Why cloudpathlib?¶

We 😍 pathlib¶

pathlib a wonderful tool for working with filesystem paths, available from the Python 3 standard library.

In [1]:

from pathlib import Path

from pathlib import Path

For example, we can easily list all the files in a directory.

In [2]:

list(Path(".").glob("*"))

list(Path(".").glob("*"))

Out[2]:

[PosixPath('caching.ipynb'),
 PosixPath('why_cloudpathlib.ipynb'),
 PosixPath('api-reference'),
 PosixPath('.ipynb_checkpoints'),
 PosixPath('logo-no-text.svg'),
 PosixPath('logo.svg'),
 PosixPath('authentication.ipynb'),
 PosixPath('favicon.svg'),
 PosixPath('stylesheets')]

There are methods to quickly learn everything there is to know about a filesystem path, and even do simple file manipulations.

In [3]:

notebook = Path("why_cloudpathlib.ipynb").resolve()

print(f"{'Path:':15}{notebook}")
print(f"{'Name:':15}{notebook.name}")
print(f"{'Stem:':15}{notebook.stem}")
print(f"{'Suffix:':15}{notebook.suffix}")
print(f"{'With suffix:':15}{notebook.with_suffix('.cpp')}")
print(f"{'Parent:':15}{notebook.parent}")
print(f"{'Read_text:'}\n{notebook.read_text()[:200]}\n")

notebook = Path("why_cloudpathlib.ipynb").resolve() print(f"{'Path:':15}{notebook}") print(f"{'Name:':15}{notebook.name}") print(f"{'Stem:':15}{notebook.stem}") print(f"{'Suffix:':15}{notebook.suffix}") print(f"{'With suffix:':15}{notebook.with_suffix('.cpp')}") print(f"{'Parent:':15}{notebook.parent}") print(f"{'Read_text:'}\n{notebook.read_text()[:200]}\n")

Path:          /Users/bull/code/cloudpathlib/docs/docs/why_cloudpathlib.ipynb
Name:          why_cloudpathlib.ipynb
Stem:          why_cloudpathlib
Suffix:        .ipynb
With suffix:   /Users/bull/code/cloudpathlib/docs/docs/why_cloudpathlib.cpp
Parent:        /Users/bull/code/cloudpathlib/docs/docs
Read_text:
{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Why cloudpathlib?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## We 😍 pathl

If you're new to pathlib, we highly recommend it over the older os.path module. We find that it has a much more intuitive and convenient interface. The official documentation is a helpful reference, and we also recommend this excellent cheat sheet by Chris Moffitt.

Cross-platform support¶

One great feature about using pathlib over regular strings is that it lets you write code with cross-platform file paths. It "just works" on Windows too. Write path manipulations that can run on anyone's machine!

path = Path.home()
path
>>> C:\Users\DrivenData\

docs = path / 'Documents'
docs
>>> C:\Users\DrivenData\Documents

We also 😍 cloud storage¶

This is great, but I live in the future. Not every file I care about is on my machine. What do I do when I am working on S3? Do I have to explicitly download every file before I can do things with them?

Of course not, if you use cloudpathlib!

In [4]:

# load environment variables from .env file;
# not required, just where we keep our creds
from dotenv import load_dotenv, find_dotenv

load_dotenv(find_dotenv())

# load environment variables from .env file; # not required, just where we keep our creds from dotenv import load_dotenv, find_dotenv load_dotenv(find_dotenv())

Out[4]:

True

In [5]:

from cloudpathlib import S3Path

s3p = S3Path("s3://cloudpathlib-test-bucket/why_cloudpathlib/file.txt")
s3p.name

from cloudpathlib import S3Path s3p = S3Path("s3://cloudpathlib-test-bucket/why_cloudpathlib/file.txt") s3p.name

Out[5]:

'file.txt'

In [6]:

# Nothing there yet...
s3p.exists()

# Nothing there yet... s3p.exists()

Out[6]:

False

In [7]:

# Touch (just like with `pathlib.Path`)
s3p.touch()

# Touch (just like with `pathlib.Path`) s3p.touch()

In [8]:

# Bingo!
s3p.exists()

# Bingo! s3p.exists()

Out[8]:

True

In [9]:

