add upload

app.py

@@ -1,4 +1,7 @@
 import spaces
+import boto3
+from botocore.exceptions import NoCredentialsError, ClientError
+from botocore.client import Config
 
 import os, pathlib
 
@@ -51,7 +54,53 @@ try:
 except OSError as e:
     sys.exit(f"❌  Could not load {cnn_so} : {e}")
 
+S3_ENDPOINT = os.getenv("S3_ENDPOINT")
+S3_ACCESS_KEY = os.getenv("S3_ACCESS_KEY")
+S3_SECRET_KEY = os.getenv("S3_SECRET_KEY")
 
+
+
+# Function to upload file to Cloudflare R2
+def upload_to_r2(file_path, bucket_name, object_name):
+    """
+    Upload a file to a Cloudflare R2 bucket.
+
+    :param file_path: Path to the file to upload.
+    :param bucket_name: Name of the R2 bucket.
+    :param object_name: Name of the object to save in the bucket.
+    :param r2_access_key: Cloudflare R2 access key.
+    :param r2_secret_key: Cloudflare R2 secret key.
+    :param r2_account_id: Cloudflare R2 account ID.
+    :return: True if file was uploaded, else False.
+    """
+    try:
+        # Initialize a session using Cloudflare R2 credentials
+        session = boto3.session.Session()
+        s3 = session.client('s3',
+            endpoint_url=f'https://{S3_ENDPOINT}',
+            aws_access_key_id=S3_ACCESS_KEY,
+            aws_secret_access_key=S3_SECRET_KEY,
+            config = Config(s3={"addressing_style": "virtual"}, signature_version='s3v4'),
+            #region_name = 'auto'
+        )
+
+        # Upload the file to R2 bucket
+        s3.upload_file(file_path, bucket_name, object_name)
+        print(f"File '{file_path}' uploaded to R2 bucket '{bucket_name}' as '{object_name}'")
+        return True
+    except FileNotFoundError:
+        print(f"The file {file_path} was not found")
+        return False
+    except NoCredentialsError:
+        print("Credentials not available")
+        return False
+    except ClientError as e:
+        print(f"Failed to upload file to R2 bucket: {e}")
+        return False
+    except Exception as e:
+        print(f"An unexpected error occurred: {e}")
+        return False
+    
 from huggingface_hub import snapshot_download
 MODEL_REPO = "deepdml/faster-whisper-large-v3-turbo-ct2"   # CT2 format
 LOCAL_DIR  = f"{CACHE_ROOT}/whisper_turbo"
@@ -667,7 +716,7 @@ class WhisperTranscriber:
                 transcription_results = self.group_segments_by_speaker(transcription_results)
             
             # Step 6: Return results
-            return {
+            result = {
                 "segments": transcription_results,
                 "language": detected_language,
                 "num_speakers": detected_num_speakers,
@@ -675,6 +724,12 @@ class WhisperTranscriber:
                 "batch_size": batch_size,
                 "speaker_embeddings": speaker_embeddings,
             }
+            filekey = f"ai-transcribe/split/{task_json["job_id"]}-{task_json["task_id"]["chunk"]["idx"]}.json"
+            ret = upload_to_r2(json.dumps(result), "intermediate", filekey)
+            if ret:
+                return {"filekey": filekey}
+            else:
+                return {"error": "Failed to upload to R2"}
                 
         except Exception as e:
             import traceback

requirements.txt

@@ -20,4 +20,5 @@ soundfile>=0.12.0
 ffmpeg-python>=0.2.0
 requests>=2.28.0
 nvidia-cudnn-cu12==9.1.0.70  # any 9.1.x that pip can find is fine
-webrtcvad>=2.0.10
+webrtcvad>=2.0.10
+boto3