algomancy-data 0.4.1

Data management model for the Algomancy library

algomancy-data

Data layer for Algomancy dashboards: schemas, extract/transform/load (ETL) primitives, validators, and data containers used by the GUI and scenario packages.

Features

• DataSource and BaseDataSource containers with table management and JSON (de)serialization
• Pluggable ETL pipeline building blocks: Extractor, Transformer, Validator, Loader
• DataManager orchestrators (stateful/stateless) to drive ETL and manage datasets
• Declarative InputFileConfiguration for file inputs (CSV, XLSX, JSON)

Installation

pip install -e packages/algomancy-data

Requires Python >= 3.14. Core dependency: pandas.

Quick start: use DataSource directly

import pandas as pd
from algomancy_data import DataSource, DataClassification

ds = DataSource(ds_type=DataClassification.MASTER_DATA, name="warehouse")
ds.add_table("inventory", pd.DataFrame({"sku": ["A", "B"], "qty": [10, 5]}))

# JSON roundtrip
json_str = ds.to_json()
ds2 = DataSource.from_json(json_str)
assert ds2.get_table("inventory").equals(ds.get_table("inventory"))

Quick start: orchestrate ETL with a DataManager

DataManager wires Extractor → Transformer → Validator → Loader. You provide an ETLFactory that builds these parts for each input configuration.

from typing import List
from algomancy_data import (
    DataSource, DataClassification,
    DataManager, StatelessDataManager, ETLFactory,
    SingleInputFileConfiguration, FileExtension
)

class MyETLFactory(ETLFactory):
    # Implement factory methods to build Extractor/Transformer/Validator/Loader
    ...

input_cfgs: List[SingleInputFileConfiguration] = [
    SingleInputFileConfiguration(
        tag="inventory", file_name="inventory", extension=FileExtension.CSV
    )
]

dm: DataManager = StatelessDataManager(
    etl_factory=MyETLFactory,
    input_configs=input_cfgs,
    save_type="json",  # or other configured type
    data_object_type=DataSource,
)

files = dm.prepare_files(file_items_with_path=[("inventory", "./data/inventory.csv")])
ds: DataSource = dm.etl_data(files=files, dataset_name="warehouse")

Documentation and examples

• Root docs: documentation/1_data.md
• End‑to‑end usage in the example app: example/ (see example/data_handling and example/main.py)