Convert Your Savings Bond Wizard File to CSV

You can see this code running here: Convert Your Savings Bond Wizard File

The US Treasury Department used to have a Savings Bond Wizard that let you track and identify the purchased bonds. But the SBW file is no longer supported so this app will let you convert them to CSV files.

Thanks to Jim Evins and the GBonds project for the nucleolus of this code.

How Its Made

Vue.js + @carbon/vue

This part was the simplest to execute. The app itself is pretty simple with just a landing page and simple privacy statement. I use a custom theme for Carbon that includes a few tweaks to match some colors from the US Treasury site. Everything else is standard.

You can launch the site locally with:

cd ui
yarn
yarn serve

Deploying the UI

You can find the code for this in the ui directory.

Deployment of the UI is straightforward with just a bit of a reverse proxy incantation to to hide the services and cloud object storage layers. The config in ui/ce creates a reverse proxy for /services/ which points to the deployed services layer which is automatically calculated from the environment variable CE_SUBDOMAIN. It also creates a reverse proxy to the Cloud Object Storage under /downloads/ the URL for which is defined by the environment variable COS_DOWNLOAD

Build the image with either docker or podman and push it to the registry. Make sure each image has both the latest tag and another tag. I have a automatic retention policy set on my registry but that policy will not prune untagged images. Setting another tag on the image when you push it allows it to be pruned later.

• Install the code engine command line

• Login and target the project

  ibmcloud login --apikey @~/projects/ibmcloud-apikey.json -g Default -r us-south
  ibmcloud ce project select --name sbw2csv

podman build -t sbw2csv-ui . #Red Hat or Fedora
podman tag localhost/sbw2csv-ui:latest us.icr.io/sbw2csv/codeengine-sbw2csv-ui:latest
podman push us.icr.io/sbw2csv/codeengine-sbw2csv-ui:latest
podman push us.icr.io/sbw2csv/codeengine-sbw2csv-ui:latest us.icr.io/sbw2csv/codeengine-sbw2csv-ui:$(date '+%FT%H%M%S')

# or
docker build -t sbw2csv-ui .
docker tag localhost/sbw2csv-ui:latest us.icr.io/sbw2csv/codeengine-sbw2csv-ui:latest
docker push us.icr.io/sbw2csv/codeengine-sbw2csv-ui:latest
docker push us.icr.io/sbw2csv/codeengine-sbw2csv-ui:latest us.icr.io/sbw2csv/codeengine-sbw2csv-ui:$(date '+%FT%H%M%S')

Roll out new version

ibmcloud ce app update --name sbw2csv

Show the latest logs

ibmcloud ce app logs -f -n sbw2csv

Set UI Environment variables

# use your public COS endpoint here i.e. https://s3.us-south.cloud-object-storage.appdomain.cloud/your-bucket/
ibmcloud ce app update --name sbw2csv \
  --env COS_DOWNLOAD=YOUR-COS-INSTANCE-PUBLIC-ENDPOINT

Deploying the Services layer

You can find the code for this in the services directory. I bootstrapped this layer with express --view=pug services and then customized it for this app. The most important endpoint is services/routes/convert.js which receives uploaded files, converts them to CSV, and then stores them in the COS.

The only other interesting endpoint is in services/routes/analytics.js which accepts analytic posts from the UI. It takes that data and stores it in a connected cloudant db. This is a experiment at this point and not really relevant to this solution. If you want real analytics you should look elsewhere.

Build the image with either docker or podman and push it to the registry. Make sure each image has both the latest tag and another tag.

podman build -t sbw2csv-services . #Red Hat or Fedora
podman tag localhost/sbw2csv-services:latest us.icr.io/sbw2csv/codeengine-sbw2csv-services:latest
podman push us.icr.io/sbw2csv/codeengine-sbw2csv-services:latest
podman push us.icr.io/sbw2csv/codeengine-sbw2csv-services:latest us.icr.io/sbw2csv/codeengine-sbw2csv-services:$(date '+%FT%H%M%S')

# or
docker build -t sbw2csv-services .
docker tag localhost/sbw2csv-services:latest us.icr.io/sbw2csv/codeengine-sbw2csv-services:latest
docker push us.icr.io/sbw2csv/codeengine-sbw2csv-services:latest
docker push us.icr.io/sbw2csv/codeengine-sbw2csv-services:latest us.icr.io/sbw2csv/codeengine-sbw2csv-services:$(date '+%FT%H%M%S')

Roll out new version

ibmcloud ce app update --name services

Show the latest logs

ibmcloud ce app logs -f -n services

Services Environment

The services layer expects to be connected to a COS instance and to a Cloudant instance (for the analytics). Use your cloud account to configure these services to the Code Engine app.

