voice-router-dev 0.2.6 → 0.2.8

package/dist/index.mjs

@@ -1935,12 +1935,27 @@ var GladiaAdapter = class extends BaseAdapter {
         utterances: this.extractUtterances(transcription),
         summary: result?.summarization?.results || void 0,
         metadata: {
-          requestParams: response.request_params,
-          customMetadata: response.custom_metadata
+          requestParams: response.request_params
         },
         createdAt: response.created_at,
         completedAt: response.completed_at || void 0
       },
+      // Extended data - fully typed from OpenAPI specs
+      extended: {
+        translation: result?.translation || void 0,
+        moderation: result?.moderation || void 0,
+        entities: result?.named_entity_recognition || void 0,
+        sentiment: result?.sentiment_analysis || void 0,
+        audioToLlm: result?.audio_to_llm || void 0,
+        chapters: result?.chapterization || void 0,
+        speakerReidentification: result?.speaker_reidentification || void 0,
+        structuredData: result?.structured_data_extraction || void 0,
+        customMetadata: response.custom_metadata || void 0
+      },
+      // Request tracking
+      tracking: {
+        requestId: response.request_id
+      },
       raw: response
     };
   }
@@ -1968,7 +1983,7 @@ var GladiaAdapter = class extends BaseAdapter {
       }))
     );
     return extractWords(allWords, (item) => ({
-      text: item.word.word,
+      word: item.word.word,
       start: item.word.start,
       end: item.word.end,
       confidence: item.word.confidence,
@@ -1988,11 +2003,11 @@ var GladiaAdapter = class extends BaseAdapter {
       end: utterance.end,
       speaker: utterance.speaker?.toString(),
       confidence: utterance.confidence,
-      words: utterance.words.map((word) => ({
-        text: word.word,
-        start: word.start,
-        end: word.end,
-        confidence: word.confidence
+      words: utterance.words.map((w) => ({
+        word: w.word,
+        start: w.start,
+        end: w.end,
+        confidence: w.confidence
       }))
     }));
   }
@@ -2124,11 +2139,11 @@ var GladiaAdapter = class extends BaseAdapter {
             text: utterance.text,
             isFinal: messageData.is_final,
             confidence: utterance.confidence,
-            words: utterance.words.map((word) => ({
-              text: word.word,
-              start: word.start,
-              end: word.end,
-              confidence: word.confidence
+            words: utterance.words.map((w) => ({
+              word: w.word,
+              start: w.start,
+              end: w.end,
+              confidence: w.confidence
             })),
             data: message
           });
@@ -2142,11 +2157,11 @@ var GladiaAdapter = class extends BaseAdapter {
             end: utterance.end,
             speaker: utterance.speaker?.toString(),
             confidence: utterance.confidence,
-            words: utterance.words.map((word) => ({
-              text: word.word,
-              start: word.start,
-              end: word.end,
-              confidence: word.confidence
+            words: utterance.words.map((w) => ({
+              word: w.word,
+              start: w.start,
+              end: w.end,
+              confidence: w.confidence
             }))
           };
           callbacks?.onUtterance?.(utteranceData);
@@ -2796,18 +2811,29 @@ var AssemblyAIAdapter = class extends BaseAdapter {
         status,
         language: response.language_code,
         duration: response.audio_duration ? response.audio_duration / 1e3 : void 0,
-        // Convert ms to seconds
         speakers: this.extractSpeakers(response),
         words: this.extractWords(response),
         utterances: this.extractUtterances(response),
         summary: response.summary || void 0,
         metadata: {
-          audioUrl: response.audio_url,
-          entities: response.entities,
-          sentimentAnalysis: response.sentiment_analysis_results,
-          contentModeration: response.content_safety_labels
+          audioUrl: response.audio_url
         }
       },
+      // Extended data - fully typed from OpenAPI specs
+      extended: {
+        chapters: response.chapters || void 0,
