{
  "id": 238712,
  "title": "Can anyone help me? Why is my score 0?",
  "url": "/competitions/coleridgeinitiative-show-us-the-data/discussion/238712",
  "author_name": "",
  "post_date": "2021-05-13T07:19:07.465891400Z",
  "votes": 4,
  "comment_count": 5,
  "views": 0,
  "content": "<p>This is my notebook:<br>\n<a href=\"https://www.kaggle.com/paoloantonuccio/notebooka5c6393d44?scriptVersionId=62734993\" target=\"_blank\">https://www.kaggle.com/paoloantonuccio/notebooka5c6393d44?scriptVersionId=62734993</a></p>",
  "messages": [
    {
      "id": "1305254",
      "postDate": "05/13/2021 07:19:07",
      "content": "<p>This is my notebook:<br>\n<a href=\"https://www.kaggle.com/paoloantonuccio/notebooka5c6393d44?scriptVersionId=62734993\" target=\"_blank\">https://www.kaggle.com/paoloantonuccio/notebooka5c6393d44?scriptVersionId=62734993</a></p>",
      "rawMarkdown": "This is my notebook:\nhttps://www.kaggle.com/paoloantonuccio/notebooka5c6393d44?scriptVersionId=62734993",
      "votes": null
    },
    {
      "id": "1306892",
      "postDate": "05/14/2021 06:09:33",
      "content": "<p>that could be because your code do not make enough number of dataset label prediction for file<br>\n2f392438-e215-4169-bebf-21ac4ff253e1<br>\n3f316b38-1a24-45a9-8d8c-4e05a42257c6</p>",
      "rawMarkdown": "that could be because your code do not make enough number of dataset label prediction for file\n2f392438-e215-4169-bebf-21ac4ff253e1\n3f316b38-1a24-45a9-8d8c-4e05a42257c6",
      "votes": null
    },
    {
      "id": "1307661",
      "postDate": "05/14/2021 15:32:33",
      "content": "<p>You can't use only the 4 ids from the sample submissions, you have to get the  ids from the files names: </p>\n<p>Something like this:</p>\n<pre><code>test_files=os.listdir(test_files_path)\ntest_files=[test_files[i].split('.')[0] for i in range(len(test_files))]\ntest_df=pd.DataFrame(test_files,columns=['Id'])\ntest_df['text'] = test['Id'].apply(get_text)\n</code></pre>\n<p>This code will generate the same results in your notebook but when they re-run it privately, will get over 8000 documents that are hidden from us.</p>\n<p>Hope this works</p>",
      "rawMarkdown": "You can't use only the 4 ids from the sample submissions, you have to get the  ids from the files names: \n\nSomething like this:\n\n```\ntest_files=os.listdir(test_files_path)\ntest_files=[test_files[i].split('.')[0] for i in range(len(test_files))]\ntest_df=pd.DataFrame(test_files,columns=['Id'])\ntest_df['text'] = test['Id'].apply(get_text)\n```\n\nThis code will generate the same results in your notebook but when they re-run it privately, will get over 8000 documents that are hidden from us.\n\nHope this works",
      "votes": null
    },
    {
      "id": "1308297",
      "postDate": "05/15/2021 05:42:58",
      "content": "<p>I get it. Their test data are different from the submission data! Thanks, it will help me!</p>",
      "rawMarkdown": "I get it. Their test data are different from the submission data! Thanks, it will help me!",
      "votes": null
    },
    {
      "id": "1308357",
      "postDate": "05/15/2021 06:28:36",
      "content": "<p>But with this code it gives me a score of 0.532</p>\n<p>for i in tqdm(range(len(submission))):<br>\n    text_id = submission.iloc[i,0]<br>\n    text = clean_text(get_text(text_id, test_path))<br>\n    submission.loc[i,\"PredictionString\"] = find_dataset(text)</p>",
      "rawMarkdown": "But with this code it gives me a score of 0.532\n\nfor i in tqdm(range(len(submission))):\n    text_id = submission.iloc[i,0]\n    text = clean_text(get_text(text_id, test_path))\n    submission.loc[i,\"PredictionString\"] = find_dataset(text)",
      "votes": null
    },
    {
