utilities

Running

App Files Files Community

Reggie commited on Oct 14, 2023

Commit

6c7ce7e

1 Parent(s): 6aa994f

Drag and drop

Browse files

Files changed (3) hide show

app.py +33 -1
requirements.txt +1 -1
templates/index.html +71 -2

app.py CHANGED Viewed

@@ -12,13 +12,23 @@ import os, time, sys
 from datetime import datetime as dt
 from datetime import timedelta
 from datetime import timezone
 app = Flask(__name__)
 # Initialize Qdrant Client and other required settings
 qdrant_api_key = os.environ.get("qdrant_api_key")
 qdrant_url = os.environ.get("qdrant_url")
 client = QdrantClient(url=qdrant_url, port=443, api_key=qdrant_api_key, prefer_grpc=False)
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
@@ -70,7 +80,7 @@ def search():
     else: results = client.search(collection_name=collection_name, query_vector=(qvector, sq), with_payload=True, limit=100)
     print('SEARCH TIME: ', time.time() - timh)
-    print(results[0].payload['text'].split('\n'))
     try:
         results = [{"text": x.payload['text'], "date": str(int(x.payload['date'])), "id": x.id} for x in results]  # Implement your Qdrant search here
         return jsonify(results)
@@ -84,5 +94,27 @@ def delete_joke():
     client.delete(collection_name="jks", points_selector=models.PointIdsList(points=[int(joke_id)],),)
     return jsonify({"deleted": True})
 if __name__ == "__main__":
     app.run(host="0.0.0.0", debug=True, port=7860)

 from datetime import datetime as dt
 from datetime import timedelta
 from datetime import timezone
+from faster_whisper import WhisperModel
+import io
 app = Flask(__name__)
+# Faster Whisper setup
+# model_size = 'small'
+beamsize = 2
+wmodel = WhisperModel("guillaumekln/faster-whisper-small", device="cpu", compute_type="int8")
 # Initialize Qdrant Client and other required settings
 qdrant_api_key = os.environ.get("qdrant_api_key")
 qdrant_url = os.environ.get("qdrant_url")
+qdrant_api_key = "WaGH94-bo_CzlxTNHFjBAGPvWRhbsWEKUKbMz6YQtYt4oTD1ZXTvwg"
+qdrant_url = "https://c9bee7c7-2bf3-4e1b-8838-2f6f23372ab5.us-east-1-0.aws.cloud.qdrant.io"
 client = QdrantClient(url=qdrant_url, port=443, api_key=qdrant_api_key, prefer_grpc=False)
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
     else: results = client.search(collection_name=collection_name, query_vector=(qvector, sq), with_payload=True, limit=100)
     print('SEARCH TIME: ', time.time() - timh)
+    # print(results[0].payload['text'].split('\n'))
     try:
         results = [{"text": x.payload['text'], "date": str(int(x.payload['date'])), "id": x.id} for x in results]  # Implement your Qdrant search here
         return jsonify(results)
     client.delete(collection_name="jks", points_selector=models.PointIdsList(points=[int(joke_id)],),)
     return jsonify({"deleted": True})
+@app.route("/whisper_transcribe", methods=["POST"])
+def whisper_transcribe():
+    if 'audio' not in request.files: return jsonify({'error': 'No file provided'}), 400
+    audio_file = request.files['audio']
+    allowed_extensions = {'mp3', 'wav', 'ogg', 'm4v'}
+    if not (audio_file and audio_file.filename.lower().split('.')[-1] in allowed_extensions): return jsonify({'error': 'Invalid file format'}), 400
+    print('Transcribing audio')
+    audio_bytes = audio_file.read()
+    audio_file = io.BytesIO(audio_bytes)
+    segments, info = wmodel.transcribe(audio_file, beam_size=beamsize) # beamsize is 2.
+    text = ''
+    starttime = time.time()
+    for segment in segments:
+        text += segment.text
+    print('Time to transcribe:', time.time() - starttime, 'seconds')
+    return jsonify({'transcription': text})
 if __name__ == "__main__":
     app.run(host="0.0.0.0", debug=True, port=7860)

