oc-cf-template.yml

AWSTemplateFormatVersion: "2010-09-09"
Description: Deploys opencravat software and full complement of annotators on separate volume, takes in 2 files and processes based on parameters in YML file
Parameters:
  InstanceType:
    Type: String
    Description: 'The default, m5.2xlarge, should be fine unless you have millions of RSIDs'
    AllowedValues:
      - m5.large
      - m5.xlarge
      - m5.2xlarge
      - m5.4xlarge
      - m5.8xlarge
    ConstraintDescription: 'Valid instance type in m5 family which will appropriately handle workload'
    Default: m5.2xlarge
  ImageId:
    Type: 'AWS::EC2::Image::Id'
    Description: >-
      OpenCRAVAT AMI with all modules installed.
    Default: ami-0894930d2a802712f

  VpcId:
    Type: 'AWS::EC2::VPC::Id'
    Description: Any should work, we just can't pick this for you!
  OCS3Bucket:
    Type: String
    Description: 'Name of your s3 bucket, ex. "haplocravat2". If your files are nested inside the bucket ex. "haplocravat2/files/", append the full path after the bucket name. No need to add any protocol such as s3://'
  OCInputFile:
    Type: String
    Description: "List of rsIDs or VCF file"
  OCYMLFile:
    Type: String
    Description: >-
      Your YML File to control the OC Run. Be sure this has been uploaded to the S3 bucket
  KeyPair:
    Type: 'AWS::EC2::KeyPair::KeyName'
    Description: Amazon EC2 Key Pair
    Default: 0
  SubnetId:
    Type: 'AWS::EC2::Subnet::Id'
    Description: >-
      Subnet ID your instance will launch in. Should be Internet accessible for
      the purposes of this demo.
Metadata:
  AWS::CloudFormation::Interface:
    ParameterGroups:
      -
        Label:
          default: "Bucket Configuration"
        Parameters:
          - OCS3Bucket
      -
        Label:
          default: "Input File Configuration"
        Parameters:
          - OCInputFile
          - OCYMLFile
      -
        Label:
          default: "Run Size Configuration"
        Parameters:
          - InstanceType
      -
        Label:
          default: "Amazon EC2 Configuration Specifics"
        Parameters:
          - ImageId
          - KeyPair
          - VpcId
          - SubnetId
    ParameterLabels:
      VpcId:
        default: "Which VPC should this be deployed to?"
      OCS3Bucket:
        default: "Your S3 Bucket"
      OCYMLFile:
        default: "Your configuration file"
      OCInputFile:
        default: "Your input file"
      InstanceType:
        default: "What size instance do you need?"
      ImageId:
        default: "Select the correct AMI"
      KeyPair:
        default: "Which of your Amazon keys do you want to use?"
      SubnetId:
        default: "Select one of your subnets for the computation"

Resources:
  OCS3ReadProfile:
    DependsOn: OCS3ReadRole
    Type: 'AWS::IAM::InstanceProfile'
    Properties:
      Path: "/"
      Roles:
        #- OCS3ReadRole
        - !Ref OCS3ReadRole
  OCS3ReadPolicy:
    Type: 'AWS::IAM::Policy'
    Properties:
      PolicyName: !Join
        - "-"
        - - "name"
          - !Select
            - 0
            - !Split
              - "-"
              - !Select
                - 2
                - !Split
                  - "/"
                  - !Ref "AWS::StackId"
      PolicyDocument:
        Version: "2012-10-17"
        Statement:
          - Effect: Allow
            Action:
              - 's3:ListBucket'
            Resource:
              - !Sub 'arn:aws:s3:::${OCS3Bucket}'
          - Effect: Allow
            Action:
              - 's3:PutObject'
              - 's3:GetObject'
              - 's3:DeleteObject'
            Resource:
              - !Sub 'arn:aws:s3:::${OCS3Bucket}/*'
      Roles:
        - !Ref OCS3ReadRole
  OCS3ReadRole:
    Type: 'AWS::IAM::Role'
    Properties:
      RoleName: !Join
        - "-"
        - - "name"
          - !Select
            - 0
            - !Split
              - "-"
              - !Select
                - 2
                - !Split
                  - "/"
                  - !Ref "AWS::StackId"
      AssumeRolePolicyDocument:
        Version: "2012-10-17"
        Statement:
          - Effect: Allow
            Principal:
              Service:
                - ec2.amazonaws.com
            Action:
              - 'sts:AssumeRole'
  OCEC2SecurityGroup:
    Type: 'AWS::EC2::SecurityGroup'
    Properties:
      SecurityGroupIngress:
        - CidrIp: 0.0.0.0/0
          FromPort: 22
          ToPort: 22
          IpProtocol: tcp
        - CidrIp: 0.0.0.0/0
          FromPort: 8060
          ToPort: 8060
          IpProtocol: tcp
      GroupDescription: SSH access for TS and cravat-view ports
      VpcId: !Ref VpcId
  OCNetworkInterface:
    Type: 'AWS::EC2::NetworkInterface'
    Properties:
      SubnetId: !Ref SubnetId
      Description: Interface for OC Machine
      GroupSet:
        - !Ref OCEC2SecurityGroup
      SourceDestCheck: true
  OCEC2Instance:
    Type: 'AWS::EC2::Instance'
    DependsOn: OCS3ReadProfile
    Properties:
      ImageId: !Ref ImageId
      InstanceType: !Ref InstanceType
      KeyName: !Ref KeyPair
      NetworkInterfaces:
        - NetworkInterfaceId: !Ref OCNetworkInterface
          DeviceIndex: 0
      Tags:
        - Key: Name
          Value: OpenCravat-Machine
      IamInstanceProfile: !Ref OCS3ReadProfile
      UserData:
        'Fn::Base64': !Sub |
            #!/bin/bash
            # Download the source files to the local machine
            cd /tmp
            aws s3 cp s3://${OCS3Bucket}/${OCInputFile} .
            aws s3 cp s3://${OCS3Bucket}/${OCYMLFile} .
            #Establish some time stamping
            stamp=`date +%s`
            # Run the cravat command based on parameters in YAML file
            cravat /tmp/${OCInputFile} -c /tmp/${OCYMLFile} 2>&1 | tee OCrun-$stamp-commandout.txt
            # Upload the sqlite output and log file to s3 using s3cmd
            aws s3api put-object --bucket ${OCS3Bucket} --key ${OCInputFile}-$stamp/
            aws s3 cp ${OCInputFile}.sqlite s3://${OCS3Bucket}/${OCInputFile}-$stamp/
            aws s3 cp ${OCInputFile}.log s3://${OCS3Bucket}/${OCInputFile}-$stamp/
            aws s3 cp ${OCInputFile}.err s3://${OCS3Bucket}/${OCInputFile}-$stamp/
            aws s3 cp OCrun-$stamp-commandout.txt s3://${OCS3Bucket}/${OCInputFile}-$stamp/
            # Wait 1 minute just to make sure the uploads complete
            sleep 60
            # Shut the instance down
            /sbin/shutdown -h now
Outputs:
  PublicIp:
    Value: !GetAtt OCEC2Instance.PublicIp