aiofs 0.3.0

Wrap data sources like files with asyncio python

Python Anyncronous File System

There are a broad variety of data sources in the current software: sql databases, files, no-sql databases, cloud storages, ... and every one has its own methods and ways to accesing the data.

A good software architecture should not be have a strong dependency to these concrete methods and this is the purpose of this library: Abstract the acccess to data sources from a software application

There already are som libraries for this purpose in python (like filesystem ) but it is not fully oriented to asyncronous programming. It has taken as reference the library aiofiles as pattern to define the system

Usage

There are only two basic classes which wrapp any data source:

1. FileLike, which is an object with basic file methods:
   1. Binary read method to access to data bytes of the file.
   2. Asyncronous binary write method for writing bytes to the file.
   3. Context management: read and write shall be ran within a context for assuring the proper closing and handling of file inside filesystem.
2. FileLikeSystem, which is a file system for the files. They have the following methods:
   1. open is the main method because it is the way to create and access to FileLike object. It is very important to have clear the future handling of the file:
      1. If the file will be used to only read, the mode should be r. This the default mode.
      2. If the file exist and will be written, the mode should be r+ if you will have a reading + writing process or w if it only will be written.
3. rm to remove one or some files
4. ls to list the filenames accesibles in the file system.

There are the following possible exception:

• BlockingIOError if two clients are writing at the same time
• FileNotFound if the file does not exist in the file system

Installation

Depending of what data source will be used, it is necessary to define extras:

1. If you are using a Operating File System it is not necessary any extras
2. If you are using Redis Data System, you'll need to add redis extra. The directory structure will be stored within the name of the variable. Ex. directory/path/filename.bin will be mapped as directory:path:filename.bin.
3. If you are using Azure blobs, you will need to add extra azure. The directory structure will go directly to the blob name. The file system is mapped to an unique blob container.

An example of manual installation for azure environment could be: pip install aiofs[azure]

TODOS

• It is intended to access to random access to files.
• Amplify the file system methods to a better file handling
• Add new data sources