# list all the files in the directory
[p for p in s3p.parent.iterdir()]

# list all the files in the directory [p for p in s3p.parent.iterdir()]

Out[9]:

[S3Path('s3://cloudpathlib-test-bucket/why_cloudpathlib/file.txt')]

In [10]:

stat = s3p.stat()
print(f"File size in bytes: {stat.st_size}")
stat

stat = s3p.stat() print(f"File size in bytes: {stat.st_size}") stat

File size in bytes: 0

Out[10]:

os.stat_result(st_mode=None, st_ino=None, st_dev='s3://', st_nlink=None, st_uid=None, st_gid=None, st_size=0, st_atime=None, st_mtime=1601853143.0, st_ctime=None)

In [11]:

s3p.write_text("Hello to all of my friends!")

s3p.write_text("Hello to all of my friends!")

Out[11]:

27

In [12]:

stat = s3p.stat()
print(f"File size in bytes: {stat.st_size}")
stat

stat = s3p.stat() print(f"File size in bytes: {stat.st_size}") stat

File size in bytes: 27

Out[12]:

os.stat_result(st_mode=None, st_ino=None, st_dev='s3://', st_nlink=None, st_uid=None, st_gid=None, st_size=27, st_atime=None, st_mtime=1601853144.0, st_ctime=None)

In [13]:

# Delete (again just like with `pathlib.Path`)
s3p.unlink()

# Delete (again just like with `pathlib.Path`) s3p.unlink()

In [14]:

s3p.exists()

s3p.exists()

Out[14]:

False

Cross-cloud support¶

That's cool, but I use Azure Blob Storage―what can I do?

In [15]:

from cloudpathlib import AzureBlobPath

azp = AzureBlobPath("az://cloudpathlib-test-container/file.txt")
azp.name

from cloudpathlib import AzureBlobPath azp = AzureBlobPath("az://cloudpathlib-test-container/file.txt") azp.name

Out[15]:

'file.txt'

In [16]:

azp.exists()

azp.exists()

Out[16]:

False

In [17]:

azp.write_text("I'm on Azure, boss.")

azp.write_text("I'm on Azure, boss.")

Out[17]:

19

In [18]:

azp.exists()

azp.exists()

Out[18]:

True

In [19]:

# list all the files in the directory
[p for p in azp.parent.iterdir()]

# list all the files in the directory [p for p in azp.parent.iterdir()]

Out[19]:

[AzureBlobPath('az://cloudpathlib-test-container/file.txt')]

In [20]:

azp.exists()

azp.exists()

Out[20]:

True

In [21]:

azp.unlink()

azp.unlink()

Cloud hopping¶

Moving between cloud storage providers should be a simple as moving between disks on your computer. Let's say that the Senior Vice President of Tomfoolery comes to me and says, "We've got a mandate to migrate our application to Azure Blob Storage from S3!"

No problem, if I used cloudpathlib! The CloudPath class constructor automatically dispatches to the appropriate concrete class, the same way that pathlib.Path does for different operating systems.

In [22]:

from cloudpathlib import CloudPath

cloud_directory = CloudPath("s3://cloudpathlib-test-bucket/why_cloudpathlib/")

upload = cloud_directory / "user_upload.txt"
upload.write_text("A user made this file!")

assert upload.exists()
upload.unlink()
assert not upload.exists()

from cloudpathlib import CloudPath cloud_directory = CloudPath("s3://cloudpathlib-test-bucket/why_cloudpathlib/") upload = cloud_directory / "user_upload.txt" upload.write_text("A user made this file!") assert upload.exists() upload.unlink() assert not upload.exists()

In [23]:

from cloudpathlib import CloudPath

# Changing this root path is the ONLY change!
cloud_directory = CloudPath("az://cloudpathlib-test-container/why_cloudpathlib/")

upload = cloud_directory / "user_upload.txt"
upload.write_text("A user made this file!")

assert upload.exists()
upload.unlink()
assert not upload.exists()

from cloudpathlib import CloudPath # Changing this root path is the ONLY change! cloud_directory = CloudPath("az://cloudpathlib-test-container/why_cloudpathlib/") upload = cloud_directory / "user_upload.txt" upload.write_text("A user made this file!") assert upload.exists() upload.unlink() assert not upload.exists()