Compare commits

..
4 Commits
10 changed files with 210 additions and 54 deletions
+17 -11
View File
@@ -20,6 +20,9 @@ jobs:
runs-on: macos-15 runs-on: macos-15
env: env:
VERSION: ${{ inputs.version }} VERSION: ${{ inputs.version }}
APPLE_ID: ${{ secrets.APPLE_ID }}
APPLE_TEAM_ID: ${{ secrets.APPLE_TEAM_ID }}
APPLE_APP_PASSWORD: ${{ secrets.APPLE_APP_PASSWORD }}
SPARKLE_PRIVATE_KEY: ${{ secrets.SPARKLE_PRIVATE_KEY }} SPARKLE_PRIVATE_KEY: ${{ secrets.SPARKLE_PRIVATE_KEY }}
steps: steps:
- uses: actions/checkout@v7 - uses: actions/checkout@v7
@@ -31,7 +34,7 @@ jobs:
APPLE_CERTIFICATE_P12: ${{ secrets.APPLE_CERTIFICATE_P12 }} APPLE_CERTIFICATE_P12: ${{ secrets.APPLE_CERTIFICATE_P12 }}
APPLE_CERTIFICATE_PASSWORD: ${{ secrets.APPLE_CERTIFICATE_PASSWORD }} APPLE_CERTIFICATE_PASSWORD: ${{ secrets.APPLE_CERTIFICATE_PASSWORD }}
run: | run: |
for variable in APPLE_CERTIFICATE_P12 APPLE_CERTIFICATE_PASSWORD SPARKLE_PRIVATE_KEY; do for variable in APPLE_CERTIFICATE_P12 APPLE_CERTIFICATE_PASSWORD APPLE_ID APPLE_TEAM_ID APPLE_APP_PASSWORD SPARKLE_PRIVATE_KEY; do
if [[ -z "${!variable:-}" ]]; then if [[ -z "${!variable:-}" ]]; then
echo "Missing GitHub Actions secret: $variable" >&2 echo "Missing GitHub Actions secret: $variable" >&2
exit 1 exit 1
@@ -53,7 +56,7 @@ jobs:
fi fi
echo "SIGNING_IDENTITY=$signing_identity" >> "$GITHUB_ENV" echo "SIGNING_IDENTITY=$signing_identity" >> "$GITHUB_ENV"
- name: Build and sign release - name: Build, sign, and notarize release
timeout-minutes: 30 timeout-minutes: 30
run: scripts/package_release.sh run: scripts/package_release.sh
@@ -70,13 +73,16 @@ jobs:
GH_TOKEN: ${{ github.token }} GH_TOKEN: ${{ github.token }}
run: | run: |
tag="v$VERSION" tag="v$VERSION"
if git rev-parse "$tag" >/dev/null 2>&1; then if gh release view "$tag" >/dev/null 2>&1; then
echo "Tag already exists: $tag" >&2 gh release upload "$tag" \
exit 1 "$RUNNER_TEMP/meetingnotes-release/release/Meetingnotes-$VERSION.zip" \
"$RUNNER_TEMP/meetingnotes-release/release/appcast.xml" \
--clobber
else
gh release create "$tag" \
"$RUNNER_TEMP/meetingnotes-release/release/Meetingnotes-$VERSION.zip" \
"$RUNNER_TEMP/meetingnotes-release/release/appcast.xml" \
--target "$GITHUB_SHA" \
--title "Meetingnotes $VERSION" \
--generate-notes
fi fi
gh release create "$tag" \
"$RUNNER_TEMP/meetingnotes-release/release/Meetingnotes-$VERSION.zip" \
"$RUNNER_TEMP/meetingnotes-release/release/appcast.xml" \
--target "$GITHUB_SHA" \
--title "Meetingnotes $VERSION" \
--generate-notes
+4 -4
View File
@@ -276,7 +276,7 @@
CODE_SIGN_IDENTITY = "Apple Development"; CODE_SIGN_IDENTITY = "Apple Development";
CODE_SIGN_STYLE = Automatic; CODE_SIGN_STYLE = Automatic;
COMBINE_HIDPI_IMAGES = YES; COMBINE_HIDPI_IMAGES = YES;
CURRENT_PROJECT_VERSION = 29; CURRENT_PROJECT_VERSION = 32;
DEVELOPMENT_ASSET_PATHS = "\"meetingnotes/Preview Content\""; DEVELOPMENT_ASSET_PATHS = "\"meetingnotes/Preview Content\"";
DEVELOPMENT_TEAM = G9LVHZAJNX; DEVELOPMENT_TEAM = G9LVHZAJNX;
ENABLE_HARDENED_RUNTIME = YES; ENABLE_HARDENED_RUNTIME = YES;
@@ -290,7 +290,7 @@