+        entities: response.entities || void 0,
+        sentimentResults: response.sentiment_analysis_results || void 0,
+        highlights: response.auto_highlights_result || void 0,
+        contentSafety: response.content_safety_labels || void 0,
+        topics: response.iab_categories_result || void 0,
+        languageConfidence: response.language_confidence ?? void 0,
+        throttled: response.throttled ?? void 0
+      },
+      // Request tracking
+      tracking: {
+        requestId: response.id
+      },
       raw: response
     };
   }
@@ -2840,14 +2866,14 @@ var AssemblyAIAdapter = class extends BaseAdapter {
     if (!transcript.words || transcript.words.length === 0) {
       return void 0;
     }
-    return transcript.words.map((word) => ({
-      text: word.text,
-      start: word.start / 1e3,
+    return transcript.words.map((w) => ({
+      word: w.text,
+      start: w.start / 1e3,
       // Convert ms to seconds
-      end: word.end / 1e3,
+      end: w.end / 1e3,
       // Convert ms to seconds
-      confidence: word.confidence,
-      speaker: word.speaker || void 0
+      confidence: w.confidence,
+      speaker: w.speaker || void 0
     }));
   }
   /**
@@ -2865,11 +2891,11 @@ var AssemblyAIAdapter = class extends BaseAdapter {
       // Convert ms to seconds
       speaker: utterance.speaker || void 0,
       confidence: utterance.confidence,
-      words: utterance.words.map((word) => ({
-        text: word.text,
-        start: word.start / 1e3,
-        end: word.end / 1e3,
-        confidence: word.confidence
+      words: utterance.words.map((w) => ({
+        word: w.text,
+        start: w.start / 1e3,
+        end: w.end / 1e3,
+        confidence: w.confidence
       }))
     }));
   }
@@ -2962,12 +2988,12 @@ var AssemblyAIAdapter = class extends BaseAdapter {
             text: turnMsg.transcript,
             isFinal: turnMsg.end_of_turn,
             confidence: turnMsg.end_of_turn_confidence,
-            words: turnMsg.words.map((word) => ({
-              text: word.text,
-              start: word.start / 1e3,
+            words: turnMsg.words.map((w) => ({
+              word: w.text,
+              start: w.start / 1e3,
               // Convert ms to seconds
-              end: word.end / 1e3,
-              confidence: word.confidence
+              end: w.end / 1e3,
+              confidence: w.confidence
             })),
             data: turnMsg
           });
@@ -3278,14 +3304,20 @@ var DeepgramAdapter = class extends BaseAdapter {
         speakers: this.extractSpeakers(response),
         words: this.extractWords(alternative),
         utterances: this.extractUtterances(response),
-        summary: this.extractSummary(alternative),
-        metadata: {
-          modelInfo: response.metadata?.model_info,
-          channels: response.metadata?.channels,
-          sentiment: response.results.sentiments,
-          intents: response.results.intents,
-          topics: response.results.topics
-        }
+        summary: this.extractSummary(alternative)
+      },
+      // Extended data - fully typed from OpenAPI specs
+      extended: {
+        metadata: response.metadata,
+        requestId: response.metadata?.request_id,
+        sha256: response.metadata?.sha256,
+        modelInfo: response.metadata?.model_info,
+        tags: response.metadata?.tags
+      },
+      // Request tracking
+      tracking: {
+        requestId: response.metadata?.request_id,
+        audioHash: response.metadata?.sha256
       },
       raw: response
     };
@@ -3320,11 +3352,11 @@ var DeepgramAdapter = class extends BaseAdapter {
       return void 0;
     }
     return alternative.words.map(
-      (word) => ({
-        text: word.word || "",
-        start: word.start || 0,
-        end: word.end || 0,
-        confidence: word.confidence,
+      (w) => ({
+        word: w.word || "",
+        start: w.start || 0,
+        end: w.end || 0,
+        confidence: w.confidence,
         speaker: void 0
         // Speaker info is at utterance level, not word level
       })
@@ -3344,11 +3376,11 @@ var DeepgramAdapter = class extends BaseAdapter {
       end: utterance.end || 0,
       speaker: utterance.speaker?.toString(),