      "id": "1308368",
      "postDate": "05/15/2021 06:33:20",
      "content": "<p>I think you gave me the right answer! Thank you!</p>",
      "rawMarkdown": "I think you gave me the right answer! Thank you!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1306892,
      "author_name": "ritesh256",
      "author_url": "",
      "post_date": "05/14/2021 06:09:33",
      "content": "<p>that could be because your code do not make enough number of dataset label prediction for file<br>\n2f392438-e215-4169-bebf-21ac4ff253e1<br>\n3f316b38-1a24-45a9-8d8c-4e05a42257c6</p>",
      "votes": null,
      "replies": [
        {
          "id": 1308368,
          "author_name": "paoloantonuccio",
          "author_url": "",
          "post_date": "05/15/2021 06:33:20",
          "content": "<p>I think you gave me the right answer! Thank you!</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1307661,
      "author_name": "sneyderyepes",
      "author_url": "",
      "post_date": "05/14/2021 15:32:33",
      "content": "<p>You can't use only the 4 ids from the sample submissions, you have to get the  ids from the files names: </p>\n<p>Something like this:</p>\n<pre><code>test_files=os.listdir(test_files_path)\ntest_files=[test_files[i].split('.')[0] for i in range(len(test_files))]\ntest_df=pd.DataFrame(test_files,columns=['Id'])\ntest_df['text'] = test['Id'].apply(get_text)\n</code></pre>\n<p>This code will generate the same results in your notebook but when they re-run it privately, will get over 8000 documents that are hidden from us.</p>\n<p>Hope this works</p>",
      "votes": null,
      "replies": [
        {
          "id": 1308297,
          "author_name": "paoloantonuccio",
          "author_url": "",
          "post_date": "05/15/2021 05:42:58",
          "content": "<p>I get it. Their test data are different from the submission data! Thanks, it will help me!</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1308357,
          "author_name": "paoloantonuccio",
          "author_url": "",
          "post_date": "05/15/2021 06:28:36",
          "content": "<p>But with this code it gives me a score of 0.532</p>\n<p>for i in tqdm(range(len(submission))):<br>\n    text_id = submission.iloc[i,0]<br>\n    text = clean_text(get_text(text_id, test_path))<br>\n    submission.loc[i,\"PredictionString\"] = find_dataset(text)</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1305254": "This is my notebook:\nhttps://www.kaggle.com/paoloantonuccio/notebooka5c6393d44?scriptVersionId=62734993",
    "1306892": "that could be because your code do not make enough number of dataset label prediction for file\n2f392438-e215-4169-bebf-21ac4ff253e1\n3f316b38-1a24-45a9-8d8c-4e05a42257c6",
    "1307661": "You can't use only the 4 ids from the sample submissions, you have to get the  ids from the files names: \n\nSomething like this:\n\n```\ntest_files=os.listdir(test_files_path)\ntest_files=[test_files[i].split('.')[0] for i in range(len(test_files))]\ntest_df=pd.DataFrame(test_files,columns=['Id'])\ntest_df['text'] = test['Id'].apply(get_text)\n```\n\nThis code will generate the same results in your notebook but when they re-run it privately, will get over 8000 documents that are hidden from us.\n\nHope this works",
    "1308297": "I get it. Their test data are different from the submission data! Thanks, it will help me!",
    "1308357": "But with this code it gives me a score of 0.532\n\nfor i in tqdm(range(len(submission))):\n    text_id = submission.iloc[i,0]\n    text = clean_text(get_text(text_id, test_path))\n    submission.loc[i,\"PredictionString\"] = find_dataset(text)",
    "1308368": "I think you gave me the right answer! Thank you!"
  },
  "source": "meta"
}