"@executable_path/../Frameworks", "@executable_path/../Frameworks",
); );
MACOSX_DEPLOYMENT_TARGET = 15.0; MACOSX_DEPLOYMENT_TARGET = 15.0;
MARKETING_VERSION = 1.1.17; MARKETING_VERSION = 1.1.20;
ONLY_ACTIVE_ARCH = NO; ONLY_ACTIVE_ARCH = NO;
OTHER_SWIFT_FLAGS = "$(inherited) -D ENABLE_TCC_SPI"; OTHER_SWIFT_FLAGS = "$(inherited) -D ENABLE_TCC_SPI";
PRODUCT_BUNDLE_IDENTIFIER = net.jamesbone.meetingnotes; PRODUCT_BUNDLE_IDENTIFIER = net.jamesbone.meetingnotes;
@@ -312,7 +312,7 @@
CODE_SIGN_IDENTITY = "Apple Development"; CODE_SIGN_IDENTITY = "Apple Development";
CODE_SIGN_STYLE = Automatic; CODE_SIGN_STYLE = Automatic;
COMBINE_HIDPI_IMAGES = YES; COMBINE_HIDPI_IMAGES = YES;
CURRENT_PROJECT_VERSION = 29; CURRENT_PROJECT_VERSION = 32;
DEVELOPMENT_ASSET_PATHS = "\"meetingnotes/Preview Content\""; DEVELOPMENT_ASSET_PATHS = "\"meetingnotes/Preview Content\"";
DEVELOPMENT_TEAM = G9LVHZAJNX; DEVELOPMENT_TEAM = G9LVHZAJNX;
ENABLE_HARDENED_RUNTIME = YES; ENABLE_HARDENED_RUNTIME = YES;
@@ -326,7 +326,7 @@
"@executable_path/../Frameworks", "@executable_path/../Frameworks",
); );
MACOSX_DEPLOYMENT_TARGET = 15.0; MACOSX_DEPLOYMENT_TARGET = 15.0;
MARKETING_VERSION = 1.1.17; MARKETING_VERSION = 1.1.20;
ONLY_ACTIVE_ARCH = YES; ONLY_ACTIVE_ARCH = YES;
OTHER_SWIFT_FLAGS = "$(inherited) -D ENABLE_TCC_SPI"; OTHER_SWIFT_FLAGS = "$(inherited) -D ENABLE_TCC_SPI";
PRODUCT_BUNDLE_IDENTIFIER = net.jamesbone.meetingnotes; PRODUCT_BUNDLE_IDENTIFIER = net.jamesbone.meetingnotes;
+16 -6
View File
@@ -87,8 +87,8 @@ final class AudioManager: NSObject, ObservableObject {
} }
let model = UserDefaultsManager.shared.transcriptionModel let model = UserDefaultsManager.shared.transcriptionModel
async let micResult = transcribe(files[0], model: model) async let micResult = transcribe(files[0], model: model, diarization: false)
async let systemResult = transcribe(files[1], model: model) async let systemResult = transcribe(files[1], model: model, diarization: true)
let (micTranscription, systemTranscription) = await (micResult, systemResult) let (micTranscription, systemTranscription) = await (micResult, systemResult)
let results = [micTranscription, systemTranscription] let results = [micTranscription, systemTranscription]
@@ -125,8 +125,8 @@ final class AudioManager: NSObject, ObservableObject {
let model = UserDefaultsManager.shared.transcriptionModel let model = UserDefaultsManager.shared.transcriptionModel
let micURL = recoveryFiles.first(where: { $0.source == .mic })?.url let micURL = recoveryFiles.first(where: { $0.source == .mic })?.url
let systemURL = recoveryFiles.first(where: { $0.source == .system })?.url let systemURL = recoveryFiles.first(where: { $0.source == .system })?.url
async let micResult = transcribe(micURL, model: model) async let micResult = transcribe(micURL, model: model, diarization: false)
async let systemResult = transcribe(systemURL, model: model) async let systemResult = transcribe(systemURL, model: model, diarization: true)
let (micTranscription, systemTranscription) = await (micResult, systemResult) let (micTranscription, systemTranscription) = await (micResult, systemResult)
