mirror of
https://github.com/anomalyco/hex.git
synced 2026-11-21 17:56:51 +08:00
* iOS: Italian, French and Portuguese full-keyboard layouts that autocorrect in their own language A TestFlight tester on build 31 with an Italian iPhone reported that Hex kept autocorrecting their Italian into English. KeyboardLanguage only knew en, nb, es and de, so Automatic mapped an it-IT dictation language (and any other language without a layout) to English: English letters, and UITextChecker's en_US dictionary rewriting Italian words. Adds italian/french/portuguese cases. They type on the English QWERTY rows (Apple's Italian and Portuguese keyboards are QWERTY; Hex has no AZERTY yet, so French is QWERTY too), with their own long-press accent order (Italian à è ì ò ù, é right after è), their UITextChecker dictionary (it, fr, pt_BR then pt_PT), and an Italian common two-letter list. Automatic now resolves it-*, fr-* and pt-* to these; the dictation language still wins over the iPhone's keyboards, and Hebrew or Russian still get English. Automatic deliberately ignores Hex's own input mode: its primaryLanguage is the Info.plist's en-US whatever you type. Spatial autocorrect (#2059) only has en and es models; the new layouts get no model and fall back to the dictionary engine in their own language (the decoder also refuses a model of another language). The space bar already names the dictation language ("Italiano"). Snapshot scenes get Italian, French and Portuguese drafts (--dictation-language it-IT, --keyboard-layout it|fr|pt). Verified on an iOS 27.2 iPhone 17 Pro simulator: full HEXMobileTests suite (272 tests, 0 failures, 1 skipped as before). New cases in KeyboardLayoutTests, AutocorrectEngineTests and FullKeyboardEngineTests. Measured on 12 Italian sentences (19 one-slip typos, 59 correct words) with the real dictionary: English fallback fixed 4, miscorrected 4 and broke 19 correct words (Domani->Domain, stai->stay, Sono->Soon, molto->moto); Italian fixed 17, miscorrected 0 and broke none. Renders of the Italian keyboard (typing with the "ritardo" suggestion, the E and I accent menus, "Italiano" on the space bar) checked by eye. * iOS: spoken "at the rate" and "at sign" become "@", joining a spoken email address An Indian English TestFlight tester (build 32) said "at the rate", India's name for @, and got the words. Apple's on-device models don't fix it: checked on macOS 27.2 with Hex's settings and an en_IN voice, SpeechTranscriber (Hex's default) writes "My email is John at the rategmail.com." and "Send it to Priya at signyahoo.com." (gluing the domain to the word before), and DictationTranscriber writes "at the rate gmail.com", though it already writes "priya@yahoo.com" for "at sign" and "john@gmail.com" for "john at gmail dot com". SpokenCleanupRules.spokenSymbols (English only) turns "at the rate" and "at sign" into "@". Before a domain ("gmail dot com", "gmail.com", "co dot in", or glued as above), the word before (with any "dot"-joined parts) and the domain join into one lowercase address, "dot" becoming "." only there: "John at the rate Gmail dot com." becomes "john@gmail.com.". "at the rate" counts only before a domain, so "at the rate of 5%" and "at the rate we planned" stay, and "at" alone is never touched; "at sign" elsewhere is a lone "@" (but not in "at sign language"). Only the matched words change. It runs from CleanupTagger.cleanIfEnabled, which both the app's and in-keyboard dictation call: in every style but Verbatim, with Clean up what I say on or off (the experiment's guard moved below it; the tagger itself still only runs when on). TranscriptFormatter: Polished capitalized after every full stop, which would have written "john@gmail.Com."; a sentence now starts only after ".", "!" or "?" and a space, an address keeps its case (Polished and Notes), and a lone address gets no full stop. Verified on an iOS 27.2 simulator: full HEXMobileTests suite (272 tests, 0 failures, 1 skipped as before), with new cases in TranscriptCleanupTests (Apple's real outputs included) and CleanupTaggerTests.