       confidence: utterance.confidence,
-      words: utterance.words?.map((word) => ({
-        text: word.word || "",
-        start: word.start || 0,
-        end: word.end || 0,
-        confidence: word.confidence
+      words: utterance.words?.map((w) => ({
+        word: w.word || "",
+        start: w.start || 0,
+        end: w.end || 0,
+        confidence: w.confidence
       }))
     }));
   }
@@ -3439,11 +3471,11 @@ var DeepgramAdapter = class extends BaseAdapter {
           if (channel) {
             const transcript = channel.transcript;
             const isFinal = message.is_final;
-            const words = channel.words?.map((word) => ({
-              text: word.word,
-              start: word.start,
-              end: word.end,
-              confidence: word.confidence
+            const words = channel.words?.map((w) => ({
+              word: w.word,
+              start: w.start,
+              end: w.end,
+              confidence: w.confidence
             }));
             callbacks?.onTranscript?.({
               type: "transcript",
@@ -3784,12 +3816,12 @@ var AzureSTTAdapter = class extends BaseAdapter {
     const recognizedPhrases = transcriptionData.recognizedPhrases || [];
     const fullText = combinedPhrases.map((phrase) => phrase.display || phrase.lexical).join(" ") || "";
     const words = recognizedPhrases.flatMap(
-      (phrase) => (phrase.nBest?.[0]?.words || []).map((word) => ({
-        text: word.word,
-        start: word.offsetInTicks / 1e7,
+      (phrase) => (phrase.nBest?.[0]?.words || []).map((w) => ({
+        word: w.word,
+        start: w.offsetInTicks / 1e7,
         // Convert ticks to seconds
-        end: (word.offsetInTicks + word.durationInTicks) / 1e7,
-        confidence: word.confidence,
+        end: (w.offsetInTicks + w.durationInTicks) / 1e7,
+        confidence: w.confidence,
         speaker: phrase.speaker !== void 0 ? phrase.speaker.toString() : void 0
       }))
     );
@@ -3801,11 +3833,12 @@ var AzureSTTAdapter = class extends BaseAdapter {
       id: String(speakerId),
       label: `Speaker ${speakerId}`
     })) : void 0;
+    const transcriptionId = transcription.self?.split("/").pop() || "";
     return {
       success: true,
       provider: this.name,
       data: {
-        id: transcription.self?.split("/").pop() || "",
+        id: transcriptionId,
         text: fullText,
         confidence: recognizedPhrases[0]?.nBest?.[0]?.confidence,
         status: "completed",
@@ -3816,6 +3849,10 @@ var AzureSTTAdapter = class extends BaseAdapter {
         createdAt: transcription.createdDateTime,
         completedAt: transcription.lastActionDateTime
       },
+      extended: {},
+      tracking: {
+        requestId: transcriptionId
+      },
       raw: {
         transcription,
         transcriptionData
@@ -4010,16 +4047,21 @@ var OpenAIWhisperAdapter = class extends BaseAdapter {
    */
   normalizeResponse(response, model, isDiarization) {
     if ("text" in response && Object.keys(response).length === 1) {
+      const requestId2 = `openai-${Date.now()}`;
       return {
         success: true,
         provider: this.name,
         data: {
-          id: `openai-${Date.now()}`,
+          id: requestId2,
           text: response.text,
           status: "completed",
           language: void 0,
           confidence: void 0
         },
+        extended: {},
+        tracking: {
+          requestId: requestId2
+        },
         raw: response
       };
     }
@@ -4038,11 +4080,12 @@ var OpenAIWhisperAdapter = class extends BaseAdapter {
         end: segment.end,
         confidence: void 0
       }));
+      const requestId2 = `openai-${Date.now()}`;
       return {
         success: true,
         provider: this.name,
         data: {
-          id: `openai-${Date.now()}`,
+          id: requestId2,
           text: diarizedResponse.text,
           status: "completed",
           language: void 0,
@@ -4050,39 +4093,53 @@ var OpenAIWhisperAdapter = class extends BaseAdapter {
           speakers,
           utterances
         },
+        extended: {},
+        tracking: {
+          requestId: requestId2
+        },
         raw: response
       };
     }
