typecast-rust 0.3.5

#![cfg(feature = "e2e")]
//! End-to-end tests for Typecast Rust SDK
//!
//! These tests hit the real Typecast API and require a valid TYPECAST_API_KEY
//! environment variable. They are gated behind the `e2e` cargo feature so that
//! the default `cargo test` (and `cargo llvm-cov`) only runs unit tests with
//! mocked HTTP and never requires network access or credentials.
//!
//! Run with: TYPECAST_API_KEY=your_key cargo test --features e2e

use typecast_rust::{
    AudioFormat, ClientConfig, EmotionPreset, Gender, Output, PresetPrompt, SmartPrompt, TTSModel,
    TTSRequest, TTSRequestWithTimestamps, TypecastClient, VoicesV2Filter,
};

fn get_client() -> TypecastClient {
    // Load .env file if present (for local development)
    let _ = dotenvy::dotenv();

    TypecastClient::from_env().expect("Failed to create client. Make sure TYPECAST_API_KEY is set")
}

fn get_test_voice_id() -> &'static str {
    // Using a well-known voice ID from the API
    "tc_60e5426de8b95f1d3000d7b5"
}

#[tokio::test]
async fn test_get_voices_v2() {
    let client = get_client();

    let voices = client
        .get_voices_v2(None)
        .await
        .expect("Failed to get voices");

    assert!(!voices.is_empty(), "Should have at least one voice");

    // Check first voice has required fields
    let first_voice = &voices[0];
    assert!(
        !first_voice.voice_id.is_empty(),
        "voice_id should not be empty"
    );
    assert!(
        !first_voice.voice_name.is_empty(),
        "voice_name should not be empty"
    );
    assert!(!first_voice.models.is_empty(), "models should not be empty");

    println!("Found {} voices", voices.len());
    println!(
        "First voice: {} ({})",
        first_voice.voice_name, first_voice.voice_id
    );
}

#[tokio::test]
async fn test_get_voices_v2_with_filter() {
    let client = get_client();

    let filter = VoicesV2Filter::new()
        .model(TTSModel::SsfmV30)
        .gender(Gender::Female);

    let voices = client
        .get_voices_v2(Some(filter))
        .await
        .expect("Failed to get filtered voices");

    assert!(
        !voices.is_empty(),
        "Should have at least one female voice with ssfm-v30"
    );

    // Verify all returned voices are female
    for voice in &voices {
        if let Some(gender) = &voice.gender {
            assert_eq!(*gender, Gender::Female, "All voices should be female");
        }
    }

    println!("Found {} female voices with ssfm-v30", voices.len());
}

#[tokio::test]
async fn test_get_voice_v2_by_id() {
    let client = get_client();
    let voice_id = get_test_voice_id();

    let voice = client
        .get_voice_v2(voice_id)
        .await
        .expect("Failed to get voice by ID");

    assert_eq!(voice.voice_id, voice_id, "Voice ID should match");
    assert!(
        !voice.voice_name.is_empty(),
        "voice_name should not be empty"
    );
    assert!(!voice.models.is_empty(), "models should not be empty");

    println!("Voice: {} ({})", voice.voice_name, voice.voice_id);
    println!("Gender: {:?}, Age: {:?}", voice.gender, voice.age);
    println!("Use cases: {:?}", voice.use_cases);

    for model in &voice.models {
        println!(
            "  Model {:?}: emotions = {:?}",
            model.version, model.emotions
        );
    }
}

#[tokio::test]
async fn test_get_voice_v2_not_found() {
    let client = get_client();

    let result = client.get_voice_v2("invalid_voice_id").await;

    assert!(
        result.is_err(),
        "Should return an error for invalid voice ID"
    );

    let error = result.unwrap_err();
    // API may return different error codes for invalid voice IDs
    assert!(
        error.is_not_found() || error.is_validation_error() || error.is_bad_request(),
        "Should be NotFound, ValidationError, or BadRequest, got: {:?}",
        error
    );

    println!("Expected error: {}", error);
}

#[tokio::test]
async fn test_text_to_speech_basic() {
    let client = get_client();
    let voice_id = get_test_voice_id();

    let request = TTSRequest::new(
        voice_id,
        "Hello, this is a test of the Typecast text to speech API.",
        TTSModel::SsfmV30,
    )
    .language("eng");

    let response = client
        .text_to_speech(&request)
        .await
        .expect("Failed to generate speech");

    assert!(
        !response.audio_data.is_empty(),
        "Audio data should not be empty"
    );
    // Duration may be 0 if the header is not provided by the server
    assert!(response.duration >= 0.0, "Duration should be non-negative");
    assert_eq!(
        response.format,
        AudioFormat::Wav,
        "Default format should be WAV"
    );

    println!(
        "Generated audio: {} bytes, {:.2} seconds, format: {:?}",
        response.audio_data.len(),
        response.duration,
        response.format
    );