257 lines
12 KiB
Swift
257 lines
12 KiB
Swift
import XCTest
|
|
@testable import HEX
|
|
|
|
final class TranscriptCleanupGuardTests: XCTestCase {
|
|
func testAcceptsMisheardWordsAndSpokenSelfCorrections() {
|
|
XCTAssertTrue(
|
|
TranscriptCleanupGuard.isAcceptable(
|
|
original: "we should use cloud code on tuesday, no wait, wednesday",
|
|
cleaned: "We should use Claude Code on Wednesday."
|
|
)
|
|
)
|
|
}
|
|
|
|
func testAcceptsPunctuationAndCaseChanges() {
|
|
XCTAssertTrue(
|
|
TranscriptCleanupGuard.isAcceptable(
|
|
original: "meet at ten on thursday and bring the notes",
|
|
cleaned: "Meet at ten on Thursday, and bring the notes."
|
|
)
|
|
)
|
|
}
|
|
|
|
func testRejectsDigitChanges() {
|
|
XCTAssertFalse(
|
|
TranscriptCleanupGuard.isAcceptable(
|
|
original: "call me at 555 0100",
|
|
cleaned: "call me at 555 0101"
|
|
)
|
|
)
|
|
XCTAssertFalse(
|
|
TranscriptCleanupGuard.isAcceptable(
|
|
original: "see you at ten",
|
|
cleaned: "see you at 10"
|
|
)
|
|
)
|
|
}
|
|
|
|
func testRejectsLostNegation() {
|
|
XCTAssertFalse(
|
|
TranscriptCleanupGuard.isAcceptable(
|
|
original: "do not delete the file",
|
|
cleaned: "delete the file"
|
|
)
|
|
)
|
|
XCTAssertFalse(
|
|
TranscriptCleanupGuard.isAcceptable(
|
|
original: "I don't think that is right",
|
|
cleaned: "I think that is right"
|
|
)
|
|
)
|
|
}
|
|
|
|
func testRejectsAddedNegation() {
|
|
XCTAssertFalse(
|
|
TranscriptCleanupGuard.isAcceptable(
|
|
original: "I will go tomorrow",
|
|
cleaned: "I will not go tomorrow"
|
|
)
|
|
)
|
|
}
|
|
|
|
func testRejectsRunawayLength() {
|
|
XCTAssertFalse(
|
|
TranscriptCleanupGuard.isAcceptable(
|
|
original: "ship it",
|
|
cleaned: String(repeating: "ship it ", count: 30)
|
|
)
|
|
)
|
|
}
|
|
|
|
func testRejectsEmptyOutputAndEmptyInput() {
|
|
XCTAssertFalse(TranscriptCleanupGuard.isAcceptable(original: "", cleaned: ""))
|
|
XCTAssertFalse(TranscriptCleanupGuard.isAcceptable(original: "ship it", cleaned: " "))
|
|
}
|
|
|
|
func testRejectsAnAnswerWithLittleOverlap() {
|
|
XCTAssertFalse(
|
|
TranscriptCleanupGuard.isAcceptable(
|
|
original: "what time is the meeting",
|
|
cleaned: "The weather tomorrow will be sunny and warm."
|
|
)
|
|
)
|
|
}
|
|
|
|
func testCleanupSettingIsOffByDefaultAndPersists() {
|
|
let suite = "AppleIntelligenceCleanupSettingsTests-\(UUID().uuidString)"
|
|
let defaults = UserDefaults(suiteName: suite)!
|
|
defer { defaults.removePersistentDomain(forName: suite) }
|
|
|
|
XCTAssertFalse(AppleIntelligenceCleanupSettings(defaults: defaults).isEnabled)
|
|
|
|
AppleIntelligenceCleanupSettings(defaults: defaults).isEnabled = true
|
|
XCTAssertTrue(AppleIntelligenceCleanupSettings(defaults: defaults).isEnabled)
|
|
}
|
|
}
|
|
|
|
/// Clean up what I say's word rules for Spanish, German and Norwegian: what goes, and what
|
|
/// must stay because the words mean something there.
