{
  "id": 493680,
  "title": "Why not putting a single soundscape to the test directory for testing?",
  "url": "/competitions/birdclef-2024/discussion/493680",
  "author_name": "",
  "post_date": "2024-04-14T13:17:47.730068600Z",
  "votes": 11,
  "comment_count": 4,
  "views": 0,
  "content": "<p>This was the common practice in the previous BirdCLEF competitions (also across Kaggle) and it was pretty handy to have that single file in the test directory for testing purposes. Would it be possible to have it back instead of a README file?</p>",
  "messages": [
    {
      "id": "2751730",
      "postDate": "04/14/2024 13:17:47",
      "content": "<p>This was the common practice in the previous BirdCLEF competitions (also across Kaggle) and it was pretty handy to have that single file in the test directory for testing purposes. Would it be possible to have it back instead of a README file?</p>",
      "rawMarkdown": "This was the common practice in the previous BirdCLEF competitions (also across Kaggle) and it was pretty handy to have that single file in the test directory for testing purposes. Would it be possible to have it back instead of a README file?",
      "votes": null
    },
    {
      "id": "2752593",
      "postDate": "04/15/2024 03:49:06",
      "content": "<p>The 'unlabeled data' can be used for testing purposes; like the test data, it consists of 4m ogg files.</p>",
      "rawMarkdown": "The 'unlabeled data' can be used for testing purposes; like the test data, it consists of 4m ogg files.",
      "votes": null
    },
    {
      "id": "2753481",
      "postDate": "04/15/2024 14:04:43",
      "content": "<p>Yeah, we can use the different directories as well (train) but I just think that it was pretty handy to have that. Thanks.</p>",
      "rawMarkdown": "Yeah, we can use the different directories as well (train) but I just think that it was pretty handy to have that. Thanks.",
      "votes": null
    },
    {
      "id": "2757515",
      "postDate": "04/17/2024 14:53:07",
      "content": "<p>I agree, but you can test your code like that:</p>\n<pre><code>import \nimport soundfile\n\n = .()\n () == :\n    soundfile.(, np.zeros(( *  * _000, )), _000)\n     = []\n[], ()\n</code></pre>",
      "rawMarkdown": "I agree, but you can test your code like that:\n\n```\nimport glob\nimport soundfile\n\nfiles = glob.glob('/kaggle/input/birdclef-2024/test_soundscapes/*')\nif len(files) == 1:\n    soundfile.write('/kaggle/working/dummy.wav', np.zeros((4 * 60 * 32_000, )), 32_000)\n    files = ['/kaggle/working/dummy.wav']\nfiles[0], len(files)\n```",
      "votes": null
    },
    {
      "id": "2757733",
      "postDate": "04/17/2024 17:31:59",
      "content": "<p>Thank you for sharing!</p>",
      "rawMarkdown": "Thank you for sharing!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2752593,
      "author_name": "tomdenton",
      "author_url": "",
      "post_date": "04/15/2024 03:49:06",
      "content": "<p>The 'unlabeled data' can be used for testing purposes; like the test data, it consists of 4m ogg files.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2753481,
          "author_name": "snnclsr",
          "author_url": "",
          "post_date": "04/15/2024 14:04:43",
          "content": "<p>Yeah, we can use the different directories as well (train) but I just think that it was pretty handy to have that. Thanks.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2757515,
      "author_name": "martynoveduard",
      "author_url": "",
      "post_date": "04/17/2024 14:53:07",
      "content": "<p>I agree, but you can test your code like that:</p>\n<pre><code>import \nimport soundfile\n\n = .()\n () == :\n    soundfile.(, np.zeros(( *  * _000, )), _000)\n     = []\n[], ()\n</code></pre>",
      "votes": null,
      "replies": [
        {
          "id": 2757733,
          "author_name": "snnclsr",
          "author_url": "",
          "post_date": "04/17/2024 17:31:59",
          "content": "<p>Thank you for sharing!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2751730": "This was the common practice in the previous BirdCLEF competitions (also across Kaggle) and it was pretty handy to have that single file in the test directory for testing purposes. Would it be possible to have it back instead of a README file?",
    "2752593": "The 'unlabeled data' can be used for testing purposes; like the test data, it consists of 4m ogg files.",
    "2753481": "Yeah, we can use the different directories as well (train) but I just think that it was pretty handy to have that. Thanks.",
    "2757515": "I agree, but you can test your code like that:\n\n```\nimport glob\nimport soundfile\n\nfiles = glob.glob('/kaggle/input/birdclef-2024/test_soundscapes/*')\nif len(files) == 1:\n    soundfile.write('/kaggle/working/dummy.wav', np.zeros((4 * 60 * 32_000, )), 32_000)\n    files = ['/kaggle/working/dummy.wav']\nfiles[0], len(files)\n```",
    "2757733": "Thank you for sharing!"
  },
  "source": "meta"
}