teaching-feedback-assistant/server/src/config.rs

use std::{env, net::IpAddr, path::PathBuf};

#[derive(Clone)]
pub enum SpeechProviderConfig {
    Local { api_url: String },
    Tencent(TencentSpeechConfig),
}

#[derive(Clone)]
pub struct TencentSpeechConfig {
    pub app_id: u64,
    pub secret_id: String,
    pub secret_key: String,
    pub engine_type: String,
}

#[derive(Clone)]
pub struct SpeechConfig {
    pub provider: SpeechProviderConfig,
    pub request_timeout_seconds: u64,
    pub worker_concurrency: usize,
    pub job_lease_seconds: i64,
}

#[derive(Clone)]
pub struct SummaryConfig {
    pub api_url: String,
    pub api_key: Option<String>,
    pub model: String,
}

#[derive(Clone)]
pub struct WechatAuthConfig {
    pub app_id: String,
    pub app_secret: String,
    pub code2session_url: String,
}

#[derive(Clone)]
pub struct AuthConfig {
    pub wechat: Option<WechatAuthConfig>,
    pub session_ttl_days: i64,
    pub allow_development_user_header: bool,
}

#[derive(Clone)]
pub struct Config {
    pub host: IpAddr,
    pub port: u16,
    pub database_url: Option<String>,
    pub database_max_connections: u32,
    pub cors_allowed_origin: Option<String>,
    pub audio_storage_dir: PathBuf,
    pub speech: Option<SpeechConfig>,
    pub summary: Option<SummaryConfig>,
    pub auth: AuthConfig,
}

impl Config {
    pub fn from_env() -> Result<Self, String> {
        let host = env::var("HOST")
            .unwrap_or_else(|_| "127.0.0.1".to_owned())
            .parse()
            .map_err(|_| "HOST must be a valid IP address".to_owned())?;
        let port = env::var("PORT")
            .unwrap_or_else(|_| "8080".to_owned())
            .parse()
            .map_err(|_| "PORT must be a valid u16 value".to_owned())?;
        let database_max_connections = env::var("DATABASE_MAX_CONNECTIONS")
            .unwrap_or_else(|_| "5".to_owned())
            .parse()
            .map_err(|_| "DATABASE_MAX_CONNECTIONS must be a valid u32 value".to_owned())?;

        let speech = speech_config()?;
        let summary = summary_config()?;
        let auth = auth_config()?;

        Ok(Self {
            host,
            port,
            database_url: env::var("DATABASE_URL")
                .ok()
                .filter(|value| !value.trim().is_empty()),
            database_max_connections,
            cors_allowed_origin: env::var("CORS_ALLOWED_ORIGIN")
                .ok()
                .filter(|value| !value.trim().is_empty()),
            audio_storage_dir: env::var("AUDIO_STORAGE_DIR")
                .map(PathBuf::from)
                .unwrap_or_else(|_| PathBuf::from("./data/audio")),
            speech,
            summary,
            auth,
        })
    }
}

fn auth_config() -> Result<AuthConfig, String> {
    let app_id = optional_env("WECHAT_APP_ID");
    let app_secret = optional_env("WECHAT_APP_SECRET");
    let wechat = match (app_id, app_secret) {
        (Some(app_id), Some(app_secret)) => Some(WechatAuthConfig {
            app_id,
            app_secret,
            code2session_url: optional_env("WECHAT_CODE2SESSION_URL")
                .unwrap_or_else(|| "https://api.weixin.qq.com/sns/jscode2session".to_owned()),
        }),
        (None, None) => None,
        _ => {
            return Err(
                "WECHAT_APP_ID and WECHAT_APP_SECRET must be configured together".to_owned(),
            );
        }
    };
    let session_ttl_days = parse_env("AUTH_SESSION_TTL_DAYS", 30_i64)?;
    if !(1..=365).contains(&session_ttl_days) {
        return Err("AUTH_SESSION_TTL_DAYS must be between 1 and 365".to_owned());
    }
    let allow_development_user_header = parse_bool_env("ALLOW_DEVELOPMENT_USER_HEADER", false)?;

