gmberton / VPR-datasets-downloader

Automatic download VPR datasets in a standard format
MIT License
119 stars 18 forks source link

nordland download url is out of date #11

Closed BanaBear closed 6 months ago

BanaBear commented 8 months ago

hello, i found that the url is out of date in download_nordland.py and i can't find the cleanImageNames.txt, could you fix this problem? thanks!!

util1.download_heavy_file("https://cloudstor.aarnet.edu.au/plus/s/8L7loyTZjK0FsWT/download?path=%2F&files=summer.tar.gz", join(raw_data_folder, "summer.tar.gz")) util1.download_heavy_file("https://cloudstor.aarnet.edu.au/plus/s/8L7loyTZjK0FsWT/download?path=%2F&files=winter.tar.gz", join(raw_data_folder, "winter.tar.gz")) util1.download_heavy_file("https://cloudstor.aarnet.edu.au/plus/s/8L7loyTZjK0FsWT/download?path=%2F&files=cleanImageNames.txt&downloadStartSecret=crd03ou9qji", join(raw_data_folder, "cleanImageNames.txt"))

gmberton commented 8 months ago

Hi, you can download the Nordland dataset from here. I haven't been able to automatize the download and format due to an issue of gdown when downloading large files, so you need to download summer and winter splits of Nordland from browser and then adapt the script. I will fix the script once the issue is fixed on gdown

Fragmentszz commented 7 months ago

but still I can't find where to get cleanImageNames.txt :(

Fragmentszz commented 7 months ago

oh, may the files can be download from here

oravus commented 7 months ago

If you are looking for that CloudStor hosted version of Nordland, it's here now. For download commands, please see this.

gmberton commented 7 months ago

Hi @oravus, thank you for your help. I tried to fix the download script using the links you provided, but the winter.tar.gz seems corrupted. I tried multiple times to download it and extract with multiple libraries, but I'm getting an error and only 8970 out of 35768 are downloaded. For example, when extracting with tar -xvzf winter.tar.gz I get this error (after thousands of images are correctly extracted):

winter/images-34312.png
winter/images-34373.png
winter/images-34381.png
tar: Skipping to next header

gzip: stdin: invalid compressed data--format violated
tar: Child returned status 1
tar: Error is not recoverable: exiting now

The summer tar is fine. Could you have a look at the uploaded winter tar?

oravus commented 7 months ago

I think I have encountered this extraction issue before. Probably ideal to reupload the winter tar, will message here once that's done.

oravus commented 6 months ago

@gmberton, Winter (and Spring) tars are now updated, thanks to @Ahmedest61. We have tested it at our end for download/extraction. It would be great if you can also test your script.

Also, to test if the downloaded file is correct, you can run sha256sum winter.tar.gz It should return c95a1cf99ef2cdcf9e1d9b47de1d8248a194a6c8deeb976edd099562ab20e1c2

gmberton commented 6 months ago

Thank you @oravus , the links work well now :)