {
let windowID: CGWindowID
let ownerPID: pid_t
let layer: Int32
let bounds: CGRect
}
private func getActiveWindowInfo() async -> (title: String, ownerName: String, windowID: CGWindowID)? {
private func getActiveWindowInfo() -> (title: String, ownerName: String, windowID: CGWindowID)? {
let currentPID = ProcessInfo.processInfo.processIdentifier
let frontmostPID = await NSWorkspace.shared.frontmostApplication?.processIdentifier
let frontmostPID = NSWorkspace.shared.frontmostApplication?.processIdentifier
let windowListInfo = CGWindowListCopyWindowInfo([.optionOnScreenOnly], kCGNullWindowID) as? [[String: Any]] ?? []
// Move expensive window list retrieval off the main thread
let candidates = await Task.detached(priority: .userInitiated) { () -> [WindowCandidate] in
let windowListInfo = CGWindowListCopyWindowInfo([.optionOnScreenOnly], kCGNullWindowID) as? [[String: Any]] ?? []
return windowListInfo.compactMap { info -> WindowCandidate? in
guard let windowID = info[kCGWindowNumber as String] as? CGWindowID,
let ownerName = info[kCGWindowOwnerName as String] as? String,
let ownerPIDNumber = info[kCGWindowOwnerPID as String] as? NSNumber,
let layer = info[kCGWindowLayer as String] as? Int32,
let boundsDict = info[kCGWindowBounds as String] as? [String: Any],
let width = boundsDict["Width"] as? CGFloat,
let height = boundsDict["Height"] as? CGFloat else {
return nil
}
let rawTitle = (info[kCGWindowName as String] as? String)?.trimmingCharacters(in: .whitespacesAndNewlines)
let resolvedTitle = rawTitle?.isEmpty == false ? rawTitle! : ownerName
let bounds = CGRect(
x: boundsDict["X"] as? CGFloat ?? 0,
y: boundsDict["Y"] as? CGFloat ?? 0,
width: width,
height: height
)
return WindowCandidate(
title: resolvedTitle,
ownerName: ownerName,
windowID: windowID,
ownerPID: ownerPIDNumber.int32Value,
layer: layer,
bounds: bounds
)
let candidates = windowListInfo.compactMap { info -> WindowCandidate? in
guard let windowID = info[kCGWindowNumber as String] as? CGWindowID,
let ownerName = info[kCGWindowOwnerName as String] as? String,
let ownerPIDNumber = info[kCGWindowOwnerPID as String] as? NSNumber,
let layer = info[kCGWindowLayer as String] as? Int32 else {
return nil
}
}.value
let rawTitle = (info[kCGWindowName as String] as? String)?.trimmingCharacters(in: .whitespacesAndNewlines)
let resolvedTitle = rawTitle?.isEmpty == false ? rawTitle! : ownerName
return WindowCandidate(
title: resolvedTitle,
ownerName: ownerName,
windowID: windowID,
ownerPID: ownerPIDNumber.int32Value,
layer: layer
)
}
func isEligible(_ candidate: WindowCandidate) -> Bool {
guard candidate.layer == 0 else { return false }
@ -81,7 +66,7 @@ class ScreenCaptureService: ObservableObject {
}
func captureActiveWindow() async -> NSImage? {
guard let windowInfo = await getActiveWindowInfo() else {
guard let windowInfo = getActiveWindowInfo() else {
return nil
}
@ -158,7 +143,7 @@ class ScreenCaptureService: ObservableObject {
}
}
guard let windowInfo = await getActiveWindowInfo() else {
guard let windowInfo = getActiveWindowInfo() else {
logger.notice("📸 No active window found")
return nil
}