    Ok(AuthConfig {
        wechat,
        session_ttl_days,
        allow_development_user_header,
    })
}

fn speech_config() -> Result<Option<SpeechConfig>, String> {
    let provider = optional_env("ASR_PROVIDER").unwrap_or_else(|| "local".to_owned());
    if matches!(provider.as_str(), "disabled" | "none") {
        return Ok(None);
    }

    let provider = match provider.as_str() {
        "local" | "funasr" => SpeechProviderConfig::Local {
            api_url: optional_env("LOCAL_ASR_URL")
                .unwrap_or_else(|| "http://127.0.0.1:10095".to_owned())
                .trim_end_matches('/')
                .to_owned(),
        },
        "tencent" => {
            let app_id = required_env("TENCENT_CLOUD_ASR_APP_ID")?
                .parse()
                .map_err(|_| "TENCENT_CLOUD_ASR_APP_ID must be an integer".to_owned())?;
            SpeechProviderConfig::Tencent(TencentSpeechConfig {
                app_id,
                secret_id: required_env("TENCENT_CLOUD_ASR_SECRET_ID")?,
                secret_key: required_env("TENCENT_CLOUD_ASR_SECRET_KEY")?,
                engine_type: optional_env("TENCENT_CLOUD_ASR_ENGINE_TYPE")
                    .unwrap_or_else(|| "16k_zh".to_owned()),
            })
        }
        _ => {
            return Err(
                "ASR_PROVIDER must be local, funasr, tencent, disabled, or none".to_owned(),
            );
        }
    };

    let request_timeout_seconds = parse_env("ASR_REQUEST_TIMEOUT_SECONDS", 1_800_u64)?;
    let worker_concurrency = parse_env("ASR_WORKER_CONCURRENCY", 1_usize)?;
    let job_lease_seconds = parse_env("ASR_JOB_LEASE_SECONDS", 3_600_i64)?;
    if request_timeout_seconds == 0 {
        return Err("ASR_REQUEST_TIMEOUT_SECONDS must be greater than 0".to_owned());
    }
    if worker_concurrency == 0 {
        return Err("ASR_WORKER_CONCURRENCY must be greater than 0".to_owned());
    }
    if job_lease_seconds <= 0 {
        return Err("ASR_JOB_LEASE_SECONDS must be greater than 0".to_owned());
    }

    Ok(Some(SpeechConfig {
        provider,
        request_timeout_seconds,
        worker_concurrency,
        job_lease_seconds,
    }))
}

fn summary_config() -> Result<Option<SummaryConfig>, String> {
    let Some(api_url) = optional_env("FEEDBACK_SUMMARY_API_URL") else {
        return Ok(None);
    };
    let model = optional_env("FEEDBACK_SUMMARY_MODEL").ok_or_else(|| {
        "FEEDBACK_SUMMARY_MODEL is required when summary API is configured".to_owned()
    })?;
    Ok(Some(SummaryConfig {
        api_url: api_url.trim_end_matches('/').to_owned(),
        api_key: optional_env("FEEDBACK_SUMMARY_API_KEY"),
        model,
    }))
}

fn optional_env(name: &str) -> Option<String> {
    env::var(name)
        .ok()
        .map(|value| value.trim().to_owned())
        .filter(|value| !value.is_empty())
}

fn required_env(name: &str) -> Result<String, String> {
    optional_env(name).ok_or_else(|| format!("{name} is required for the selected ASR provider"))
}

fn parse_env<T>(name: &str, default: T) -> Result<T, String>
where
    T: std::str::FromStr,
{
    optional_env(name).map_or_else(
        || Ok(default),
        |value| value.parse().map_err(|_| format!("{name} is invalid")),
    )
}

fn parse_bool_env(name: &str, default: bool) -> Result<bool, String> {
    let Some(value) = optional_env(name) else {
        return Ok(default);
    };
    match value.to_ascii_lowercase().as_str() {
        "1" | "true" | "yes" | "on" => Ok(true),
        "0" | "false" | "no" | "off" => Ok(false),
        _ => Err(format!("{name} must be true or false")),
    }
}