requirements.txt CHANGED Viewed

@@ -2,4 +2,4 @@ transformers
 torch
 qdrant-client
 flask
-Werkzeug

 torch
 qdrant-client
 flask
+faster-whisper

templates/index.html CHANGED Viewed

@@ -30,7 +30,7 @@
             border: none;
             padding: 10px 20px;
             cursor: pointer;
-            margin-left: 10px; /* Add margin to separate the query bar and button */
         }
         #results {
@@ -51,6 +51,17 @@
             cursor: pointer;
             margin: 0 10px;
         }
     </style>
 </head>
 <body>
@@ -60,6 +71,11 @@
         <button id="search">Search</button>
     </div>
     <div id="results"></div>
     <script>
         $(document).ready(function () {
@@ -67,6 +83,7 @@
                 var query = $("#query").val();
                 $.post("/search", { query: query }, function (data) {
                     var results = $("#results");
                     results.empty();
@@ -96,14 +113,66 @@
                 });
             }
             $("#search").on("click", function () {
                 performSearch();
             });
             $("#query").on("keydown", function (event) {
                 if (event.key === "Enter") {
-                    event.preventDefault(); // Prevent form submission
                     performSearch();
                 }
             });
         });

             border: none;
             padding: 10px 20px;
             cursor: pointer;
+            margin-left: 10px;
         }
         #results {
             cursor: pointer;
             margin: 0 10px;
         }
+        #drag-drop-container {
+            border: 2px dashed #4c72af;
+            padding: 20px;
+            text-align: center;
+            cursor: pointer;
+        }
+        .deleted-text {
+            color: red;
+        }
     </style>
 </head>
 <body>
         <button id="search">Search</button>
     </div>
     <div id="results"></div>
+    <!-- Drag and drop area -->
+    <div id="drag-drop-container">
+        Drag & Drop Audio File (mp3, wav, ogg, m4v) Here
+    </div>
     <script>
         $(document).ready(function () {
                 var query = $("#query").val();
                 $.post("/search", { query: query }, function (data) {
+                    // Handle search results
                     var results = $("#results");
                     results.empty();
                 });
             }
+            function displayTranscriptionResult(transcription) {
+                // Display the transcription result
+                var results = $("#results");
+                results.empty();
+                results.append("<div class='result'>" + "<p>" + transcription + "</p>" + "</div>");
+                // Hide the "Transcribing..." message
+                $("#drag-drop-container").text("Drag & Drop Audio File (mp3, wav, ogg, m4v) Here");
+                $("#drag-drop-container").css("border-color", "#4c72af");
+            }
             $("#search").on("click", function () {
                 performSearch();
+                $("#drag-drop-container").css("display", "none"); // Hide drag & drop area
             });
             $("#query").on("keydown", function (event) {
                 if (event.key === "Enter") {
+                    event.preventDefault();
                     performSearch();
+                    $("#drag-drop-container").css("display", "none"); // Hide drag & drop area
+                }
+            });
+            // Drag and drop functionality
+            $("#drag-drop-container").on("dragover", function (e) {
+                e.preventDefault();
+                $(this).css("border", "2px dashed #ff0000");
+            });
+            $("#drag-drop-container").on("dragleave", function (e) {
+                e.preventDefault();
+                $(this).css("border", "2px dashed #4c72af");
+            });
+            $("#drag-drop-container").on("drop", function (e) {
+                e.preventDefault();
+                $(this).css("border", "2px dashed #4c72af");
+                var files = e.originalEvent.dataTransfer.files;
+                if (files.length > 0) {
+                    // Display "Transcribing..." message
+                    $("#drag-drop-container").text("Transcribing the audio file...");
+                    $("#drag-drop-container").css("border-color", "#ccc");
+                    // Assume there's only one file for simplicity
+                    var file = files[0];
+                    var formData = new FormData();
+                    formData.append("audio", file);
+                    $.ajax({
+                        url: "/whisper_transcribe",
+                        type: "POST",
+                        data: formData,
+                        processData: false,
+                        contentType: false,
+                        success: function (data) {
+                            // Handle transcription result
+                            displayTranscriptionResult(data.transcription);
+                        },
+                    });
                 }
             });
         });