An Amazon Web Services SDK for Haskell with support for most public services. Parts of the code contained in this repository are auto-generated and automatically kept up to date with Amazon's latest service APIs.
- You can find the latest Haddock documentation for each respective library on the Amazonka website.
- A release changelog can be found in lib/amazonka/CHANGELOG.md.
- For problems, comments, or feedback please create an issue here on GitHub.
- License
- Directory Layout
- Supported Platforms and GHC Versions
- Getting Started
- Building the Project
- Building the Documentation
- Running the Code Generator
- Code Formatting
- Third Party Packages
Amazonka is licensed under the Mozilla Public License Version 2.0.
The AWS service descriptions are licensed under Apache 2.0. Source files derived from the service descriptions contain an additional licensing clause in their header.
This repository is organised into the following directory structure:
lib/amazonka
: The main library containing setup, authentication, and send logic. This will be your primary dependency.lib/service/amazonka-*
: A library per supported Amazon Web Service, you'll need to add a dependency on each selected service library.lib/amazonka-core
: Theamazonka-core
library upon which each of the services depends.lib/amazonka-test
: Common test functionality.examples
: Basic examples for using the service libraries.configs
: Service configuration, templates, and assets used by the code generator.docs
: The website documentation and related build code.gen
: The code and configuration generators.nix
: Nix configuration code for toolchain packages.scripts
: Scripts to manage the project, such as the release lifecycle.tools
: Custom bazel rules.third_party
: Third party packages and patches.
GHC versions 8.8.4
and 8.10.7
are officially supported and tested on NixOS, Ubuntu, and macOS. GHC 8.6.5
may also work, but is not tested by our continuous integration pipeline.
This repository is built using a combination of Nix and your choice of Bazel or Cabal. If you're just using Amazonka as a git dependency in your Cabal or Stack project, you can skip these steps. But if you plan on contributing to the codebase - welcome, read on!
git clone [email protected]:brendanhay/amazonka.git
cd amazonka
Building the code in this repository requires various development dependencies (e.g. Nix, Bazel, GHC.)
The Nix package manager is used to obtain and build the other dependencies in a hermetic environment. You can install Nix by following the official installation instructions:
sh <(curl -L https://nixos.org/nix/install) --daemon
Once Nix is setup, you can enable the cache to avoid building dependencies:
nix-env -iA cachix -f https://cachix.org/api/v1/install
cachix use amazonka
The build tools are installed and activated upon entering a Nix shell, which is achieved by running the following command in the root of the repository:
nix-shell
You can also enter a shell and explicitly specify the GHC version:
nix-shell --argstr ghcVersion 884
Optionally, if you have Direnv and lorri installed you can use the provided .envrc instead, which will also add the scripts directory to your PATH
. You can extend this by adding your own uncommitted .envrc.local
file. See the Direnv Wiki for various recipes.
The following commands assume you're already in a nix-shell outlined in the previous step.
If you're familiar with Cabal, you can build amazonka-*
packages via:
cabal build amazonka amazonka-s3
Or the entire project (which will take a very long time!):
cabal build all
Alternatively, if you plan on contributing to the project or want to perform code generation, you will need to familiarise yourself with Bazel. You can build packages by specifying one or more targets using Bazel's label syntax:
bazel build //lib/amazonka //lib/services/amazonka-s3
Or build all Haskell libraries in the project using the ...
wildcard:
bazel build //lib/...
To view what targets are available in the workspace:
bazel query //...
By default, the
bazel
command will use the same GHC version as the Nix shell'sghcVersion
argument. You can choose a different GHC version usingnix-shell --argstr ghcVersion 884
- which is just a synonym forbazel build --//tools/ghc:version=884
.
The docs Bazel package contains the Haddock target and Hugo static site definition and markdown content. To build the site locally, run:
bazel build //docs:bundle
Alternatively, you can serve the documentation site locally on http://localhost:1313
by running:
bazel run //docs:serve
The gen Bazel package contains code generators for synthesising Haskell data types, packages, and configuration from the botocore service definitions.
scripts/generate will run the code generator for all services configured in config/services, for example:
./scripts/generate
Or, you can selectively run the generator on one or more services:
./scripts/generate ec2 s3 iam
To update the botocore service definitions used by the generator, you can run:
./scripts/update-botocore
scripts/generate-configs will run the config generator to produce placeholder config/serivces configurations for the version of botocore pinned in the WORKSPACE.
To generate any missing service configurations:
./scripts/generate-configs
Service configurations generated in this way are intended as examples only and the resulting configs/services/<name>.json:libraryName
(Haskell package name) and configs/annexes/<name>.json:serviceAbbreviation
(Haskell package namespace) should be manually verified and curated as necessary.
For pull requests which affect generated output please do not include the regenerated amazonka-*
packages, only commit updates to the build rules, documentation, generator, and related configuration. This ensures the Continuous Integration process is the single source of truth for the generated code and reduces noise in pull requests, keeping them reviewable and focused on actual generator code/logic changes.
Please use ./scripts/format
frequently - it's OK, I hate 2 spaces too, we're in this together.
When naming an additional library which provides supplemental functionality to amazonka
, if you want to use the amazonka-*
namespace, then please consider prefixing your package names with amazonka-contrib-*
. For example, amazonka-contrib-rds-utils.
This minimises potential future collisions with auto-generated package names and new AWS service and product releases.