|
|
final class SpokenCleanupRulesTests: XCTestCase {
|
|
func testSpanishFillersAndCorrections() {
|
|
for (spoken, cleaned) in [
|
|
("Nos vemos mañana a las 3, digo, a las 4.", "Nos vemos mañana a las 4."),
|
|
("Quedamos el martes, perdón, el miércoles en la oficina.", "Quedamos el miércoles en la oficina."),
|
|
("Compra dos kilos, mejor dicho, tres kilos de tomates.", "Compra tres kilos de tomates."),
|
|
("Te llamo hoy, no, mañana por la tarde.", "Te llamo mañana por la tarde."),
|
|
("Llegué a las 8, perdón, a las 9 de la mañana.", "Llegué a las 9 de la mañana."),
|
|
("Llego a las 3. Digo, a las 4.", "Llego a las 4."),
|
|
("Ven el lunes, no, mejor el martes.", "Ven el martes."),
|
|
("A las 3, no, espera, a las 5.", "A las 5."),
|
|
("¿El martes, perdón, el miércoles?", "¿El miércoles?"),
|
|
("Este, quería preguntarte si vienes a la cena.", "Quería preguntarte si vienes a la cena."),
|
|
("Es muy caro, o sea, no lo voy a comprar.", "Es muy caro, no lo voy a comprar."),
|
|
("Pues, este, no sé qué decirte.", "Pues, no sé qué decirte."),
|
|
("Y, este, fuimos al cine.", "Y fuimos al cine."),
|
|
("Mmm, creo que sí, eh, pero no estoy seguro.", "Creo que sí, pero no estoy seguro."),
|
|
("Eh, nos vemos a las 3, digo, a las 4.", "Nos vemos a las 4."),
|
|
] {
|
|
XCTAssertEqual(SpokenCleanupRules.clean(spoken, language: "es-ES"), cleaned)
|
|
}
|
|
}
|
|
|
|
func testSpanishWordsInTheirOwnMeaningStay() {
|
|
for spoken in [
|
|
"Este libro me gustó mucho.",
|
|
"Quiero este, el azul.",
|
|
"Te digo que no puedo ir.",
|
|
"Vamos a la playa, digo yo, si hace buen tiempo.",
|
|
"No, gracias, ya comí.",
|
|
"Perdón por llegar tarde, había mucho tráfico.",
|
|
"Lo siento, perdón, Ana.",
|
|
"O sea que al final no vienes.",
|
|
"Le dije a Juan que no, que mejor otro día.",
|
|
"Un coche no mejor que el otro.",
|
|
"¿Vienes el sábado o el domingo?",
|
|
"No te olvides, ¿eh?",
|
|
"Padre e hijo.",
|
|
"Uno\ndos",
|
|
] {
|
|
XCTAssertEqual(SpokenCleanupRules.clean(spoken, language: "es-419"), spoken)
|
|
}
|
|
}
|
|
|
|
func testGermanAndNorwegianCorrections() {
|
|
XCTAssertEqual(SpokenCleanupRules.clean("Wir treffen uns um 3, nein, um 4.", language: "de-DE"), "Wir treffen uns um 4.")
|
|
XCTAssertEqual(
|
|
SpokenCleanupRules.clean("Ich komme am Montag, ich meine, am Dienstag.", language: "de-AT"),
|
|
"Ich komme am Dienstag."
|
|
)
|
|
XCTAssertEqual(SpokenCleanupRules.clean("Nein, ich meine das ernst.", language: "de-DE"), "Nein, ich meine das ernst.")
|
|
XCTAssertEqual(SpokenCleanupRules.clean("Vi sender den tirsdag, nei vent, torsdag.", language: "nb-NO"), "Vi sender den torsdag.")
|
|
XCTAssertEqual(SpokenCleanupRules.clean("Møtet er klokka 3, nei, klokka 4.", language: "nn"), "Møtet er klokka 4.")
|
|
XCTAssertEqual(SpokenCleanupRules.clean("Nei, jeg vil ikke.", language: "nb-NO"), "Nei, jeg vil ikke.")
|
|
XCTAssertEqual(SpokenCleanupRules.clean("Eh, jeg kommer snart.", language: "nb-NO"), "Jeg kommer snart.")
|
|
XCTAssertEqual(SpokenCleanupRules.clean("Äh, das ist eh egal.", language: "de-DE"), "Das ist eh egal.")
|
|
}
|
|
|
|
func testOnlyItsLanguagesHaveRules() {
|
|
XCTAssertTrue(SpokenCleanupRules.supports("es-US"))
|
|
XCTAssertTrue(SpokenCleanupRules.supports("no"))
|
|
XCTAssertFalse(SpokenCleanupRules.supports("en-US"))
|
|
XCTAssertFalse(SpokenCleanupRules.supports("pt-BR"))
|
|
XCTAssertEqual(SpokenCleanupRules.clean("A las 3, digo, a las 4.", language: "pt-BR"), "A las 3, digo, a las 4.")
|
|
}
|
|
|
|
/// "at the rate" is how India says "@": an address joins into one, "dot" becoming "."
|
|
/// only inside it. "at sign" is "@" anywhere.
|
|
func testSpokenAtSignsBecomeAddresses() {
|
|
for (spoken, written) in [
|
|
("john at the rate gmail dot com", "john@gmail.com"),
|
|
("Mail it to John at the rate Gmail dot com.", "Mail it to john@gmail.com."),
|
|
("My email is priya dot sharma at the rate company dot co dot in, call me too",
|
|
"My email is priya.sharma@company.co.in, call me too"),
|
|
("Write to support at the rate hex.app today", "Write to support@hex.app today"),
|
|
("rahul123 at sign yahoo dot com", "rahul123@yahoo.com"),
|
|
("At the rate gmail dot com addresses work", "@gmail.com addresses work"),
|
|
("Type the at sign, then your name", "Type the @, then your name"),
|
|
("Press at sign.", "Press @."),
|
|
("ravi at the rate, gmail dot com", "ravi@gmail.com"),
|
|
// What Apple's on-device models wrote for an Indian English voice: SpeechTranscriber
|
|
// glues the domain to "rate" or "sign"; DictationTranscriber keeps them apart.
