using deepgram :)

apps/AppWithWearable/lib/backend/api_requests/api_calls.dart

@@ -1,6 +1,7 @@
 import 'dart:convert';
 import 'dart:io';
 
+import 'package:deepgram_speech_to_text/deepgram_speech_to_text.dart';
 import 'package:flutter/foundation.dart';
 import 'package:flutter/material.dart';
 import 'package:friend_private/backend/database/memory.dart';
@@ -389,6 +390,50 @@ Future<List<TranscriptSegment>> transcribeAudioFile(File file, String uid) async
   }
 }
 
+Future<List<TranscriptSegment>> transcribeAudioFile2(File file) async {
+  var startTime = DateTime.now();
+  Deepgram deepgram = Deepgram(Env.deepgramApiKey!, baseQueryParams: {
+    'model': 'nova-2-general',
+    'detect_language': false,
+    'language': SharedPreferencesUtil().recordingsLanguage,
+    'filler_words': false,
+    'punctuate': true,
+    'diarize': true,
+    'smart_format': true,
+    // TODO: try more options, sentiment analysis, intent, topics
+  });
+
+  DeepgramSttResult res = await deepgram.transcribeFromFile(file);
+  debugPrint('transcribeAudioFile2 took: ${DateTime.now().difference(startTime).inSeconds} seconds');
+  var data = jsonDecode(res.json);
+  var result = data['results']['channels'][0]['alternatives'][0];
+  List<TranscriptSegment> segments = [];
+  for (var word in result['words']) {
+    if (segments.isEmpty) {
+      segments.add(TranscriptSegment(
+          speaker: 'SPEAKER_${word['speaker']}',
+          start: word['start'],
+          end: word['end'],
+          text: word['word'],
+          isUser: false));
+    } else {
+      var lastSegment = segments.last;
+      if (lastSegment.speakerId == word['speaker']) {
+        lastSegment.text += ' ${word['word']}';
+        lastSegment.end = word['end'];
+      } else {
+        segments.add(TranscriptSegment(
+            speaker: 'SPEAKER_${word['speaker']}',
+            start: word['start'],
+            end: word['end'],
+            text: word['word'],
+            isUser: false));
+      }
+    }
+  }
+  return segments;
+}
+
 Future<bool> userHasSpeakerProfile(String uid) async {
   var response = await makeApiCall(
     url: '${Env.customTranscriptApiBaseUrl}profile?uid=$uid',
@@ -436,3 +481,12 @@ Future<bool> uploadSample(File file, String uid) async {
     throw Exception('An error occurred uploadSample: $e');
   }
 }
+
+//curl \
+//   --request POST \
+//   --header 'Authorization: Token YOUR_DEEPGRAM_API_KEY' \
+//   --header 'Content-Type: application/json' \
+//   --data '{"url":"https://dpgr.am/spacewalk.wav"}' \
+//   --url 'https://api.deepgram.com/v1/listen?model=nova-2&smart_format=true'
+
+Future transcribeFromDeepgram(String url) async {}

apps/AppWithWearable/lib/pages/capture/widgets/transcript.dart

@@ -101,7 +101,8 @@ class TranscriptWidgetState extends State<TranscriptWidget> {
           // var containsAudio = await vad.predict(f.readAsBytesSync());
           // debugPrint('Processing audio bytes: ${f.toString()}');
           try {
-            List<TranscriptSegment> segments = await transcribeAudioFile(f, SharedPreferencesUtil().uid);
+            // List<TranscriptSegment> segments = await transcribeAudioFile(f, SharedPreferencesUtil().uid);
+            List<TranscriptSegment> segments = await transcribeAudioFile2(f);
             processCustomTranscript(segments);
           } catch (e) {
             debugPrint(e.toString());

apps/AppWithWearable/pubspec.yaml

@@ -15,7 +15,7 @@ publish_to: 'none' # Remove this line if you wish to publish to pub.dev
 # In iOS, build-name is used as CFBundleShortVersionString while build-number used as CFBundleVersion.
 # Read more about iOS versioning at
 # https://developer.apple.com/library/archive/documentation/General/Reference/InfoPlistKeyReference/Articles/CoreFoundationKeys.html
-version: 1.0.5+17
+version: 1.0.6+19
 
 environment:
   sdk: ">=3.0.0 <4.0.0"
@@ -74,6 +74,7 @@ dependencies:
   objectbox: ^4.0.1
   objectbox_flutter_libs: any
   flutter_foreground_task: ^6.5.0
+  deepgram_speech_to_text: ^2.1.1
 
 dependency_overrides:
   http: ^1.2.1