let results = [micTranscription, systemTranscription] let results = [micTranscription, systemTranscription]
let (chunks, failures) = buildTranscriptChunks( let (chunks, failures) = buildTranscriptChunks(
@@ -149,10 +149,19 @@ final class AudioManager: NSObject, ObservableObject {
lastRecoveryAudioFolderName = nil lastRecoveryAudioFolderName = nil
} }
private func transcribe(_ fileURL: URL?, model: String) async -> Result<CoderAPIClient.Transcription, Error>? { private func transcribe(
_ fileURL: URL?,
model: String,
diarization: Bool
) async -> Result<CoderAPIClient.Transcription, Error>? {
guard let fileURL else { return nil } guard let fileURL else { return nil }
do { do {
return .success(try await CoderAPIClient.shared.transcribe(fileURL: fileURL, model: model)) return .success(try await CoderAPIClient.shared.transcribe(
fileURL: fileURL,
model: model,
diarization: diarization,
maxSpeakerCount: 4
))
} catch { } catch {
return .failure(error) return .failure(error)
} }
@@ -181,6 +190,7 @@ final class AudioManager: NSObject, ObservableObject {
updated.append(TranscriptChunk( updated.append(TranscriptChunk(
timestamp: captureStartedAt.addingTimeInterval(max(0, segment.start)), timestamp: captureStartedAt.addingTimeInterval(max(0, segment.start)),
source: source, source: source,
speaker: source == .system ? segment.speaker : nil,
text: text, text: text,
isFinal: true isFinal: true
)) ))
@@ -79,7 +79,7 @@ class UserDefaultsManager {
} }
var transcriptionModel: String { var transcriptionModel: String {
get { userDefaults.string(forKey: Keys.transcriptionModel) ?? "groq/whisper-large-v3-turbo" } get { userDefaults.string(forKey: Keys.transcriptionModel) ?? "local-parakeet/parakeet-tdt-0.6b-v3" }
set { userDefaults.set(newValue, forKey: Keys.transcriptionModel) } set { userDefaults.set(newValue, forKey: Keys.transcriptionModel) }
} }
+21 -4
View File
@@ -48,30 +48,46 @@ struct TranscriptChunk: Codable, Identifiable, Hashable {
let id: UUID let id: UUID
let timestamp: Date let timestamp: Date
let source: AudioSource let source: AudioSource
let speaker: Int?
let text: String let text: String
let isFinal: Bool let isFinal: Bool
init(id: UUID = UUID(), timestamp: Date = Date(), source: AudioSource, text: String, isFinal: Bool = false) { init(id: UUID = UUID(), timestamp: Date = Date(), source: AudioSource, speaker: Int? = nil, text: String, isFinal: Bool = false) {
self.id = id self.id = id
self.timestamp = timestamp self.timestamp = timestamp
self.source = source self.source = source
self.speaker = speaker
self.text = text self.text = text
self.isFinal = isFinal self.isFinal = isFinal
} }
var displayName: String {
if source == .mic { return "Me" }
if let speaker { return "Speaker \(speaker)" }
return source.displayName
}
} }
struct CollapsedTranscriptChunk: Identifiable { struct CollapsedTranscriptChunk: Identifiable {
let id: UUID let id: UUID
let timestamp: Date let timestamp: Date
let source: AudioSource let source: AudioSource
let speaker: Int?