|
|
("My email is John at the rategmail.com.", "My email is john@gmail.com."),
|
|
("Send it to Priya at signyahoo.com.", "Send it to priya@yahoo.com."),
|
|
("My email is John at the rate gmail.com", "My email is john@gmail.com"),
|
|
] {
|
|
XCTAssertEqual(SpokenCleanupRules.spokenSymbols(in: spoken, language: "en-IN"), written, spoken)
|
|
}
|
|
}
|
|
|
|
func testAtInItsOwnMeaningStays() {
|
|
for spoken in [
|
|
"Prices are growing at the rate of 5% a year.",
|
|
"At the rate of knots, we'll be there soon.",
|
|
"We are spending at the rate we planned.",
|
|
"At the rate you're going, you'll finish early.",
|
|
"Meet me at the office at 5.",
|
|
"Look at that dot on the map.",
|
|
"I'm good at sign language.",
|
|
"john at the rate gmail",
|
|
"Send it at the rate",
|
|
"Wait at the gate dot com",
|
|
// Apple's own output that must stay: "@" already there, "at" alone, "at the rate of".
|
|
"Send it to priya@yahoo.com",
|
|
"My email is john at gmail.com.",
|
|
"Prices Are growing at the rate of5% a year",
|
|
"Rates went up at the ratepayers' meeting.",
|
|
] {
|
|
XCTAssertEqual(SpokenCleanupRules.spokenSymbols(in: spoken, language: "en-IN"), spoken)
|
|
}
|
|
// English only: other languages say it their own way, and keep their words.
|
|
XCTAssertEqual(
|
|
SpokenCleanupRules.spokenSymbols(in: "john at the rate gmail dot com", language: "es-ES"),
|
|
"john at the rate gmail dot com"
|
|
)
|
|
}
|
|
|
|
/// Every style but Verbatim keeps the address whole: no capital after its dots, none on
|
|
/// it at a sentence start, and no full stop after an address said alone.
|
|
func testStylesKeepAnAddressWhole() {
|
|
XCTAssertEqual(TranscriptFormatter.format("john@gmail.com", style: .polished), "john@gmail.com")
|
|
XCTAssertEqual(
|
|
TranscriptFormatter.format("send it to john@gmail.com", style: .polished),
|
|
"Send it to john@gmail.com."
|
|
)
|
|
XCTAssertEqual(
|
|
TranscriptFormatter.format("john@gmail.com is mine. call me", style: .polished),
|
|
"john@gmail.com is mine. Call me."
|
|
)
|
|
XCTAssertEqual(TranscriptFormatter.format("Mail john@gmail.com", style: .casual), "mail john@gmail.com")
|
|
XCTAssertEqual(
|
|
TranscriptFormatter.format("bullet john@gmail.com bullet call Ravi", style: .notes),
|
|
"- john@gmail.com\n- Call Ravi"
|
|
)
|
|
XCTAssertEqual(TranscriptFormatter.format("Hello. world! ok? yes", style: .polished), "Hello. World! Ok? Yes.")
|
|
}
|
|
|
|
/// The styles trim each language's own hesitation sounds, never a word ("e" is "and" in
|
|
/// Spanish, "eh" is "anyway" in German).
|
|
func testStylesTrimSpanishAndGermanHesitations() {
|
|
XCTAssertEqual(
|
|
TranscriptFormatter.format("Mmm, creo que sí, eh, pero no estoy seguro.", style: .polished, language: "es-ES"),
|
|
"Creo que sí, pero no estoy seguro."
|
|
)
|
|
XCTAssertEqual(
|
|
TranscriptFormatter.format("Eh, ¿qué tal? Padre e hijo.", style: .polished, language: "es-ES"),
|
|
"¿Qué tal? Padre e hijo."
|
|
)
|
|
XCTAssertEqual(
|
|
TranscriptFormatter.format("Äh, das ist eh egal.", style: .polished, language: "de-DE"),
|
|
"Das ist eh egal."
|
|
)
|
|
XCTAssertEqual(
|
|
TranscriptFormatter.format("Eh, nos vemos.", style: .verbatim, language: "es-ES"),
|
|
"Eh, nos vemos."
|
|
)
|
|
}
|
|
}
|