    // Verify WAV header
    assert!(
        response.audio_data.len() > 44,
        "WAV file should have header"
    );
    assert_eq!(
        &response.audio_data[0..4],
        b"RIFF",
        "Should start with RIFF"
    );
}

#[tokio::test]
async fn test_text_to_speech_with_mp3_format() {
    let client = get_client();
    let voice_id = get_test_voice_id();

    let request = TTSRequest::new(
        voice_id,
        "This is a test with MP3 output format.",
        TTSModel::SsfmV30,
    )
    .language("eng")
    .output(Output::new().audio_format(AudioFormat::Mp3));

    let response = client
        .text_to_speech(&request)
        .await
        .expect("Failed to generate MP3 speech");

    assert!(
        !response.audio_data.is_empty(),
        "Audio data should not be empty"
    );
    assert_eq!(response.format, AudioFormat::Mp3, "Format should be MP3");

    println!(
        "Generated MP3: {} bytes, {:.2} seconds",
        response.audio_data.len(),
        response.duration
    );
}

#[tokio::test]
async fn test_text_to_speech_with_preset_emotion() {
    let client = get_client();
    let voice_id = get_test_voice_id();

    let request = TTSRequest::new(
        voice_id,
        "I am so happy today! Everything is wonderful!",
        TTSModel::SsfmV30,
    )
    .language("eng")
    .prompt(
        PresetPrompt::new()
            .emotion_preset(EmotionPreset::Happy)
            .emotion_intensity(1.5),
    );

    let response = client
        .text_to_speech(&request)
        .await
        .expect("Failed to generate happy speech");

    assert!(
        !response.audio_data.is_empty(),
        "Audio data should not be empty"
    );

    println!(
        "Generated happy speech: {} bytes, {:.2} seconds",
        response.audio_data.len(),
        response.duration
    );
}

#[tokio::test]
async fn test_text_to_speech_with_smart_prompt() {
    let client = get_client();
    let voice_id = get_test_voice_id();

    let request = TTSRequest::new(
        voice_id,
        "Everything is so incredibly perfect that I feel like I'm dreaming.",
        TTSModel::SsfmV30,
    )
    .language("eng")
    .prompt(
        SmartPrompt::new()
            .previous_text("I feel like I'm walking on air and I just want to scream with joy!")
            .next_text(
                "I am literally bursting with happiness and I never want this feeling to end!",
            ),
    );

    let response = client
        .text_to_speech(&request)
        .await
        .expect("Failed to generate smart speech");

    assert!(
        !response.audio_data.is_empty(),
        "Audio data should not be empty"
    );

    println!(
        "Generated smart speech: {} bytes, {:.2} seconds",
        response.audio_data.len(),
        response.duration
    );
}

#[tokio::test]
async fn test_text_to_speech_with_output_settings() {
    let client = get_client();
    let voice_id = get_test_voice_id();

    let request = TTSRequest::new(
        voice_id,
        "Testing audio output settings with custom volume and tempo.",
        TTSModel::SsfmV30,
    )
    .language("eng")
    .output(Output::new().volume(120).audio_pitch(2).audio_tempo(1.2));

    let response = client
        .text_to_speech(&request)
        .await
        .expect("Failed to generate customized speech");

    assert!(
        !response.audio_data.is_empty(),
        "Audio data should not be empty"
    );

    println!(
        "Generated customized speech: {} bytes, {:.2} seconds",
        response.audio_data.len(),
        response.duration
    );
}

#[tokio::test]
async fn test_text_to_speech_with_seed() {
    let client = get_client();
    let voice_id = get_test_voice_id();

    let request = TTSRequest::new(
        voice_id,
        "Testing reproducible output with a seed value.",
        TTSModel::SsfmV30,
    )
    .language("eng")
    .seed(42);

    let response = client
        .text_to_speech(&request)
        .await
        .expect("Failed to generate seeded speech");

    assert!(
        !response.audio_data.is_empty(),
        "Audio data should not be empty"
    );

    println!(
        "Generated seeded speech: {} bytes, {:.2} seconds",
        response.audio_data.len(),
        response.duration
    );
}

#[tokio::test]
async fn test_text_to_speech_korean() {
    let client = get_client();
    let voice_id = get_test_voice_id();

    let request = TTSRequest::new(
        voice_id,
        "안녕하세요, 타입캐스트 텍스트 투 스피치 API 테스트입니다.",
        TTSModel::SsfmV30,
    )
    .language("kor");

    let response = client
        .text_to_speech(&request)
        .await
        .expect("Failed to generate Korean speech");

    assert!(
        !response.audio_data.is_empty(),
        "Audio data should not be empty"
    );

    println!(
        "Generated Korean speech: {} bytes, {:.2} seconds",
        response.audio_data.len(),
        response.duration
    );
}

