`stor` provides a cross-compatible CLI and Python API for accessing block and
object storage. `stor` was created so you could write one piece of code to work
with local or remote files, without needing to write specialized code to handle
failure modes, retrying or temporarily system unavailability. The functional
API (i.e., `stor.copytree`, `stor.rmtree`, `stor.remove`, `stor.listdir`) will
work with the same semantics across all storage backends. This makes it really
easy to develop/test code locally with files and then take advantage of robust
and cheaper object storage when you push to remote.
View full docs for stor at https://counsyl.github.io/stor/ .
pip install stor
`stor` provides both a CLI and a Python library for manipulating Posix and OBS
with a single, cross-compatible API.
## Quickstart - CLI
usage: stor [-h] [-c CONFIG_FILE] [--version] ...
A command line interface for stor.
list List contents using the path as a prefix.
ls List path as a directory.
cp Alias for copy.
rm Remove file at a path.
walkfiles List all files under a path that match an optional
cat Output file contents to stdout.
cd Change directory to a given OBS path.
pwd Get the present working directory of a service or all
clear Clear current directories of a specified service.
You can `ls` local and remote directories
›› stor ls s3://stor-test-bucket
Copy files locally or remotely or upload from stdin
›› echo "HELLO WORLD" | stor cp - swift://AUTH_stor_test/hello_world.txt
starting upload of 1 objects
upload complete - 1/1 0:00:00 0.00 MB 0.00 MB/s
›› stor cat swift://AUTH_stor_test/hello_world.txt
›› stor cp swift://AUTH_stor_test/hello_world.txt hello_world.txt
›› stor cat hello_world.txt
## Quickstart - Python
List files in a directory, taking advantage of delimiters
List all objects in a bucket
Or in a local path
Read and write files from POSIX or OBS, using python file objects.
with stor.open('/my/exciting.json') as fp:
data1 = json.load(fp)
data1['read'] = True
with stor.open('s3://bestbucket/exciting.json') as fp:
## Testing code that uses stor
The key design consideration of `stor` is that your code should be able to
transparently use POSIX or any object storage system to read and update files.
So, rather than use mocks, we suggest that you structure your test code to point
to local filesystem paths and restrict yourself to the functional API. E.g.,
in your prod settings, you could set `DATADIR = 's3://bestbucketever'`and when
you test, you could use `DATADIR = '/somewhat/cool/path/to/test/data'`, while
your actual code just says:
with stor.open(stor.join(DATADIR, experiment)) as fp:
data = json.load(fp)
Easy! and no mocks required!
## Running the Tests
## Contributing and Semantic Versioning
We use semantic versioning to communicate when we make API changes to the
library. See CONTRIBUTING.md for more details on contributing to stor.
TODO: Brief introduction on what you do with files - including link to relevant help section.