build: unified deployment script (#18)

docs: links, script details

README.md

@@ -9,9 +9,9 @@ AWS Lambdas are fun, but often the amount of boilerplate involved in getting a p
 
 ## Local development
 
-The base Lambda handler is at `src/base.py` and all the Terraform configurations are in `infrastructure`.
+The base Lambda handler is at `src/base.py` and all the Terraform configuration files are in `infrastructure`.
 
-[Read more about Sceptre](https://sceptre.cloudreach.com/latest/index.html://www.terraform.io/docs/index.html)
+[Read more about Terraform](https://www.terraform.io/docs/index.html)
 
 [Read more about AWS Lambda](https://docs.aws.amazon.com/lambda/latest/dg/lambda-python.html)
 
@@ -22,7 +22,16 @@ All](https://github.com/github/scripts-to-rule-them-all) paradigm and can be fou
 
 ## Deployment
 
-Deployment is in three steps: on first setup, you will need to make sure that your `bootstrap` environment is ready via `PROJECT=bootstrap . script/apply`. Then, you should prepare your source code package (dependent on the language and framework you're using, you have to supply this bit!) and `ARCHIVE=<path-to-zip> . script/push`. Finally, you can deploy your application resources with `PROJECT=app . script/apply`.
+Provided your runtime source code is prepackages and ready to go, you can simply `ARCHIVE=<path-to-function-zip> . script/deploy` to deploy your function in seconds!
+
+By default, the resources are identified with an environment name `dev-$USER`, but you can change this in staging /
+production deployments by specifying the `ENV_NAME` environment variable.
+
+Individual parts of the application (i.e. the bootstrap resources or the application itself) can be deployed separately
+using `PROJECT=<bootstrap|app> script/apply`.
+
+Prepackaging the source code depends on what language is used as a runtime, more details can be found in [AWS Lambda
+documentation](https://docs.aws.amazon.com/lambda/latest/dg/gettingstarted-package.html).
 
 ## Contributing
 

script/deploy

@@ -0,0 +1,52 @@
+#!/usr/bin/bash -e
+
+# Deploys the whole project (bootstrap + application resources). This
+# also takes care of pushing the lambda function's packed source code
+# to S3, provided ARCHIVE is provided.
+#
+# Example usage:
+#   ARCHIVE=./lambda_function.zip . script/deploy
+
+source $(dirname $0)/../.config
+
+BOOTSTRAP_INFRA_ROOT=$(realpath $BOOTSTRAP_ROOT/infrastructure/bootstrap)
+APPLICATION_INFRA_ROOT=$(realpath $BOOTSTRAP_ROOT/infrastructure/app)
+DEFAULT_ENVNAME="dev-$USER"
+
+# Bootstrap resources deployment
+
+(
+    cd $BOOTSTRAP_INFRA_ROOT
+    terraform init
+    terraform apply --var-file $VARIABLES_PATH
+
+    BUCKET_NAME=$(terraform output --json | jq .artifacts_bucket_name.value -r)
+    cd -
+    aws s3 cp $ARCHIVE s3://$BUCKET_NAME
+) || exit $?
+
+# Application resources deployment
+
+(
+    cd $APPLICATION_INFRA_ROOT
+    terraform init
+
+    # Some resources are always marked as tainted
+    # to force their recreation.
+
+    declare -a ALWAYS_TAINT_RESOURCES=(
+        "aws_lambda_function.apgnd_lambda_func"
+        "aws_lambda_permission.apigw"
+    )
+ 
+    for RESOURCE in $ALWAYS_TAINT_RESOURCES 
+    do
+        terraform taint --allow-missing $RESOURCE
+    done
+
+    terraform apply \
+        --var-file $VARIABLES_PATH \
+        -var="env_name=${ENV_NAME:-$DEFAULT_ENVNAME}" \
+        -var="commit_sha=$(git log --pretty=format:'%H' -n 1)" \
+        -var="lambda_archive_name=$(basename $ARCHIVE)"
+) || exit $?