This page provides documentation for our command line tools.
stock CLI provides a git-like experience (whenever possible) to interact with
the stock repository. It also means that the current working directory is where
the stock repository would exist (like git 😊 ).
stock [OPTIONS] COMMAND [ARGS]...
--version display current stockroom version
It does a stock commit. Stock commit consists of two actions
- Make a hangar commit
- Update the
head.stockfile (git will track this file if you are using git)
stock commit [OPTIONS]
-m, --message TEXT The commit message. If multiple arguments are provided, each of them gets converted into a new line
Downloads and add a pytorch dataset (from torchvision, torchtext or torchaudio) to StockRoom. It creates the repo if it doesn't exist and loads the dataset into a repo for you
stock import [OPTIONS] DATASET_NAME
-d, --download-dir PATH If you have the dataset downloaded in a non-default path or want to download it to a non-default path, pass it here
Init stockroom repository. This will create a .hangar directory and a
file in your
stock init would be triggered implicitly if you are making
a stock repository by using
stock import but in all other case, you'd need to
initialize a stock repository to start operating on it with the python APIs
stock init [OPTIONS]
--username TEXT Username of the user --email TEXT Email address of the user --overwrite overwrite a repository if it exists at the current path
Release the writer lock forcefully and make the repository available for writing.
If another process, that has the writer lock, is writing to the repo, releasing the lock leads to an exception in that process. Use it carefully
stock liberate [OPTIONS]
Lists the items in the 3 shelves(Data, Model, Experiment) for the current stock head.
stock list [OPTIONS]
-d, --data Flag to show the data shelf and show a summary of the data -m, --model Flag to show the model shelf and list the models. -e, --experiment Flag to show the tags and their values