Search Assessment Tool

Overview

The Search Assessment Tool simply analyzes and assesses the relevancy performance of the existing search services (Canada.ca, known internal or global search, or department specific search, known as contextual or internal specific, and external search, Google for now). This tool fetches the top 100 search terms and the expected result of each from Airtable (3rd party SaaS) using its API, and checks how these search terms perform with respect to their target URLs. Finally, it provides a score for the search page in question.

Assessments only run once, and the results are archived in another Airtable base (a table/spreadsheet). This tool was meant to automatically get the top 100 search terms from Adobe Analytics and populate them in Airtable, but no access was allowed to Adobe Analytics as of this writing, and the workflow may change as a result. For this reason, this tool is still a work in progress.

This is a back-end service that exposes its services through API endpoints using Rest. The UI was built in a separate project using Angular search-assessment-ui.

Getting Started

This is a standalone Java application that has an embedded server. It doesn't require a container such as Tomcat, jetty, etc. It was built using a simple lightweight framework called Dropwizard. As a result, running this application is very simple.

1. Installation process

After the code is cloned from its repository, it can be ran as a Java application. To run the jar file, you need to pass the following arguments: server configuration.yml

2. Software dependencies

All the dependencies are in the pom file.

3. Latest releases

This is still in alpha and hasn't been deployed to production.

4. API references

Refer to the code as this changes very frequently.

---

As of this writing, the application is deployed in the Azure alpha environment:

dev: https://search-assessment.azurewebsites.net/

Build and Deploy

To build the jar file for deployment purposes, just run maven install. An uber jar will be built with all the dependencies. After that, a Docker container can be built for deployment purposes.

To build and deploy on Azure, use the following:

Update Docker Container

az acr build --registry CRAsearch --subscription "Pay-As-You-Go (cra-arc.alpha.canada.ca)" --image search-assessment . (latest - used for dev)

az acr build --registry CRAsearch --subscription Pay-As-You-Go (cra-arc.alpha.canada.ca) --image search-assessment:stage . (stage)

Using the Azure console, you can deploy the changes if "App Service" is not configured to do so automatically.