#[tokio::test]
async fn test_text_to_speech_invalid_voice() {
    let client = get_client();

    let request = TTSRequest::new("invalid_voice_id", "This should fail.", TTSModel::SsfmV30);

    let result = client.text_to_speech(&request).await;

    assert!(
        result.is_err(),
        "Should return an error for invalid voice ID"
    );

    let error = result.unwrap_err();
    println!("Expected error: {}", error);
}

#[tokio::test]
async fn test_client_with_custom_config() {
    let config =
        ClientConfig::new(std::env::var("TYPECAST_API_KEY").expect("TYPECAST_API_KEY not set"))
            .timeout(std::time::Duration::from_secs(120));

    let client = TypecastClient::new(config).expect("Failed to create client with custom config");

    let voices = client
        .get_voices_v2(None)
        .await
        .expect("Failed to get voices");

    assert!(!voices.is_empty(), "Should have at least one voice");
    println!(
        "Client with custom config works! Found {} voices",
        voices.len()
    );
}

// ─────────────────────────────────────────────────────────────────────────────
// Text-to-Speech with Timestamps E2E tests
// ─────────────────────────────────────────────────────────────────────────────

const TIMESTAMP_VOICE: &str = "tc_60e5426de8b95f1d3000d7b5";

fn build_timestamp_request(text: &str, language: &str) -> TTSRequestWithTimestamps {
    TTSRequestWithTimestamps::new(TIMESTAMP_VOICE, text, TTSModel::SsfmV30)
        .language(language)
        .prompt(serde_json::json!({
            "emotion_type": "preset",
            "emotion_preset": "normal",
            "emotion_intensity": 1.0
        }))
        .seed(42)
}

#[tokio::test]
async fn test_with_timestamps_no_granularity() {
    let client = get_client();
    let req = build_timestamp_request("Hello.", "eng");

    let resp = client
        .text_to_speech_with_timestamps(&req, None)
        .await
        .expect("text_to_speech_with_timestamps failed");

    assert!(resp.audio_duration > 0.0, "audio_duration should be > 0");
    let words = resp.words.as_ref().expect("words should not be None");
    assert!(!words.is_empty(), "words should be non-empty");
    let chars = resp
        .characters
        .as_ref()
        .expect("characters should not be None");
    assert!(!chars.is_empty(), "characters should be non-empty");
    println!(
        "no_granularity: duration={:.2} words={} chars={}",
        resp.audio_duration,
        words.len(),
        chars.len()
    );
}

#[tokio::test]
async fn test_with_timestamps_word_granularity() {
    let client = get_client();
    let req = build_timestamp_request("Hello.", "eng");

    let resp = client
        .text_to_speech_with_timestamps(&req, Some("word"))
        .await
        .expect("text_to_speech_with_timestamps (word) failed");

    let words = resp.words.as_ref().expect("words should not be None");
    assert!(
        !words.is_empty(),
        "words should be non-empty for word granularity"
    );
    let chars_empty = resp
        .characters
        .as_ref()
        .map(|c| c.is_empty())
        .unwrap_or(true);
    assert!(
        chars_empty,
        "characters should be None/empty for word granularity"
    );
    println!("word granularity: words={}", words.len());
}

#[tokio::test]
async fn test_with_timestamps_char_granularity() {
    let client = get_client();
    let req = build_timestamp_request("Hello.", "eng");

    let resp = client
        .text_to_speech_with_timestamps(&req, Some("char"))
        .await
        .expect("text_to_speech_with_timestamps (char) failed");

    let chars = resp
        .characters
        .as_ref()
        .expect("characters should not be None");
    assert!(
        !chars.is_empty(),
        "characters should be non-empty for char granularity"
    );
    let words_empty = resp.words.as_ref().map(|w| w.is_empty()).unwrap_or(true);
    assert!(
        words_empty,
        "words should be None/empty for char granularity"
    );
    println!("char granularity: chars={}", chars.len());
}

#[tokio::test]
async fn test_with_timestamps_jpn_char() {
    let client = get_client();
    let req = build_timestamp_request("こんにちは。お元気ですか?", "jpn");

    let resp = client
        .text_to_speech_with_timestamps(&req, Some("char"))
        .await
        .expect("text_to_speech_with_timestamps (jpn+char) failed");

    let chars = resp
        .characters
        .as_ref()
        .expect("characters should not be None for jpn+char");
    assert!(
        chars.len() >= 5,
        "Expected >= 5 character segments for Japanese, got {}",
        chars.len()
    );
    println!("jpn+char: chars={}", chars.len());
}