[ADD] locations of the detections & signed URL to the mail

.gitignore

@@ -2,5 +2,5 @@ audioclip/assets/*
 **/*/__pycache__/*
 singing_in_noise/Figures_MS/*
 cloud_analysis/gmail_logs.env
-cloud_analysis/key-file.json
+cloud_analysis/*.json
 logs/logfile.log

cloud_analysis/README.md

@@ -126,7 +126,7 @@ gcloud run deploy \
     --memory 8Gi \
     --region europe-north1 \
     --platform managed \
-    --cpus 4
+    --cpu 4
 ```
 
 Be sure to allocate enough memory for the service to be able to process the files and run the model.
@@ -177,4 +177,10 @@ Explanation of the command:
 
 `--entry-point trigger_audio_analysis`: Specifies that the function to execute is named trigger_audio_analysis within your source code.
 
-`--source .`: Specifies that the source code that should be deployed is in the current directory.
+`--source .`: Specifies that the source code that should be deployed is in the current directory.
+
+## USEFUL POINTS
+
+- It is possible to create a **signed URL** which can be included in the output email and that allow the end user to download the file even though he / she does not have credentials to access the Google Cloud Bucket
+- In the `config.json` file of the audio recorder, add **project_id** as the location so that the folder created has the name of the location. When sending the email, it is then possible to infer **where** the detections have been made.
+

cloud_analysis/main.py

@@ -20,6 +20,9 @@
 from utils.utils import AudioList
 
 from google.cloud import storage
+from google.oauth2 import service_account
+import datetime
+
 import io
 
 import logging
@@ -108,6 +111,38 @@ def fetch_audio_data(bucket_name, blob_name):
 
     return wav_file_object
 
+def generate_signed_url(bucket_name, blob_name, expiration_time=600):
+    """
+    Generates a signed URL for a GCS object.
+    
+    Parameters:
+        bucket_name (str): Name of the GCS bucket.
+        blob_name (str): Name of the blob (file) in the GCS bucket.
+        expiration_time (int): Time, in seconds, until the signed URL expires.
+
+    Returns:
+        str: A signed URL to download the object.
+    """
+    # Path to the service account key file
+    key_path = '/app/cloud_analysis/key-file.json'
+
+    # Initialize a GCS client
+    credentials = service_account.Credentials.from_service_account_file(key_path)
+    client = storage.Client(credentials=credentials)
+
+    # Get the bucket and blob objects
+    bucket = client.get_bucket(bucket_name)
+    blob = bucket.blob(blob_name)
+
+    # Generate the signed URL
+    url = blob.generate_signed_url(
+        version="v4",
+        expiration=datetime.timedelta(seconds=expiration_time),
+        method="GET"
+    )
+
+    return url
+
 
 def analyseAudioFile(
         audio_file_object, min_hr, min_conf, batch_size=1, num_workers=2,
@@ -178,7 +213,7 @@ def on_process_audio(audio_id: str, audio_rec: dict, bucket_name: str, blob_name
             #u"analysisId": analysis_id,
             # Add any other information you want to record here
         })
-
+    
     return count
 
 
@@ -195,7 +230,17 @@ def process_audio_endpoint():
     results = on_process_audio(audio_id, audio_rec, bucket_name, blob_name, hr, conf)
 
     if results > 0:
-        send_email("Snowmobile Detection Alert", f"{results} snowmobile detections were made in the audio file!")
+        # Create a signed URL
+        download_url = generate_signed_url(bucket_name, blob_name)
+
+        # Extract folder name (location) from the blob name
+        location = blob_name.split("/")[0]
+
+        # Write and send the email
+        email_body = f"{results} snowmobile detections were made in the audio file!\n"
+        email_body += f"Detections come from: {location}\n"
+        email_body += f"Download the audio file here: {download_url}"
+        send_email("Snowmobile Detection Alert", email_body)
 
     return jsonify({"message": "Audio processing completed!"})
 

cloud_function_directory/main.py

@@ -56,8 +56,8 @@ def fetch_service_account_key(bucket_name, blob_name):
 def trigger_audio_analysis(data, context):
     file_name = data['name']
     bucket_name = data['bucket']
-    harmonic_ratio = 0.05
-    confidence = 0.80
+    harmonic_ratio = 0.1
+    confidence = 0.95
 
     # URL of your Cloud Run service
     cloud_run_url = "https://model-4uhtnq5xla-lz.a.run.app/process-audio"