RawDet-7 Dataset Download Instructions

Option 1: Downloading Part Zip Files

This method involves downloading the dataset as a set of split zip archive parts. You will need to download all parts, combine them, and then extract the main archive.

The part zip files are located in the directory: part_zip_files/

Steps:

1. Download all part files:

   Create a directory for the parts and download them. For example:

   mkdir dataset_parts
   cd dataset_parts

   Download each part file (e.g., RawDet-7.zip.partaa, RawDet-7.zip.partab, etc.) into this directory. You can use wget for each file. Replace with actual file names and add all parts:

   wget https://data.dws.informatik.uni-mannheim.de/machinelearning/RawDet-7/part_zip_files/FILENAME.partaa
   wget https://data.dws.informatik.uni-mannheim.de/machinelearning/RawDet-7/part_zip_files/FILENAME.partab
   # ... continue for all parts

   Note: Please replace FILENAME.partaa etc. with the actual names of your part files. Ensure all parts are downloaded into the same directory. If there are many parts, you can list all URLs in a text file (e.g., urls.txt, one URL per line) and use wget -i urls.txt.

2. Combine the part files:

   In the directory containing all downloaded parts, combine them into a single archive. If the files were split using a command like split (e.g., resulting in .partaa, .partab suffixes):

   cat FILENAME.parta* > FullDataset.zip

   Ensure the wildcard (e.g., FILENAME.parta*) correctly matches all your part files in the correct order. The output file (FullDataset.zip) can be named as you wish.

3. Extract the combined zip file:

   Finally, unzip the combined archive:

   unzip FullDataset.zip

Option 2: Downloading Individual Files

This method allows you to download the dataset as individual files and folders, preserving the original directory structure. This is useful if you need only a subset of the data or prefer direct file access.

The individual files are located in the directory: individual_files/

Steps:

1. Download all files recursively:

   To download the entire dataset while preserving the directory structure under individual_files/, you can use wget. This command will create a local directory named individual_files containing the dataset.

   wget -r -np -nH --cut-dirs=2 -R "index.html*" https://data.dws.informatik.uni-mannheim.de/machinelearning/RawDet-7/individual_files/

   Explanation of wget options:

   • -r: Recursive download.
   • -np: No parent. Prevents ascending to parent directories.
   • -nH: No host directories. Disables generation of host-prefixed directories (e.g., data.dws.informatik.uni-mannheim.de/).
   • --cut-dirs=2: Ignores the first 2 components of the remote directory path (machinelearning/RawDet-7/), so the local structure starts with individual_files/.
   • -R "index.html*": Rejects (skips downloading) any index.html files, which are often server-generated directory listings.

2. Download specific files or subdirectories:

   If you only need specific files, you can browse the individual_files/ directory in your web browser and download files manually. Alternatively, use wget with the direct URL to a file or sub-directory:

   # To download a specific file:
   wget https://data.dws.informatik.uni-mannheim.de/machinelearning/RawDet-7/individual_files/path/to/your/file.ext
   
   # To download a specific subdirectory (add -P to specify local directory):
   wget -r -np -nH --cut-dirs=X -R "index.html*" https://data.dws.informatik.uni-mannheim.de/machinelearning/RawDet-7/individual_files/path/to/subdir/
   # Adjust --cut-dirs=X accordingly for the subdirectory URL.

Note: Recursive download relies on the server allowing directory listing or having a predictable structure.