Quickstart¶

This walks through your first storix session end to end. Everything here runs against real local disk, but the exact same code runs against any backend, which is the whole point.

Open a session¶

A session is anchored at a backend. LocalBackend("~/storix-data") makes that directory the root, so inside the session / means ~/storix-data and you can never wander above it.

from storix import Storix
from storix.backends import LocalBackend

fs = Storix(LocalBackend("~/storix-data"))
print(fs.pwd())   # / , the session starts at its home

Create and read¶

fs.mkdir("/notes")
fs.echo("buy milk", "/notes/todo.txt")

print(fs.cat("/notes/todo.txt"))   # b'buy milk'
print(fs.exists("/notes/todo.txt"))  # True

echo writes, cat reads. echo accepts native Python: bytes, str, or an iterator of chunks (more on that in Reading and writing).

Navigate like a shell¶

A session has a working directory, so relative paths and cd work exactly as in a terminal.

fs.cd("/notes")
print(fs.pwd())          # /notes
print(fs.ls())           # [StorixPath('todo.txt')]
fs.echo("...", "idea.txt")   # relative to the cwd -> /notes/idea.txt

ls hides dotfiles by default, like the real thing. Pass all=True to show them.

Move, copy, remove¶

Move and copy are variadic, and the last argument is the destination, just like unix mv and cp.

fs.mkdir("/archive", parents=True)  # mkdir -p; safe if it already exists
fs.mv("/notes/todo.txt", "/notes/idea.txt", "/archive")  # both into /archive
fs.rm("/notes", recursive=True)                          # rm -r

Every one of these raises a typed error on failure (for example PathNotFoundError), so you never have to check a boolean return.

Swap the backend, keep the code¶

Here is the payoff. Point the same code at the in-memory backend, and nothing else changes. This is what makes tests fast and hermetic.

from storix.backends import MemoryBackend

fs = Storix(MemoryBackend())   # nothing on disk
fs.mkdir("/notes")
fs.echo("buy milk", "/notes/todo.txt")
print(fs.cat("/notes/todo.txt"))   # b'buy milk'

Swap MemoryBackend() for AzureBackend(...) and the same code talks to Azure Data Lake. See Backends.

The async flavor¶

Every operation has an awaitable twin under storix.aio, with identical names. Storix is async-first, and the sync package is generated from that source so the two never diverge.

from storix.aio import Storix
from storix.aio.backends import MemoryBackend

async def main():
    async with Storix(MemoryBackend()) as fs:
        await fs.echo("buy milk", "/todo.txt")
        print(await fs.ls())
        print(await fs.cat("/todo.txt"))

Where to next¶

• Reading and writing: streaming, iterators, and keeping memory bounded on large files.
• Backends: the port, configuration, and going to the cloud.
• Layers: sandboxing, caching, and capabilities.