# navigate to your cloud object store instance to create and API key and get your instance id
ibmcloud ce secret create --name cos-secrets \
  -l COS_INSTANCE_ID=crn:your-instance-id:: \
  -l COS_APIKEY=your-api-key

ibmcloud ce app bind --name services --service-instance YOUR-CLOUDANT-INSTANCE

ibmcloud ce app update --name services \
  --env-from-secret cos-secrets \
  --env COS_ENDPOINT=s3.private.YOUR-END-POINT \
  --env COS_BUCKET=your-bucket \
  --env NODE_DB_PREFIX=prod_ \
  --env DEBUG="services:server services:cos services:convert"

The services layer uses the parameter COS_ENDPOINT to find the COS endpoint so make sure that it defined correctly for your environment. You can see in the manifest file that it defaults to the private endpoint for us-east so the private endpoint for your region should work for you. If the app fails to connect via the private COS end point try the direct endpoint.

The services layer also uses the parameter COS_BUCKET to find the bucket in the COS. Define this for your environment and be sure to create the bucket in your environment. The code does not attempt to create the bucket. It expects it to already exist.

When you create the bucket be sure to set the expiration rule to automatically remove objects.

C++ Command line utility

The convert endpoint uses this utility to do its work. The code from the gbonds application already had an import for SBW files so those dependencies are in the services/cli/deps/gbonds-2.0.3/ directory. I added a little glue around that in services/cli/main.cpp to export the files to CSV format.

Then the only trick is to compile it so that it is compatible with the stack running the in Code Engine. The servces/Dokerfile handles that by using the latest CentOS image in the build phase.

Incidentally, you can build the command line utility locally like this:

Build

cd services/cli
cmake -S . -Bbuild
cmake --build build/

That works great for me on Ubuntu 20 and Fedora 35 but likely needs updates for OSX and Windows. LMK if you make changes and I'll merge them in.

Local db

If you want to try out the simple analytics locally, you should create a local db for this. You can do that like this:

cd database
yarn
yarn serve

This will automatically create and run local couchdb image that the services layer can connect to.

How to Make the How Its Made Illustration

• Edit the doc/HowItsMade.odp

• Record illustration with recordmydesktop

• Edit recording with OpenShot

  

• Process to gif How to make GIFs with FFMPEG

  ffmpeg -y -i Howitsmade.mpeg -filter_complex "[0:v] palettegen" palette.png
  ffmpeg -y -i Howitsmade.mpeg -i palette.png  -filter_complex "[0:v][1:v] paletteuse" Howitsmade.gif

Deployment notes

Make sure you are fully targeted including the resource group

ibmcloud ce project select -n YOUR_PROJECT
ibmcloud target

API endpoint: https://cloud.ibm.com
Region: us-south
User: [email protected]
Account: aaa
Resource group: ggg

Development setup

pre-reqs

• yarn

• nvm

• node 16 nvm install 16

• docker

  • podman on Fedora or RedHat works too

• cmake

• gcc

• boost

  Fedora

  sudo dnf install cmake boost-static boost-devel glib-devel libstdc++-devel libstdc++-static

• start database - open a new terminal

  cd database
  yarn
  yarn serve # this takes a few minutes

• build the sbw2csv cli

  cd services/cli
  cmake -S . -Bbuild/
  cmake --build build/

• start services - open a new terminal

  cd services
  yarn
  yarn serve

  NOTE: If you want your local environment to connect to a remote COS, add your CE_SERVICES value found in the runtime environment variables to a .env.local.json file and restart your services.

• start UI - open a new terminal

  cd ui
  yarn
  yarn serve