Adding support for CodeBuild to pull from docker hub

docs/providers-guide.md

@@ -198,6 +198,14 @@ Provider type: `codebuild`.
   > pipeline to consume a custom image if required.
   > Along with `repository_arn`, we also support a `tag` key which can be used
   > to define which image should be used (defaults to `latest`).
+  > 
+  > Image can also take an object that contains a reference to a
+  > public docker hub image such as `bitnami/mongodb`. This allows your
+  > pipeline to consume a public docker hub image if required.
+  > Along with the docker hub image name, we also support using a tag which can
+  > be provided after the docker hub image name to define which image should
+  > be used (defaults to `latest`).
+  > 
 - *size* *(String)* **(small|medium|large)** - default: `small`.
   > The Compute type to use for the build, types can be found
   > [here](https://docs.aws.amazon.com/codebuild/latest/userguide/build-env-ref-compute-types.html).

docs/user-guide.md

@@ -319,6 +319,22 @@ pipelines:
       - ...
 ```
 
+Public images from docker hub can be defined in your deployment map like so:
+
+```yaml
+pipelines:
+  - name: example-custom-image
+    default_providers:
+      source:
+        ...
+      build:
+        provider: codebuild
+        properties:
+          image: bitnami/mongodb
+    targets:
+      - ...
+```
+
 ### CloudFormation Parameters and Tagging
 
 When you define CloudFormation templates as artifacts to push through a pipeline you might want to have a set of parameters associated with the templates. You can utilize the `params` folder in your repository to add in parameters as you see fit. To avoid having to create a parameter file for each of the stacks you wish to deploy to, you can create a parameter file called `global.yml` *(or .json)* any parameters defined in this file will be merged into the parameters for any specific account parameter file at build time. For example you might have a single parameter for a template called `CostCenter` the value of this will be the same across every deployment of your application however you might have another parameter called `InstanceType` that you want to be different per account. Using this example we can create a `global.yml` file that contains the following content:

src/lambda_codebase/initial_commit/bootstrap_repository/adf-build/shared/cdk/cdk_constructs/adf_codebuild.py

@@ -158,6 +158,15 @@ def determine_build_spec(codebuild_id, default_props, target=None):
                 else DEFAULT_BUILD_SPEC_FILENAME
             ),
         )
+
+    @staticmethod
+    def get_image_by_name(specific_image: str):
+        if hasattr(_codebuild.LinuxBuildImage,
+            (specific_image or DEFAULT_CODEBUILD_IMAGE).upper()):    
+            return getattr(_codebuild.LinuxBuildImage,
+                (specific_image or DEFAULT_CODEBUILD_IMAGE).upper())
+        else:
+            return _codebuild.LinuxBuildImage.from_docker_registry(specific_image)
 
     @staticmethod
     def determine_build_image(scope, target, map_params):
@@ -183,10 +192,7 @@ def determine_build_image(scope, target, map_params):
                 repo_arn,
                 specific_image.get('tag', 'latest'),
             )
-        return getattr(
-            _codebuild.LinuxBuildImage,
-            (specific_image or DEFAULT_CODEBUILD_IMAGE).upper(),
-        )
+        return CodeBuild.get_image_by_name(specific_image) 
 
     @staticmethod
     def generate_build_env_variables(codebuild, shared_modules_bucket, map_params, target=None):