Adding a command line client

client/client.py

@@ -0,0 +1,127 @@
+'''
+Simple command line client to remove vocals from audio/video tracks.
+'''
+
+import os
+import json
+import time
+
+import click
+import requests
+from pygments import highlight, lexers, formatters
+
+
+API_ENDPOINT_URL = os.getenv('API_ENDPOINT_URL')
+
+
+def api_req(ctx: click.core.Context, method: str, url: str, **kwargs):
+    method = method.lower()
+    if method not in ['get', 'post']:
+        raise ValueError(f'Unsupported HTTP method "{method}"')
+
+    actual_url = ctx.obj['url'](url)
+
+    if ctx.obj['verbose']:
+        print(f'GET {actual_url}')
+
+    response = getattr(requests, method)(actual_url, **kwargs)
+
+    if ctx.obj['verbose']:
+        print(f'Server response status: {response.status_code}')
+        echo_obj(response.json())
+
+    return response
+
+
+def api_get(ctx: click.core.Context, url: str, **kwargs):
+    return api_req(ctx, 'GET', url, **kwargs)
+
+
+def api_post(ctx: click.core.Context, url: str, **kwargs):
+    return api_req(ctx, 'POST', url, **kwargs)
+
+
+def highlight_json(obj):
+    formatted_json = json.dumps(obj, sort_keys=True, indent=4)
+    return highlight(formatted_json, lexers.JsonLexer(), formatters.TerminalFormatter())
+
+
+def echo_obj(obj):
+    click.echo(highlight_json(obj))
+
+
+@click.group()
+@click.option('--verbose', '-v', is_flag=True, help='Enables verbose output.')
+@click.option('--endpoint', '-e', default=API_ENDPOINT_URL, show_default=True)
+@click.pass_context
+def cli(ctx: click.core.Context, verbose: bool, endpoint: str):
+    ctx.obj = {
+        'verbose': verbose,
+        'endpoint': endpoint,
+        'url': lambda s: f'{endpoint}{s}'
+    }
+
+
+@cli.command('version')
+@click.pass_context
+def get_version(ctx: click.core.Context):
+    '''Gets the service version'''
+    response = api_get(ctx, '/')
+    echo_obj(response.json())
+
+
+@cli.command('list')
+@click.pass_context
+def list_jobs(ctx: click.core.Context):
+    '''List conversion jobs'''
+    response = api_get(ctx, '/jobs')
+    echo_obj(response.json())
+
+
+@cli.command('remove-vocals')
+@click.option('--file', type=click.File(mode='rb'), required=True)
+@click.option('--output-path', '-o', type=click.Path(writable=True))
+@click.pass_context
+def remove_vocals(ctx: click.core.Context, file: click.File, output_path: click.Path):
+    '''Removes the vocal part from an audio/video file'''
+    response = api_post(ctx, '/jobs')
+    job = response.json()
+    job_id = job['job_id']
+    upload_data = job['upload_data']
+    files = {'file': file}
+
+    upload_started = time.time()
+    response = requests.post(upload_data['url'],
+        data=upload_data['fields'],
+        files=files)
+
+    upload_time = time.time() - upload_started
+    print(f"File uploaded in {upload_time}s")
+
+    response = api_post(ctx, f'/jobs/{job_id}/process')
+    print(f'Processing file with job id {job_id}')
+
+    processing_started = time.time()
+
+    status_changed = False
+    job = None
+    while not status_changed:
+        response = api_get(ctx, f'/jobs/{job_id}')
+
+        job = response.json()
+        status_changed = job['status'] != 'processing'
+        time.sleep(5)
+
+    processing_time = time.time() - processing_started
+    print(f"File processed in {processing_time}s")
+
+    response = requests.get(job['output_url'], stream=True)
+    with open(output_path, 'wb') as f:
+        for chunk in response.raw:
+            f.write(chunk)
+
+    print(f"Saved file to: {output_path}")
+
+
+if __name__ == '__main__':
+    cli()

client/requirements.txt

@@ -0,0 +1,3 @@
+click
+pygments
+requests

worker/separator.py

@@ -21,6 +21,7 @@ OUTPUT_FILENAME_FORMAT = os.getenv('OUTPUT_FILENAME_FORMAT', '{instrument}.{code
 QUEUE_NAME = os.getenv('QUEUE_NAME')
 POLLING_INTERVAL = int(os.getenv('POLLING_INTERVAL', 5))
 OUTPUT_BUCKET_NAME = os.getenv('OUTPUT_BUCKET_NAME')
+OUTPUT_BUCKET_REGION= os.getenv('OUTPUT_BUCKET_REGION', 'eu-west-1')
 TRACKS_TABLE_NAME = os.getenv('TRACKS_TABLE_NAME')
 USE_MULTICHANNEL_WIENER_FILTERING = False
 
@@ -79,6 +80,7 @@ def poll_for_sqs_message(queue_name: str):
         body = json.loads(message['Body'])
         job_id = body['job_id']
         output_s3_url = f"s3://{OUTPUT_BUCKET_NAME}/track_{job_id}.mp3"
+        download_url = f'https://{OUTPUT_BUCKET_NAME}.s3-{OUTPUT_BUCKET_REGION}.amazonaws.com/track_{job_id}.mp3'
         logging.info(f'Start separating {job_id} -> {output_s3_url}')
 
         try:
@@ -94,7 +96,7 @@ def poll_for_sqs_message(queue_name: str):
                 Key={'id': {'S': job_id}},
                 AttributeUpdates={
                     'status': {'Value': {'S': 'successful'}},
-                    'output_url': {'Value': {'S': output_s3_url}}
+                    'output_url': {'Value': {'S': download_url}}
                 })
         except:
             logging.exception('Processing failed for some reason')