pynidus 0.0.134

A handful of utilities.

pynidus

A handful of utilities predominantly made to develop basic Cloud Run services that connect to the same databases. Any configuration variable should be passed as an environment variable from Cloud Run.

Usage

Training a ML model usually require multiple databases access. Although it is doable to instantiate a connection to Postgresql or Elasticsearch everytime one needs to, it can actually become quickly cumbersome to do it for a certain number or models. To make everything a tiny bit more DRY, pynidus wraps very small chunks of code to deal with config files and clients instantiation.

from pynidus.base import MLTBase

config = {
    "postgresql": {
        "host": "host",
        "user": "user",
        "password": "password",
        "database": "database"
    },
    "elasticsearch": {
        "host": "host",
        "user": "user",
        "password": "password"
    }
}

class TestClassifier(MLTBase):

    def __init__(self, **kwargs):
        super().__init__(**kwargs)

    def some_method_fetching_data_from_db(self):
        return self.pg_client.query("SELECT * FROM table LIMIT 1")

    def some_other_method_fetching_data_from_es(self):
        return self.es_client.query("...")

testclassifier = TestClassifier(
    pg_config = config.get("postgresql"),
    es_config = config.get("elasticsearch")
)

TODO

• Add a dev branch