https://github.com/dequis/lutetium

move along nothing to see here
https://github.com/dequis/lutetium

Last synced: 3 months ago
JSON representation

move along nothing to see here

• Host: GitHub
• URL: https://github.com/dequis/lutetium
• Owner: dequis
• Created: 2018-08-19T20:51:13.000Z (almost 7 years ago)
• Default Branch: master
• Last Pushed: 2018-08-21T07:33:51.000Z (almost 7 years ago)
• Last Synced: 2025-03-07T03:30:55.759Z (3 months ago)
• Language: Python
• Size: 25.4 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# lutetium

Move along, nothing to see here.

## Installation and usage

    docker-compose up

The current directory will be mounted in the container, with the output being

written to `out.jsonl` (line-separated json)

### Alternate deployment modes

The default commandline is the following:

    python -m lutetium -vv all

This runs the `all` task set with debug-level verbosity. This is equivalent to:

    python -m lutetium -vv meter pvsim writer

All three tasks can be run individually in separate processes, if desired.

Other options:

    Options:

      -v, --verbose

      -f, --filename TEXT

      --meter-source [random|absurd_polynomial|noisy_absurd_polynomial]

      --pvsim-source [random|absurd_polynomial|noisy_absurd_polynomial]

      --help                          Show this message and exit.

`-f` changes the output filename (default `out.jsonl`)

`--meter-source` and `--pvsim-source` change the data sources for each of those

components. Available sources:

 * `random`: A meter source that targets random values and moves 10%-30% of the

    way in each step

    Default meter source.

 * `absurd_polynomial`: A PV data source that pretends this is a data science

    problem. 

    Uses data extracted from the PDF using [WebPlotDigitizer][1] (data set

    available at `data/pv_curve_pdf.json`), then fits that data into a cubic

    polynomial using numpy, excluding source data points under a threshold of

    y=0.1 so the curve doesn't have to deal with the flat night portion. Then,

    when evaluating the polynomial, it uses max(0, x) to skip the negative part

    of the resulting curve.

    To avoid the numpy dependency at runtime, the generated polynomial is just

    a constant and evaluated with a list comprehension. The code to generate it

    is available below, but not executed.

 * `noisy_absurd_polynomial`: Adds a bit of random noise over

   `AbsurdPolynomialFitSource`

    Default pvsim source.

### RabbitMQ configuration

The RabbitMQ connection can be configured through the `LUTETIUM_AMQP_CONFIG`

environment variable. It should be a JSON object, which for the defaults would

be `{"host": "127.0.0.1", "port": 5672}`

Available keys, from the aioamqp config:

> * host (str) – the host to connect to

> * port (int) – broker port

> * login (str) – login

> * password (str) – password

> * virtualhost (str) – AMQP virtualhost to use for this connection

> * ssl (bool) – create an SSL connection instead of a plain unencrypted one

> * verify_ssl (bool) – verify server’s SSL certificate (True by default)

> * login_method (str) – AMQP auth method

## Development usage with virtualenv

    docker-compose up -d rabbitmq

    ./setup_virtualenv

    . ./activate

    python -m lutetium

Running tests:

    python -m pytest

## What's the name?

Random word generated by [asd][2], which is what I use to avoid getting stuck

thinking names for things.

[1]: https://apps.automeris.io/wpd/

[2]: https://github.com/dequis/dotfiles/blob/master/bin/asd