climate-cooperative / home-energy-assessment-database Goto Github PK

Description

TypeScript introduces a layer of type safety to the repo. It also comes with additional features which can help with payloads/models, to improve code quality.

While the repo is still relatively small, this won't be a huge refactor.

Technical Notes

• import typescript library as devDep
• refactor to typescript
• convert models to interface's
• ensure building still works

Risks

[Add Risks]

Related Issues Or Discussions

[...Issues]

Right now the lambda builds a sample js image. Need to:

• Build image of express app. (src/lambda.js)
• update Dockerfile.lambda
• update .env to reflect any new env vars
• create scripts to locally run express app as lambda

Lambdas performance is is dictated by its memory allocation. This is a setting defined in the tf module with memory_size attribute. Default is 128MB.

By Right Sizing this memory, can find a balance of performance and price. faster lambdas mean less cold/hot/warm response times. Lambda pricing is partially number of reqs, but also by execution time. Increasing mem decreases execution time.

Technical Notes:

https://s3.amazonaws.com/lambda-tools/pricing-calculator.html

Capture metrics regarding to plug into calculator and change mem to see what fits best (cheapest)

The pipeline now uses a lot of hard coded variables. Moving these to parameters/variables allows for more re-usable code and configurable builds.

Add Variables for:

• ACM cert name
• domain name
• application name
• environment (PROD, DEVTEST..)

To get the best performance from lambda, its best to keep the amount of code put into the lambda as small as possible. IE bundling EVERYTHING to a single entrypoint

Technical notes:

• Explore esbuild (primary option)
• Explore webpack (meh it'll work)

Description

An OpenAPi spec is a useful tool to describe an api. Describes available paths, and payloads.

https://swagger.io/specification/

Technical Notes

Create endpoint as well, to programmatically provide it. /openapi???

Maybe use https://www.npmjs.com/package/swagger-autogen for auto generating openapi/swagger specs

Risks

[Add Risks]

Related Issues Or Discussions

[...Issues]

make it open source hospitable

Best practices (to include in checks during build_deploy) is to test the api works as intended with a set of integration tests.

Using Jest Create:

• A Test suite that can be run in the pipeline
• A test suite that can be run locally (maybe the same as above)

Description

Right now the API is tooled to integrate with dynamodb. It will be useful to abstract this layer out, so that any db/orm client can be used for implementation purposes.

Technical Notes

• Create custom abstract service layer
• Possibly utilize https://www.prisma.io/orm

Risks

Related Issues Or Discussions

[...Issues]

Made the decision to switch from mongo to dynamo. Need to:

• Terraform infra for dynamo
• iam permission for lambda to access dynamo
• refactor api calls to mongo to use dynamo client https://docs.aws.amazon.com/AWSJavaScriptSDK/v3/latest/client/dynamodb/
• update hydrating scripts to PUT data into db

A new route to get State data via the zipcode

/state/:zipcode

Technical notes:

Either use query parameters or POST method to allow for this access

Switching to DynamoDB we need 1of the following:

• A set of scripts, or client to load data into dynamo. Data comes from LOCAL
• A client to migrate data from mongo to Dynamo. Data comes from MONGO
• A new repo to act as client to add data.
  • This would allow for easy access and updating of data

Open Questions:

• Where does the data come from? The origin of truth?
• How often does the data need to be updated
• automate it???? i think yes

Description

Create an endpoint (/score) to provide a calculated response that can plug in to the FE: https://github.com/climate-cooperative/home-energy-calculator.

Document req/res models as well

Technical Notes

Do similar calculations in the backend.

POST request.

Risks

Without Integration Tests, will take coordination with FE to ensure correct payload.

Related Issues Or Discussions

Move Calculations to API

Quite a bit of dead code in the repo from older times. Get ride of it...

Swapping to Dynamo opens up a lot of opportunities in the cloud, but closes some doors locally. A nice to have is a local dynamo instance to hit locally.

Needs to:

• standup local (mock) dynamo instance using docker-compose
• be able to simulate full E2E usage

Technical Notes:

• Needs to specify which dynamo instance to hit using .env files. ie .env.LOCAL & .env.PROD
• .env.PROD can probably be made during pipeline?
• Try using https://docs.localstack.cloud/user-guide/aws/dynamodb/
• simple set of mock data checked-in to help on db init