Tanzania HTI

In order to enable the THTI project, we're assembled a solution that allows complex workflows to be created an changed by ODK form submission. The current use case captured is this:

HIV Flowchart

Nodes 1, 3, and 5 each represent a submitted ODK form. (Update: two additional nodes or ODK forms were added before node 1). XLS forms can be found here: HIV XLS Forms

TB Flowchart

Nodes 1, 2, 3 and 4 each represent a submitted ODK form. XLS forms can be found here: TB XLS Forms

To accomplish this, we'll use:

• RapidPro (via the hosted textit.in service)
• Gather
• Aether Stream Consumer
• Aether Elasticsearch Consumer

The main drivers of the workflow are the Stream Consumer and RapidPro, related as described in the following diagram:

Artifacts for all of the consumers and RapidPro can be found in the appropriate folders.

Pipelines:

All the Pipelines are available in StreamConsumer/README.md. For each pipeline, you will need to set:

• The textit api token for rapidpro_auth_header available on Textit
• The right rapidpro IDs. A rapidpro flow ID can be found in the url of the flow.
• (Optional) The username and password for the Aether/Gather.

Credentials

All username and passwords are available in the HTI Credentials note in the Shared Folder on LastPass.

Useful commands

• Get the list of form submissions (ordered by submission date): https://eha-data.org/htidev/kernel/entities.json?ordering=-modified
• Get the list of pipelines: https://eha-data.org/htidev/stream-consumer/pipeline/list
• Get a specific pipeline: https://eha-data.org/htidev/stream-consumer/pipeline/get?id=pipeline_id
• Get the list of subscriptions: https://eha-data.org/htidev/es-consumer/subscription/list
• Get a specific subscription: https://eha-data.org/htidev/es-consumer/subscription/get?id=subscription_id
• Get a specific job: https://eha-data.org/htidev/stream-consumer/job/get?id=job_id
• Get the list of jobs: https://eha-data.org/htidev/stream-consumer/job/list
• Get the list of topics (ordered by last updated): https://eha-data.org/htidev/kernel/schemadecorators.json?fields=modified%2Cname%2Cid%2Ctopic&ordering=-modified
• Add/update a pipeline: http --json POST https://eha-data.org/htidev/stream-consumer/pipeline/add -a username:password < ./StreamConsumer/pipeline_file.json
• Add job to consumer: http --json POST https://eha-data.org/htidev/stream-consumer/job/add -a username:password < ./StreamConsumer/job_file.json Example of job:

{
  "id": "rural_a_1",
  "name": "HIV Rural A",
  "pipelines": [
    "rural_a"
  ]
}

• To see if a flow was started in RapidPro (textit), go to https://textit.com/flowstart/
• Kakfa topics can be found in the ehealth-africa-prod1 cluster in Confluent

Test

• You can test the entire flow based on this test file.