flops-utils

This package/library contains common pieces of code for the FLOps project.

User Facing Features

flops-utils provides FLOPs users with guiding abstract template classes for implementing their ML models and "proxy/place-holder" components to satisfy linters.

When FLOps users want to structure their ML code to match the structural requirements of FLOps they should do the following:

# data_manager.py

from flops_utils.ml_repo_building_blocks import load_dataset
from flops_utils.ml_repo_templates import DataManagerTemplate

class DataManager(DataManagerTemplate):
    def __init__(self):
        ... = self._prepare_data()

    def _prepare_data(self, partition_id=1) -> Any:
        dataset = load_dataset()
        ...
        return ...

    def get_data(self) -> Tuple[Any, Any]:
        return ...

# model_manager.py

import warnings
from typing import Any, Tuple

import numpy as np
from data_manager import DataManager
from flops_utils.ml_repo_templates import ModelManagerTemplate
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import log_loss


class ModelManager(ModelManagerTemplate):
    def __init__(self):
        self.model = ...
        self._set_init_params()

    def _set_init_params(self) -> None:
        ...

    def set_model_data(self) -> None:
        ... = DataManager().get_data()

    def get_model(self) -> Any:
        return self.model

    def get_model_parameters(self) -> Any:
        params = ...
        return params

    def set_model_parameters(self, parameters) -> None:
        ...

    def fit_model(self) -> int:
        return len(self.x_train)

    def evaluate_model(self) -> Tuple[Any, Any, int]:
        loss = ...
        accuracy = ...
        return loss, accuracy, len(self.x_test)

A tangible example implementation is available here.

Internal FLOps Commonalities

Logger
Types
MQTT Topics
Handlers
- Notifications (Project Observer & FLOps Manager)
- Environment Variables
Auxiliary Constructs
- Running Shell Commands
- Measuring Execution Time Frames

Installation Alternatives

One can also include and use these utilities by doing the following:

pip install --no-cache-dir git+https://github.com/oakestra/addon-FLOps.git@main#subdirectory=utils_library

Metadata

Release files for flops-utils 0.6.3

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for flops-utils 0.6.3
File	Size	Uploaded
flops_utils-0.6.3.tar.gz	6.7 kB	Jan 24, 2025	Details

Built distribution (wheel)

Table of built distributions (wheels) for flops-utils 0.6.3
File	Interpreter	ABI	Platform	Reset
flops_utils-0.6.3-py3-none-any.whl 10.3 kB Jan 24, 2025	Python 3	none	any	Details

Total release size: 17.0 kB

Release files / flops_utils-0.6.3.tar.gz

Download URL	flops_utils-0.6.3.tar.gz
Size	6.7 kB
Tags	Source
SHA-256 checksum How to use checksums	`545182b2994f4305def33fc2ee1e188493dceb4f0db8378efa9fccf1a8f8cd0c`
BLAKE2b-256 checksum How to use checksums	`0b65f9df95a64719643be1221a1ec80abad3f75ad158ec938b1bd303616459d6`
Upload date	Jan 24, 2025
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/6.1.0 CPython/3.12.2`

Release files / flops_utils-0.6.3-py3-none-any.whl

Download URL	flops_utils-0.6.3-py3-none-any.whl
Size	10.3 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`399168bde810a00c15d223d0a73761868d7c5610e8c3d9cb64aafdb7af521378`
BLAKE2b-256 checksum How to use checksums	`7a0acd1b2060046e5360016626834cad0441a23d557f1774bdc8f5969d18354a`
Upload date	Jan 24, 2025
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/6.1.0 CPython/3.12.2`