let combinedText: String let combinedText: String
init(id: UUID = UUID(), timestamp: Date, source: AudioSource, combinedText: String) { init(id: UUID = UUID(), timestamp: Date, source: AudioSource, speaker: Int? = nil, combinedText: String) {
self.id = id self.id = id
self.timestamp = timestamp self.timestamp = timestamp
self.source = source self.source = source
self.speaker = speaker
self.combinedText = combinedText self.combinedText = combinedText
} }
var displayName: String {
if source == .mic { return "Me" }
if let speaker { return "Speaker \(speaker)" }
return source.displayName
}
} }
struct Meeting: Codable, Identifiable, Hashable { struct Meeting: Codable, Identifiable, Hashable {
@@ -115,7 +131,7 @@ struct Meeting: Codable, Identifiable, Hashable {
var transcript: String { var transcript: String {
return transcriptChunks return transcriptChunks
.filter { $0.isFinal } .filter { $0.isFinal }
.map { "[\($0.source.rawValue)] \($0.text)" } .map { "[\($0.displayName)] \($0.text)" }
.joined(separator: " ") .joined(separator: " ")
} }
@@ -124,7 +140,7 @@ struct Meeting: Codable, Identifiable, Hashable {
let finalChunks = transcriptChunks.filter { $0.isFinal } let finalChunks = transcriptChunks.filter { $0.isFinal }
return finalChunks.map { chunk in return finalChunks.map { chunk in
"[\(TranscriptTimestampFormatter.string(from: chunk.timestamp))] \(chunk.source.copyPrefix): \(chunk.text)" "[\(TranscriptTimestampFormatter.string(from: chunk.timestamp))] \(chunk.displayName): \(chunk.text)"
}.joined(separator: "\n") }.joined(separator: "\n")
} }
@@ -135,6 +151,7 @@ struct Meeting: Codable, Identifiable, Hashable {
id: chunk.id, id: chunk.id,
timestamp: chunk.timestamp, timestamp: chunk.timestamp,
source: chunk.source, source: chunk.source,
speaker: chunk.speaker,
combinedText: chunk.text combinedText: chunk.text
) )
} }
+5 -2
View File
@@ -138,8 +138,11 @@ final class ProcessTap {
tapDescription = CATapDescription(stereoMixdownOfProcesses: [process.objectID]) tapDescription = CATapDescription(stereoMixdownOfProcesses: [process.objectID])
logger.debug("Configuring tap for single process objectID: \(process.objectID)") logger.debug("Configuring tap for single process objectID: \(process.objectID)")
case .systemAudio: case .systemAudio:
tapDescription = CATapDescription(monoGlobalTapButExcludeProcesses: []) // Keep the HAL tap's buffer layout consistent with the default
logger.debug("Configuring a global system audio tap.") // output stream. AudioManager performs the stereo-to-mono mix when
// it converts the captured audio to the 16 kHz transcription file.
tapDescription = CATapDescription(stereoGlobalTapButExcludeProcesses: [])
logger.debug("Configuring a stereo global system audio tap.")
} }
tapDescription.uuid = UUID() tapDescription.uuid = UUID()
+100 -22
View File
@@ -21,6 +21,7 @@ struct CoderModel: Codable, Identifiable, Hashable {
var supportsChat: Bool { capabilities.isEmpty || capabilities.contains("chat") } var supportsChat: Bool { capabilities.isEmpty || capabilities.contains("chat") }
var supportsTranscription: Bool { capabilities.contains("audio_transcription") } var supportsTranscription: Bool { capabilities.contains("audio_transcription") }
var supportsSpeakerDiarization: Bool { capabilities.contains("speaker_diarization") }
} }
enum CoderAPIError: LocalizedError { enum CoderAPIError: LocalizedError {
@@ -54,6 +55,7 @@ final class CoderAPIClient {
let start: TimeInterval let start: TimeInterval
let end: TimeInterval let end: TimeInterval
let text: String let text: String
let speaker: Int?
} }
let text: String let text: String
@@ -70,8 +72,16 @@ final class CoderAPIClient {
} }
private struct TranscriptionResponse: Decodable { private struct TranscriptionResponse: Decodable {
struct Word: Decodable {
let word: String
let start: TimeInterval
let end: TimeInterval
let speaker: Int?
}
let text: String let text: String
let segments: [Transcription.Segment]? let segments: [Transcription.Segment]?
let words: [Word]?