     if ("duration" in response && "language" in response) {
       const verboseResponse = response;
-      const words = verboseResponse.words?.map((word) => ({
-        text: word.word,
-        start: word.start,
-        end: word.end,
+      const words = verboseResponse.words?.map((w) => ({
+        word: w.word,
+        start: w.start,
+        end: w.end,
         confidence: void 0
       }));
+      const requestId2 = `openai-${Date.now()}`;
       return {
         success: true,
         provider: this.name,
         data: {
-          id: `openai-${Date.now()}`,
+          id: requestId2,
           text: verboseResponse.text,
           status: "completed",
           language: verboseResponse.language,
           duration: verboseResponse.duration,
           words
         },
+        extended: {},
+        tracking: {
+          requestId: requestId2
+        },
         raw: response
       };
     }
+    const requestId = `openai-${Date.now()}`;
     return {
       success: true,
       provider: this.name,
       data: {
-        id: `openai-${Date.now()}`,
+        id: requestId,
         text: "text" in response ? response.text : "",
         status: "completed"
       },
+      extended: {},
+      tracking: {
+        requestId
+      },
       raw: response
     };
   }
@@ -4314,7 +4371,7 @@ var SpeechmaticsAdapter = class extends BaseAdapter {
   normalizeResponse(response) {
     const text = response.results.filter((r) => r.type === "word" && r.alternatives).map((r) => r.alternatives[0]?.content || "").join(" ");
     const words = response.results.filter((r) => r.type === "word" && r.start_time !== void 0 && r.end_time !== void 0).map((result) => ({
-      text: result.alternatives?.[0]?.content || "",
+      word: result.alternatives?.[0]?.content || "",
       start: result.start_time,
       end: result.end_time,
       confidence: result.alternatives?.[0]?.confidence,
@@ -4381,6 +4438,10 @@ var SpeechmaticsAdapter = class extends BaseAdapter {
         summary: response.summary?.content,
         createdAt: response.job.created_at
       },
+      extended: {},
+      tracking: {
+        requestId: response.job.id
+      },
       raw: response
     };
   }
@@ -4461,12 +4522,12 @@ var GladiaWebhookHandler = class extends BaseWebhookHandler {
   /**
    * Convert Gladia WordDTO to unified Word type
    */
-  mapWord(word) {
+  mapWord(w) {
     return {
-      text: word.word,
-      start: word.start,
-      end: word.end,
-      confidence: word.confidence
+      word: w.word,
+      start: w.start,
+      end: w.end,
+      confidence: w.confidence
     };
   }
   /**
@@ -4804,11 +4865,11 @@ var DeepgramWebhookHandler = class extends BaseWebhookHandler {
           raw: payload
         };
       }
-      const words = alternative.words && alternative.words.length > 0 ? alternative.words.map((word) => ({
-        text: word.word || "",
-        start: word.start || 0,
-        end: word.end || 0,
-        confidence: word.confidence
+      const words = alternative.words && alternative.words.length > 0 ? alternative.words.map((w) => ({
+        word: w.word || "",
+        start: w.start || 0,
+        end: w.end || 0,
+        confidence: w.confidence
       })) : void 0;
       const speakers = response.results.utterances && response.results.utterances.length > 0 ? response.results.utterances.map((utterance) => ({
         id: utterance.speaker?.toString() || "unknown",
@@ -4822,11 +4883,11 @@ var DeepgramWebhookHandler = class extends BaseWebhookHandler {
         end: utterance.end || 0,
         speaker: utterance.speaker?.toString(),
         confidence: utterance.confidence,
-        words: utterance.words && utterance.words.length > 0 ? utterance.words.map((word) => ({
-          text: word.word || "",
-          start: word.start || 0,
-          end: word.end || 0,
-          confidence: word.confidence
+        words: utterance.words && utterance.words.length > 0 ? utterance.words.map((w) => ({
+          word: w.word || "",
+          start: w.start || 0,
+          end: w.end || 0,
+          confidence: w.confidence
         })) : void 0
       })) : void 0;
       const summary = alternative.summaries?.[0]?.summary;