Trunk
A file attachment/storage library for Elixir.
Installation
Add :trunk
to your list of dependencies in mix.exs
:
def deps do
[
{:trunk, "~> 1.0"},
# If you want to use Amazon S3, then add:
{:ex_aws_s3, "~> 2.0"},
{:hackney, "~> 1.7"},
{:poison, "~> 3.1"},
{:sweet_xml, "~> 0.6"},
]
end
Trunk has only one hard dependency on Briefly to handle temporary file creation (and auto-destruction)
Usage
Trunk is a behaviour that is implemented in a module. use Trunk
creates functions for storing, deleting, and generating urls for your files and their versions. It then implements callbacks in the simplest way possible. You can then override the callbacks you need to as you want to extend the behaviour.
defmodule MyTrunk do
use Trunk, versions: [:original, :thumb]
# override callbacks as needed.
end
Configuration
Trunk has been designed to be highly configurable. It can be configured in stages with each level merging with the level before it.
See the documentation for all config options.
Global configuration
config :trunk,
storage: Trunk.Storage.Filesystem,
storage_opts: [path: "/tmp"]
App specific configuration for umbrella type configs
config :my_app, :trunk,
storage: Trunk.Storage.S3,
storage_opts: [bucket: "test-trunk"]
in order for these options to be used, you need to pass the otp_app
option when calling use Trunk
as follows:
defmodule MyTrunk do
use Trunk, otp_app: :my_app
end
Module configuration
defmodule MyTrunk do
use Trunk, versions: [:original, :trunk],
storage: Trunk.Storage.Filesystem,
storage_opts: [path: "/tmp"]
end
Function options
Caution: If you override options during the storage call, you need to be sure to pass the same options to other calls
MyTrunk.store("/path/to/file.ext", storage: Trunk.Storage.S3, storage_opts: [bucket: "test-trunk"])
Storage
Storage is handled by a behaviour.
Two storage handlers are included: Trunk.Storage.Filesystem, and Trunk.Storage.S3 Additional storage systems can be handled by creating a module that implements the Trunk.Storage behaviour.
When storing files, Trunk runs the file through a transformation pipeline allowing you to generate different versions of a file. Each stage in the pipeline is handled via a callback allowing you to configure what transformations take place, where the version is stored and how it is named. Full information can be found in the documentation
Scope
You have the option of passing a scope (usually a struct or map) into the transform functions. This scope object will then be available in each callback allowing you to further customise the handling of each version.
Example:
defmodule MyTrunk do
use Trunk, versions: [:thumb]
def storage_dir(%Trunk.State{scope: %{id: model_id}}, :thumb)
do: "my_models/#{model_id}"
end
MyTrunk.store("/path/to/file.ext", %{id: 42})
# will save to <storage>/my_models/42/<filename>
State
State about the file and the version transformations is kept in a Trunk.State struct which is passed to each callback. Each version also keeps track of its own state in a Trunk.VersionState struct which is also available through the Trunk.State.versions
map.
Transformation
One of the key features of Trunk is the ability to take a file and produce transformed versions. Perhaps you want to take an uploaded photo and produce a thumbnail, take a video and extract thumbnails, or take an XLS file and produce a CSV file for easier processing. This is all handled easily with the flexible version and transform system.
See full documentation on Trunk.transform/2
Credits
A shout out to stavro who created arc which I used in many of my projects, and which provided much inspiration for what resulted in Trunk.
Photos used in testing
Copyright and License
Copyright (c) 2017 Andrew Timberlake
This work is free. You can redistribute it and/or modify it under the terms of the MIT License. See the LICENSE.md file for more details.