Datasets installation¶
Dataset installation is triggered with the datasets install subcommand
Output
Usage: tcbench datasets install [OPTIONS]
Install a dataset.
╭─ Options ───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
│ * --name -n [ucdavis-icdm19|utmobilenet21|mirage19|mirage22] Dataset to install. [required] │
│ --input-folder -i PATH Folder where to find pre-downloaded tarballs. │
│ --help Show this message and exit. │
╰─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯
The raw data of the datasets is either hosted on websites or cloud environments. The automatic download from those locations is available only for some of the datasets.
| Name | Auto download |
|---|---|
ucdavis-icdm19 |
|
mirage19 |
|
mirage22 |
|
utmobilenet21 |
If auto download is not possible, to install the dataset
you need to manually fetch the related archives, place them
in a folder, e.g., /download, and provide the --input-folder
option when triggering installation.
When installing a dataset, tcbench also
shows two types of reports as formatted tables.
-
Samples count: This tables collect the number of samples (i.e., flows) available.
-
Stats: The curation process can filter out flows (e.g., based on a minum number of packets or remove classes without a minimum number of flows). As such, when installing,
tcbenchis showing general stats (mean, std, percentiles) about number of packets for each flow across classes.
Please check the specific install page for each dataset for more details.
Datasets deletion¶
The datasets files are installed within the python environment where tcbench is installed.
You can delete a dataset using the following command