} }
private struct AudioChunk { private struct AudioChunk {
@@ -159,11 +169,17 @@ final class CoderAPIClient {
} }
} }
func transcribe(fileURL: URL, model: String, language: String = "en") async throws -> Transcription { func transcribe(
fileURL: URL,
model: String,
language: String = "en",
diarization: Bool = false,
maxSpeakerCount: Int = 4
) async throws -> Transcription {
let selectedModel = model.trimmingCharacters(in: .whitespacesAndNewlines) let selectedModel = model.trimmingCharacters(in: .whitespacesAndNewlines)
guard !selectedModel.isEmpty else { throw CoderAPIError.missingModel("transcription") } guard !selectedModel.isEmpty else { throw CoderAPIError.missingModel("transcription") }
let apiKey = try requiredAPIKey(KeychainHelper.shared.getCoderAPIKey() ?? "") let apiKey = try requiredAPIKey(KeychainHelper.shared.getCoderAPIKey() ?? "")
let chunks = try makeAudioChunks(from: fileURL) let chunks = try makeAudioChunks(from: fileURL, preserveSpeakerIdentity: diarization)
defer { defer {
for chunk in chunks where chunk.isTemporary { for chunk in chunks where chunk.isTemporary {
try? FileManager.default.removeItem(at: chunk.url) try? FileManager.default.removeItem(at: chunk.url)
@@ -180,13 +196,15 @@ final class CoderAPIClient {
chunk.url, chunk.url,
model: selectedModel, model: selectedModel,
language: language, language: language,
apiKey: apiKey apiKey: apiKey,
diarization: diarization,
maxSpeakerCount: maxSpeakerCount
) )
if transcription.segments.isEmpty { if transcription.segments.isEmpty {
let text = transcription.text.trimmingCharacters(in: .whitespacesAndNewlines) let text = transcription.text.trimmingCharacters(in: .whitespacesAndNewlines)
if !text.isEmpty { if !text.isEmpty {
textParts.append(text) textParts.append(text)
segments.append(.init(start: chunk.offset, end: chunk.offset, text: text)) segments.append(.init(start: chunk.offset, end: chunk.offset, text: text, speaker: nil))
} }
continue continue
} }
@@ -206,7 +224,8 @@ final class CoderAPIClient {
segments.append(.init( segments.append(.init(
start: segment.start + chunk.offset, start: segment.start + chunk.offset,
end: segment.end + chunk.offset, end: segment.end + chunk.offset,
text: text text: text,
speaker: segment.speaker
)) ))
} }
} }
@@ -218,13 +237,17 @@ final class CoderAPIClient {
_ fileURL: URL, _ fileURL: URL,
model: String, model: String,
language: String, language: String,
apiKey: String apiKey: String,
diarization: Bool,
maxSpeakerCount: Int
) async throws -> Transcription { ) async throws -> Transcription {
let boundary = "Meetingnotes-\(UUID().uuidString)" let boundary = "Meetingnotes-\(UUID().uuidString)"
let bodyURL = try makeMultipartBody( let bodyURL = try makeMultipartBody(
audioURL: fileURL, audioURL: fileURL,
model: model, model: model,
language: language, language: language,
diarization: diarization,
maxSpeakerCount: maxSpeakerCount,
boundary: boundary boundary: boundary
) )
defer { try? FileManager.default.removeItem(at: bodyURL) } defer { try? FileManager.default.removeItem(at: bodyURL) }
@@ -240,22 +263,18 @@ final class CoderAPIClient {
let (data, response) = try await transcriptionSession.upload(for: request, fromFile: bodyURL) let (data, response) = try await transcriptionSession.upload(for: request, fromFile: bodyURL)
try validate(response: response, data: data) try validate(response: response, data: data)
let decoded = try JSONDecoder().decode(TranscriptionResponse.self, from: data) let decoded = try JSONDecoder().decode(TranscriptionResponse.self, from: data)
return Transcription(text: decoded.text, segments: decoded.segments ?? []) let segments = decoded.segments ?? segments(from: decoded.words ?? [])
return Transcription(text: decoded.text, segments: segments)
} }
private func makeAudioChunks(from fileURL: URL) throws -> [AudioChunk] { private func makeAudioChunks(from fileURL: URL, preserveSpeakerIdentity: Bool) throws -> [AudioChunk] {
