branberry / PVConsumer

Consumer PV data from various sources

PVConsumer

Consumer PV data from pvoutput.org. THe idea is to expand to different data sources in the future.

Live

This application pulls live data from PV output.org and stores it in a database

To run the application locally

python pvconsumer/app.py

Logic

The app has the following high-level strucuture

  graph TD;
      A[1. Get PV system]-->B;
      B[2. Filter PV Systems]-->C;
      C[3. Pull Data]-->D[4. Save data];

1. Get PV System

   graph TD
    A0(Start) --> A1
    A0(Start) --> A2
    A1[Load local PV systems] --> A3{Are all PV system in the database}
    A2[Load Database PV systems] --> A3
    A3 --> |No| A4[Load the extra <br/> PV systems from pvoutput.org]
    A3 --> |yes| A5(Finish)
    A4 --> A5

2. Filter PV Systems

   graph TD
    B0(Start) --> B1{Is there any PV data in <br/> our database for this PV system?}
    B1 --> |No| B2[Keep PV system]
    B1--> |yes| B3{Is there any more PV data, <br/> from pv output.org, <br/>available for this PV system?}
    B3 --> |yes| B2
    B3 --> |No| B5[Dischagre PV system]
    B2 --> B6(Finish)
    B5 --> B6

3. Pull Data

   graph TD
    C0(Start) --> C1[Pull Data from pvoutput.prg]
    C1 --> C2{Is this data <br/> in our database already?}
    C2 --> |yes| C3[Keep PV data]
    C2 --> |No| C4[Dischagre PV data]
    C3 --> C5(Finish)
    C4 --> C5

🩺 Testing

Tests are run by using the following command

docker-compose -f test-docker-compose.yml run pvconsumer

These sets up postgres in a docker container and runs the tests in another docker container. This slightly more complicated testing framework is needed (compared to running pytest) as some queries can not be fully tested on a sqlite database

Environmental Variables

• DB_URL: The database url which the forecasts will be saved too
• API_KEY: API key for pvoutput.org
• SYSTEM_ID: System id for pvoutput.org
• DATA_SERVICE_URL: data service url for pvoutput.org