{ "id": 446102, "title": "Why is there no \"is_tma\" in the test csv?", "url": "/competitions/UBC-OCEAN/discussion/446102", "author_name": "Jonas Gjerris", "post_date": "2023-10-10T09:15:08.668000", "votes": 3, "comment_count": 7, "views": 0, "content": "

Can anyone explain the omission of \"is_tma\" from the test csv? Although the information seems discernible—if the image_id isn't in the test_thumbnails folder, then the image is a tma—it feels unnecessarily complicated. Am I missing something?

", "messages": [ { "id": 2482591, "postDate": "2023-10-15T06:45:58.973Z", "content": "

Yeah it doesn't make sense. You can basically do

\n
df_test[] = \ndf_test.loc[df_test[].isin(test_thumbnails), ] = \n
", "rawMarkdown": "Yeah it doesn't make sense. You can basically do\n\n```python\ndf_test['is_tma'] = True\ndf_test.loc[df_test['image_id'].isin(test_thumbnails), 'is_tma'] = False\n```\n", "votes": 3, "replies": [ { "id": 2483025, "postDate": "2023-10-15T12:31:17.727Z", "content": "

I was starting to complain but this is a very nice fix, thank you.

", "rawMarkdown": "I was starting to complain but this is a very nice fix, thank you." }, { "id": 2483073, "postDate": "2023-10-15T13:03:06.863Z", "content": "

I had to do smth like this tho

\n
\ntest_thumbnails_folder = \n\n\nthumbnails_filenames = os.listdir(test_thumbnails_folder)\n\n\nimage_ids_in_thumbnails = [filename.split()[]  filename  thumbnails_filenames]\n\n\ntest_df[] = ~test_df[].isin(image_ids_in_thumbnails)\n
", "rawMarkdown": "I had to do smth like this tho\n\n```python\n# Replace 'test_thumbnails' with the actual path to your folder\ntest_thumbnails_folder = '/kaggle/input/UBC-OCEAN/test_thumbnails'\n\n# Get a list of filenames in the folder\nthumbnails_filenames = os.listdir(test_thumbnails_folder)\n\n# Extract image IDs from filenames\nimage_ids_in_thumbnails = [filename.split('_')[0] for filename in thumbnails_filenames]\n\n# Set 'is_tma' in test_df based on image IDs\ntest_df['is_tma'] = ~test_df['image_id'].isin(image_ids_in_thumbnails)\n```", "votes": 2 }, { "id": 2546914, "postDate": "2023-12-03T01:43:05.963Z", "content": "

Does this still work? It seems that all test images have thumbnails. Am I wrong?

", "rawMarkdown": "Does this still work? It seems that all test images have thumbnails. Am I wrong?", "replies": [ { "id": 2552639, "postDate": "2023-12-07T16:18:55.973Z", "content": "

Did you figure it out?

", "rawMarkdown": "Did you figure it out?" }, { "id": 2562040, "postDate": "2023-12-15T03:59:58.570Z", "content": "

After some testing, it seem there is no TMA in the public LB test data. Johnny, right?

", "rawMarkdown": "After some testing, it seem there is no TMA in the public LB test data. Johnny, right?" }, { "id": 2574595, "postDate": "2023-12-26T05:22:41.900Z", "content": "

you can judge by the width and height

", "rawMarkdown": "you can judge by the width and height" } ] } ] }, { "id": 2476001, "postDate": "2023-10-10T09:15:08.667Z", "content": "

Can anyone explain the omission of \"is_tma\" from the test csv? Although the information seems discernible—if the image_id isn't in the test_thumbnails folder, then the image is a tma—it feels unnecessarily complicated. Am I missing something?

", "rawMarkdown": "Can anyone explain the omission of \"is_tma\" from the test csv? Although the information seems discernible—if the image_id isn't in the test_thumbnails folder, then the image is a tma—it feels unnecessarily complicated. Am I missing something?", "votes": 3 } ], "comments": [ { "id": 2482591, "author_name": "Gunes Evitan", "author_url": "", "post_date": "2023-10-15T06:45:58.973000", "content": "

Yeah it doesn't make sense. You can basically do

\n
df_test[] = \ndf_test.loc[df_test[].isin(test_thumbnails), ] = \n
", "votes": 3, "replies": [ { "id": 2483025, "author_name": "Parham Gousheh", "author_url": "", "post_date": "2023-10-15T12:31:17.727000", "content": "

I was starting to complain but this is a very nice fix, thank you.

", "votes": 0, "replies": [] }, { "id": 2483073, "author_name": "Parham Gousheh", "author_url": "", "post_date": "2023-10-15T13:03:06.863000", "content": "

I had to do smth like this tho

\n
\ntest_thumbnails_folder = \n\n\nthumbnails_filenames = os.listdir(test_thumbnails_folder)\n\n\nimage_ids_in_thumbnails = [filename.split()[]  filename  thumbnails_filenames]\n\n\ntest_df[] = ~test_df[].isin(image_ids_in_thumbnails)\n
", "votes": 2, "replies": [] }, { "id": 2546914, "author_name": "Johnny Lee", "author_url": "", "post_date": "2023-12-03T01:43:05.963000", "content": "

Does this still work? It seems that all test images have thumbnails. Am I wrong?

", "votes": 0, "replies": [ { "id": 2552639, "author_name": "Jonas Gjerris", "author_url": "", "post_date": "2023-12-07T16:18:55.973000", "content": "

Did you figure it out?

", "votes": 0, "replies": [] }, { "id": 2562040, "author_name": "Yoobin", "author_url": "", "post_date": "2023-12-15T03:59:58.570000", "content": "

After some testing, it seem there is no TMA in the public LB test data. Johnny, right?

", "votes": 0, "replies": [] }, { "id": 2574595, "author_name": "Toby", "author_url": "", "post_date": "2023-12-26T05:22:41.900000", "content": "

you can judge by the width and height

", "votes": 0, "replies": [] } ] } ] } ], "raw_markdown_by_id": { "2482591": "Yeah it doesn't make sense. You can basically do\n\n```python\ndf_test['is_tma'] = True\ndf_test.loc[df_test['image_id'].isin(test_thumbnails), 'is_tma'] = False\n```\n", "2476001": "Can anyone explain the omission of \"is_tma\" from the test csv? Although the information seems discernible—if the image_id isn't in the test_thumbnails folder, then the image is a tma—it feels unnecessarily complicated. Am I missing something?" } }