Update: 将子项目从 submodule 转为完整内容

- 移除 GovAI, nomifun-tauri, 算力盒子的 submodule 引用 - 添加所有子项目的完整源代码 - 保留原始 .git 为 .git.bak 备份
2026-07-04 19:20:46 +08:00
parent 54d6465fa7
commit f7a720204a
3360 changed files with 802660 additions and 3 deletions
@@ -0,0 +1,162 @@
+use nomifun_api_types::{DeepgramSpeechToTextConfig, SpeechToTextProvider, SpeechToTextResult};
+use reqwest::Client;
+
+use crate::error::SttError;
+
+const DEFAULT_BASE_URL: &str = "https://api.deepgram.com";
+
+pub async fn transcribe(
+    client: &Client,
+    config: &DeepgramSpeechToTextConfig,
+    audio_data: Vec<u8>,
+    mime_type: &str,
+    language_hint: Option<&str>,
+) -> Result<SpeechToTextResult, SttError> {
+    if config.api_key.is_empty() {
+        return Err(SttError::DeepgramNotConfigured);
+    }
+
+    let base_url = config
+        .base_url
+        .as_deref()
+        .unwrap_or(DEFAULT_BASE_URL)
+        .trim_end_matches('/');
+
+    let mut query_params = vec![("model", config.model.clone())];
+
+    let language = language_hint.or(config.language.as_deref()).filter(|s| !s.is_empty());
+    if let Some(lang) = language {
+        query_params.push(("language", lang.to_owned()));
+    } else if config.detect_language == Some(true) {
+        query_params.push(("detect_language", "true".to_owned()));
+    }
+
+    if config.punctuate == Some(true) {
+        query_params.push(("punctuate", "true".to_owned()));
+    }
+
+    if config.smart_format == Some(true) {
+        query_params.push(("smart_format", "true".to_owned()));
+    }
+
+    let url = format!("{base_url}/v1/listen");
+
+    let response = client
+        .post(&url)
+        .header("Authorization", format!("Token {}", config.api_key))
+        .header("Content-Type", mime_type)
+        .query(&query_params)
+        .body(audio_data)
+        .send()
+        .await
+        .map_err(|e| SttError::RequestFailed(format!("Deepgram request error: {e}")))?;
+
+    let status = response.status();
+    if !status.is_success() {
+        let body = response.text().await.unwrap_or_else(|_| "<unreadable>".to_owned());
+        return Err(SttError::RequestFailed(format!(
+            "Deepgram API returned {status}: {body}"
+        )));
+    }
+
+    let body: serde_json::Value = response
+        .json()
+        .await
+        .map_err(|e| SttError::RequestFailed(format!("failed to parse Deepgram response: {e}")))?;
+
+    let transcript = body["results"]["channels"]
+        .get(0)
+        .and_then(|ch| ch["alternatives"].get(0))
+        .and_then(|alt| alt["transcript"].as_str())
+        .unwrap_or("")
+        .to_owned();
+
+    let detected_language = body["results"]["channels"]
+        .get(0)
+        .and_then(|ch| ch["detected_language"].as_str())
+        .map(|s| s.to_owned())
+        .or_else(|| language.map(|s| s.to_owned()));
+
+    let model_name = extract_model_name(&body).unwrap_or_else(|| config.model.clone());
+
+    Ok(SpeechToTextResult {
+        text: transcript,
+        model: model_name,
+        provider: SpeechToTextProvider::Deepgram,
+        language: detected_language,
+    })
+}
+
+fn extract_model_name(body: &serde_json::Value) -> Option<String> {
+    body["metadata"]["model_info"]
+        .as_object()
+        .and_then(|map| map.values().next())
+        .and_then(|info| info["name"].as_str())
+        .map(|s| s.to_owned())
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn default_base_url_value() {
+        assert_eq!(DEFAULT_BASE_URL, "https://api.deepgram.com");
+    }
+
+    #[tokio::test]
+    async fn empty_api_key_returns_not_configured() {
+        let config = DeepgramSpeechToTextConfig {
+            api_key: String::new(),
+            base_url: None,
+            model: "nova-2".into(),
+            language: None,
+            detect_language: None,
+            punctuate: None,
+            smart_format: None,
+        };
+        let result = transcribe(&Client::new(), &config, vec![0u8; 10], "audio/wav", None).await;
+        assert!(matches!(result, Err(SttError::DeepgramNotConfigured)));
+    }
+
+    #[test]
+    fn extract_model_name_from_response() {
+        let body = serde_json::json!({
+            "metadata": {
+                "model_info": {
+                    "some-uuid": {
+                        "name": "2-general-nova",
+                        "version": "2024-01-18.26916"
+                    }
+                }
+            },
+            "results": {
+                "channels": [{
+                    "alternatives": [{ "transcript": "hello" }]
+                }]
+            }
+        });
+        assert_eq!(extract_model_name(&body), Some("2-general-nova".to_owned()));
+    }
+
+    #[test]
+    fn extract_model_name_missing_metadata() {
+        let body = serde_json::json!({
+            "results": {
+                "channels": [{ "alternatives": [{ "transcript": "hi" }] }]
+            }
+        });
+        assert_eq!(extract_model_name(&body), None);
+    }
+
+    #[test]
+    fn extract_model_name_empty_model_info() {
+        let body = serde_json::json!({
+            "metadata": { "model_info": {} },
+            "results": {
+                "channels": [{ "alternatives": [{ "transcript": "hi" }] }]
+            }
+        });
+        assert_eq!(extract_model_name(&body), None);
+    }
+}