colab_essence

Google Colaboratory utilities.

Google Drive utilities

• mount_drive()

Mount Google Drive to Colab.

• to_drive(file_location, path="")

Move file from the local Colaboratory hard disk to your Google Drive storage.

Github utilities

• github_id()

Identify yourself to GitHub.

• clone_repo(repo_name, secret=False, user=None)

Clone a public or private GitHub repository.

Download utilities

• download(url, load_cookies=False, save_cookies=False)

Download a file from a URL. Save or load cookies for more involved downloads.

• big_drive_file(link)

Download Google Drive files (publicly shared with a link) that cannot be downloaded directly without confirmation because of their size (Google does not perform antivirus analysis on them).

Function utilities

• curry(func)

Create a curried version of a function that returns as soon as all the required arguments are passed.

• F(expr)

Given an expression as a string, it returns a curried function that when passed the needed arguments in alphabetical order, returns the result of the expression. Any variable used in the expression is a local argument of the function, unless the variable is in the global scope.

Run Visual Studio Code

• run_vscode()

Prepare the environment and run Visual Studio Code.

Other utilities

• env_vars("GITHUB_NAME", "GITHUB_PASSWORD", set_environ=False, env_file_location=f"{DRIVE_PATH}.env")

Return the value of the requested enviroment variables from a given .env file. If set_environ is True (False by default), load to os.environ and is available with os.environ[var_name] from anywhere in the current environment.

• hash_file(file_location, hash_fns={"MD5": hl.md5, "SHA1": hl.sha1}, buffer_size=4*GB)

• uncompress(file_location)

Extract compressed files (.zip, .tar, .tar.gz, .tar.bz2, .tar.lz, .tar.lzma, .tar.xz).

Things to do:

• Visualization recommender system.
• See what other self-hosted services can be added.
• Improve download (local file system, Google Sheet, SQL, BigQuery)
• Add numpy, Sympy, LaTeX alternatives to F
• What about transformational programming like Bird-Meertens
• Ask or generate password if not present in .env file
• Input widgets for all functions
• Boilerplate for TPUs
• Keras + FastAI
• Original Jupyter through SSH
• Wrapping for mobile devices with text transformation and extra imput row.
• Try improving the autocompletion using Tabnine or Kite
• Add auto formatting, refactoring, debugging, profiling utilities
• Add property based testing with Hypothesis