let input = try AVAudioFile(forReading: fileURL) let input = try AVAudioFile(forReading: fileURL)
let format = input.processingFormat let format = input.processingFormat
guard format.sampleRate > 0 else { guard format.sampleRate > 0 else { throw CoderAPIError.invalidResponse }
return [AudioChunk(url: fileURL, offset: 0, isTemporary: false)]
}
let duration = Double(input.length) / format.sampleRate let framesPerChunk = preserveSpeakerIdentity
guard duration > transcriptionChunkDuration else { ? max(1, input.length)
return [AudioChunk(url: fileURL, offset: 0, isTemporary: false)] : AVAudioFramePosition(format.sampleRate * transcriptionChunkDuration)
}
let framesPerChunk = AVAudioFramePosition(format.sampleRate * transcriptionChunkDuration)
var chunks: [AudioChunk] = [] var chunks: [AudioChunk] = []
var frameOffset: AVAudioFramePosition = 0 var frameOffset: AVAudioFramePosition = 0
@@ -263,7 +282,7 @@ final class CoderAPIClient {
while frameOffset < input.length { while frameOffset < input.length {
let frameCount = min(framesPerChunk, input.length - frameOffset) let frameCount = min(framesPerChunk, input.length - frameOffset)
let chunkURL = FileManager.default.temporaryDirectory let chunkURL = FileManager.default.temporaryDirectory
.appendingPathComponent("meetingnotes-transcription-\(UUID().uuidString).m4a") .appendingPathComponent("meetingnotes-transcription-\(UUID().uuidString).wav")
try writeAudioChunk( try writeAudioChunk(
from: input, from: input,
frameCount: frameCount, frameCount: frameCount,
@@ -293,10 +312,13 @@ final class CoderAPIClient {
to outputURL: URL to outputURL: URL
) throws { ) throws {
let settings: [String: Any] = [ let settings: [String: Any] = [
AVFormatIDKey: kAudioFormatMPEG4AAC, AVFormatIDKey: kAudioFormatLinearPCM,
AVSampleRateKey: format.sampleRate, AVSampleRateKey: format.sampleRate,
AVNumberOfChannelsKey: format.channelCount, AVNumberOfChannelsKey: format.channelCount,
AVEncoderBitRateKey: 48_000 * max(1, Int(format.channelCount)) AVLinearPCMBitDepthKey: 16,
AVLinearPCMIsFloatKey: false,
AVLinearPCMIsBigEndianKey: false,
AVLinearPCMIsNonInterleaved: false
] ]
let output = try AVAudioFile( let output = try AVAudioFile(
forWriting: outputURL, forWriting: outputURL,
@@ -317,6 +339,51 @@ final class CoderAPIClient {
} }
} }
private func segments(from words: [TranscriptionResponse.Word]) -> [Transcription.Segment] {
var result: [Transcription.Segment] = []
var currentWords: [String] = []
var currentStart: TimeInterval?
var currentEnd: TimeInterval = 0
var currentSpeaker: Int?
func flush() {
guard let start = currentStart, !currentWords.isEmpty else { return }
result.append(.init(
start: start,
end: currentEnd,
text: currentWords.joined(separator: " "),
speaker: currentSpeaker
))
currentWords.removeAll(keepingCapacity: true)
currentStart = nil
currentEnd = 0
currentSpeaker = nil
}
for word in words {
let text = word.word.trimmingCharacters(in: .whitespacesAndNewlines)
guard !text.isEmpty else { continue }
let speakerChanged = currentStart != nil && word.speaker != currentSpeaker
let longPause = currentStart != nil && word.start - currentEnd > 1.5
if speakerChanged || longPause { flush() }
if currentStart == nil {
currentStart = word.start
currentSpeaker = word.speaker
}
currentWords.append(text)
currentEnd = word.end
let sentenceEnded = text.last.map { ".!?".contains($0) } ?? false
let duration = currentEnd - (currentStart ?? currentEnd)
if currentWords.count >= 40 || (sentenceEnded && (currentWords.count >= 12 || duration >= 8)) {
flush()
}
}
flush()
return result
}
private func endpoint(baseURL: String, path: String) throws -> URL { private func endpoint(baseURL: String, path: String) throws -> URL {
guard var components = URLComponents(string: baseURL.trimmingCharacters(in: .whitespacesAndNewlines)), guard var components = URLComponents(string: baseURL.trimmingCharacters(in: .whitespacesAndNewlines)),
