Is your feature request related to a problem? Please describe.
I would like to have CodeSeeder support for partitions other than aws...in particular for aws-cn and aws-us-gov

Describe the solution you'd like
I want call a codeseeder API to see if seedkit is deployed for my project (not via decorator patterns).
I want to be able to check:

• if seedkit deployed, return the cfn output in a dict
• if not, call an exposed method to deploy the seedkit and return the CFN output

** USE CASE:

• my code needs artifacts created by seedkit, but the methods decorated for codeseeder are not called until after the needed artifacts are required to be set

Due the the direct string replacements used to transform the absolute local path of the source files to a bundle relative path we bundle the entire local path when running in a windows environment. Suggest fixing as follows within _bundle.py:

zip_relative_path = os.path.relpath(pathlib.Path(path), pathlib.Path(root_dir)) zf.write(path, zip_relative_path)

Currently only support for PLAINTEXT environment variables in CodeBuild is supported. PARAMETER_STORE and SECRETS_MANAGER environment variable types are missing

Lets discuss feasibility of below...?

Is your feature request related to a problem? Please describe.
Similar to the creation of the seedkit toolkit, our code calls the creation of a project managed policy that is digging deep into the codeseeder codebase.

Describe the solution you'd like
Create a command level function that will check if the managed policy stack is deployed, return a boolean, the stackname and output.
Create a command level function that will create the stack
Create a command level function to destroy the stack

A similar solution to this issue, but for the managed policy:
#20
#21

Is your feature request related to a problem? Please describe.
At times it may be necessary to asynchronously execute the remote_function code. Currently the processes blocks until the CodeBuild execution completes or errors.

Describe the solution you'd like
A mechanism for flagging a remote_function for asynchronous execution. When the function executes, rather than blocking until CodeBuild completes or errors, the library could block until CodeBuild starts, then return the ExecutionId. A secondary function to rejoin and/or check execution status and retrieve any return values will also be required.

Additional context
The use case being targeted is to begin execution of a remote_function within AWS Lambda function and then check execution status with another Lambda function. This would enable CodeSeeder workflows to be used in AWS StepFunctions StateMachines. For this to work, the asynchronous remote_function will need to return the CodeBuild ExecutionID so that it can be returned from the Lambda. subsequent Lambda functions would take the ExecutionID as input and use it to check Execution status.

Is your feature request related to a problem? Please describe.
When a failure occurs, a Runtime error is thrown by CodeSeeder, and a customized string is added to the error message including the build id, but no other information.

Describe the solution you'd like
Create a new Error class that will contain a customized error string and a dict of any/all pertinent data (account id, region, build_id, etc)

Describe alternatives you've considered
Crafting a custom dict that serialized to a string as the error message

Describe the bug
When trying to add a file to the bundle via the decorators (extra_files) codeseeder throws an error if the file is in a directory.

For example, when adding seedfarmer.yaml to the bundle, that file is located at the root of the bundle, so it is added easily. But when trying to add an extra file located at data/somewhere/extradata.json, codeseeder issues an error:

FileNotFoundError: [Errno 2] No such file or directory: <userPath>/idf-modules/codeseeder.out/examples-extra-dummy-dummy/bundle/data/somewhere/extradata.json because the parent directory data/somewhere does not exists in the local bundle path under codeseeder.out.

Please see line 132 in aws_codeseeder/_bundle.py where the shutil.copy takes place

At this point, when we deploy a seedkit using codeseeder, it deploys AWS Codeartifact domain and AWS Codeartifact Repository by default. This default behavior of creating AWS Codeartifact will cause an issue in the regions where it is not supported. AWS Codeartifact should be created optionally

To launch ADDF modules in isolated env, as an example: EKS module, we need ability to run codebuilds in a VPC mode, for being able to talk to private API server of EKS cluster.

What needs to run in codebuild to talk to EKS cluster?

• eksctl commands from deployspec.yaml
• helm commands
• kubectl commands

We have added support for deploying on Windows OS (ref PR #81 ), but when trying to execute codeseeder deploy modules --module aws-codeseeder:. to upload and artifact, the upload fails as it is based on a bash script.

REF: aws-codeseeder/aws_codeseeder/commands/_module_commands.py
REF: aws-codeseeder/aws_codeseeder/resources/update_repo.sh

Add support for windows to upload a library?

NOTE: this does not affect deployments....only the management of CodeArtifact-enabled projects who want to upload libraries.

First invocation of codeseeder from a client fails if the seedkit has not been deployed. It would be great if the seedkit would be created automatically if codeseeder detects that one does not exist on first invocation.

Describe the bug
When deploying without a SecretsManager configured, there is a stack trace stating that an error occurred. This is true, an error occurred, but the stack trace is confusing as the build process continues.

To Reproduce
Deploy a module via seedfarmer with a module that does not use SecretsManager (is not logging into DockerHub). See the stack trace

Expected behavior
When not using DockerHub or SecretsMananger, this stack trace should not indicate an error?

** Sample Log**

An error occurred (ResourceNotFoundException) when calling the GetSecretValue operation: Secrets Manager can't find the specified secret.
Traceback (most recent call last):
  File "/var/scripts/retrieve_docker_creds.py", line 24, in get_secret
    get_secret_value_response = client.get_secret_value(SecretId=secret_name)
  File "/root/.pyenv/versions/3.7.10/lib/python3.7/site-packages/botocore/client.py", line 415, in _api_call
    return self._make_api_call(operation_name, kwargs)
  File "/root/.pyenv/versions/3.7.10/lib/python3.7/site-packages/botocore/client.py", line 745, in _make_api_call
    raise error_class(parsed_response, operation_name)
botocore.errorfactory.ResourceNotFoundException: An error occurred (ResourceNotFoundException) when calling the GetSecretValue operation: Secrets Manager can't find the specified secret.

Is your feature request related to a problem? Please describe.
In order to support cross-account or cross-region execution of the AWS CodeBuild it may be necessary to create a boto3 Session with credentials acquired from an sts:AssumeRole request or with a region other than the default. Currently the boto3 Session for all boto3 operations is created by CodeSeeder with default credentials and region.

Describe the solution you'd like
A method (function call? config?) for setting a boto3 Session that is used for boto3 calls. The method should enable updating of the boto3 Session so that short lived credentials can be refreshed during long running CodeBuild executions.

Describe alternatives you've considered
Overriding internal and/or private functions to manually set the boto3 Session

Describe the bug
When passing in a decorator to override the configuration of aws-codeseeder, the changes are not being evaluated.

For example, the codebuild_image is configured as a config parameter (default) and the image itself is changed with a remote decorator. That image override is not referenced.

REF: codeseeder.py line 217