let scheme = components.scheme?.lowercased(), let scheme = components.scheme?.lowercased(),
@@ -346,7 +413,14 @@ final class CoderAPIClient {
} }
} }
private func makeMultipartBody(audioURL: URL, model: String, language: String, boundary: String) throws -> URL { private func makeMultipartBody(
audioURL: URL,
model: String,
language: String,
diarization: Bool,
maxSpeakerCount: Int,
boundary: String
) throws -> URL {
let bodyURL = FileManager.default.temporaryDirectory.appendingPathComponent("meetingnotes-upload-\(UUID().uuidString).body") let bodyURL = FileManager.default.temporaryDirectory.appendingPathComponent("meetingnotes-upload-\(UUID().uuidString).body")
_ = FileManager.default.createFile(atPath: bodyURL.path, contents: nil) _ = FileManager.default.createFile(atPath: bodyURL.path, contents: nil)
let output = try FileHandle(forWritingTo: bodyURL) let output = try FileHandle(forWritingTo: bodyURL)
@@ -358,7 +432,11 @@ final class CoderAPIClient {
try write("--\(boundary)\r\nContent-Disposition: form-data; name=\"model\"\r\n\r\n\(model)\r\n") try write("--\(boundary)\r\nContent-Disposition: form-data; name=\"model\"\r\n\r\n\(model)\r\n")
try write("--\(boundary)\r\nContent-Disposition: form-data; name=\"language\"\r\n\r\n\(language)\r\n") try write("--\(boundary)\r\nContent-Disposition: form-data; name=\"language\"\r\n\r\n\(language)\r\n")
try write("--\(boundary)\r\nContent-Disposition: form-data; name=\"response_format\"\r\n\r\nverbose_json\r\n") try write("--\(boundary)\r\nContent-Disposition: form-data; name=\"response_format\"\r\n\r\nverbose_json\r\n")
try write("--\(boundary)\r\nContent-Disposition: form-data; name=\"file\"; filename=\"\(audioURL.lastPathComponent)\"\r\nContent-Type: audio/mp4\r\n\r\n") if diarization {
try write("--\(boundary)\r\nContent-Disposition: form-data; name=\"diarization\"\r\n\r\ntrue\r\n")
try write("--\(boundary)\r\nContent-Disposition: form-data; name=\"max_speaker_count\"\r\n\r\n\(maxSpeakerCount)\r\n")
}
try write("--\(boundary)\r\nContent-Disposition: form-data; name=\"file\"; filename=\"\(audioURL.lastPathComponent)\"\r\nContent-Type: audio/wav\r\n\r\n")
let input = try FileHandle(forReadingFrom: audioURL) let input = try FileHandle(forReadingFrom: audioURL)
defer { try? input.close() } defer { try? input.close() }
while let chunk = try input.read(upToCount: 1 << 20), !chunk.isEmpty { while let chunk = try input.read(upToCount: 1 << 20), !chunk.isEmpty {
+2 -2
View File
@@ -208,12 +208,12 @@ struct CollapsedTranscriptChunkView: View {
.font(.caption) .font(.caption)
.foregroundColor(chunk.source == .mic ? .blue : .orange) .foregroundColor(chunk.source == .mic ? .blue : .orange)
Text(chunk.source.displayName) Text(chunk.displayName)
.font(.caption) .font(.caption)
.fontWeight(.medium) .fontWeight(.medium)
.foregroundColor(chunk.source == .mic ? .blue : .orange) .foregroundColor(chunk.source == .mic ? .blue : .orange)
} }
.frame(width: 50, alignment: .leading) .frame(width: 78, alignment: .leading)
// Transcript text // Transcript text
Text(chunk.combinedText) Text(chunk.combinedText)
+12
View File
@@ -58,6 +58,18 @@ struct SettingsView: View {
Text(model.displayName).tag(model.id) Text(model.displayName).tag(model.id)
} }
} }
if let selectedModel = viewModel.coderModels.first(where: { $0.id == viewModel.settings.transcriptionModel }) {
if selectedModel.supportsSpeakerDiarization {
Label("Remote participants are labeled Speaker 14; your microphone is labeled Me.", systemImage: "person.2.wave.2")
.font(.caption)
.foregroundColor(.secondary)
} else {
Label("Remote participants are labeled Them; your microphone is labeled Me.", systemImage: "person.2")
.font(.caption)
.foregroundColor(.secondary)
}
}
} }
Text("The token is stored locally in Keychain. Audio and note generation are sent only to this Coder service.") Text("The token is stored locally in Keychain. Audio and note generation are sent only to this Coder service.")
+32 -2
View File
@@ -14,6 +14,9 @@ APP_PATH="$DERIVED_DATA/Build/Products/Release/$APP_NAME.app"
required_variables=( required_variables=(
VERSION VERSION
SIGNING_IDENTITY SIGNING_IDENTITY
APPLE_ID
APPLE_TEAM_ID
APPLE_APP_PASSWORD
SPARKLE_PRIVATE_KEY SPARKLE_PRIVATE_KEY
GITHUB_REPOSITORY GITHUB_REPOSITORY
) )
@@ -80,6 +83,33 @@ ARCHIVE_NAME="$APP_NAME-$VERSION.zip"
ARCHIVE_PATH="$RELEASE_DIR/$ARCHIVE_NAME" ARCHIVE_PATH="$RELEASE_DIR/$ARCHIVE_NAME"
ditto -c -k --sequesterRsrc --keepParent "$APP_PATH" "$ARCHIVE_PATH" ditto -c -k --sequesterRsrc --keepParent "$APP_PATH" "$ARCHIVE_PATH"
NOTARY_RESULT="$BUILD_ROOT/notary-result.json"
xcrun notarytool submit "$ARCHIVE_PATH" \
--apple-id "$APPLE_ID" \
--team-id "$APPLE_TEAM_ID" \
--password "$APPLE_APP_PASSWORD" \
--wait \
--timeout 20m \
--output-format json > "$NOTARY_RESULT"
NOTARY_STATUS=$(plutil -extract status raw -o - "$NOTARY_RESULT")
if [[ "$NOTARY_STATUS" != "Accepted" ]]; then
submission_id=$(plutil -extract id raw -o - "$NOTARY_RESULT")
xcrun notarytool log "$submission_id" \
--apple-id "$APPLE_ID" \
--team-id "$APPLE_TEAM_ID" \
--password "$APPLE_APP_PASSWORD" || true
echo "Apple notarization failed with status: $NOTARY_STATUS" >&2
exit 1
fi
xcrun stapler staple "$APP_PATH"
xcrun stapler validate "$APP_PATH"
spctl --assess --type execute --verbose=2 "$APP_PATH"
rm -f "$ARCHIVE_PATH"
ditto -c -k --sequesterRsrc --keepParent "$APP_PATH" "$ARCHIVE_PATH"
GENERATE_APPCAST=$(find "$DERIVED_DATA/SourcePackages/artifacts" -type f -name generate_appcast -print -quit) GENERATE_APPCAST=$(find "$DERIVED_DATA/SourcePackages/artifacts" -type f -name generate_appcast -print -quit)
if [[ -z "$GENERATE_APPCAST" ]]; then if [[ -z "$GENERATE_APPCAST" ]]; then
echo "Sparkle generate_appcast tool was not found" >&2 echo "Sparkle generate_appcast tool was not found" >&2
@@ -97,8 +127,8 @@ grep -q "$DOWNLOAD_URL$ARCHIVE_NAME" "$RELEASE_DIR/appcast.xml"
grep -q 'sparkle:edSignature=' "$RELEASE_DIR/appcast.xml" grep -q 'sparkle:edSignature=' "$RELEASE_DIR/appcast.xml"
if [[ -n "${GITHUB_STEP_SUMMARY:-}" ]]; then if [[ -n "${GITHUB_STEP_SUMMARY:-}" ]]; then
printf 'Built and Developer ID-signed Meetingnotes %s. The GitHub release is ready to publish.\n' \ printf 'Built, Developer ID-signed, notarized, and stapled Meetingnotes %s. The GitHub release is ready to publish.\n' \
"$VERSION" >> "$GITHUB_STEP_SUMMARY" "$VERSION" >> "$GITHUB_STEP_SUMMARY"
fi fi
echo "Signed release artifacts are ready in $RELEASE_DIR" echo "Signed and notarized release artifacts are ready in $RELEASE_DIR"