mlecho 0.0.1__tar.gz

mlecho-0.0.1/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Thanuja Liyanage
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

mlecho-0.0.1/PKG-INFO

@@ -0,0 +1,55 @@
+Metadata-Version: 2.4
+Name: mlecho
+Version: 0.0.1
+Summary: Real-time CNN-VAE poisoning detection for MLOps pipelines
+Author: Thanuja Liyanage
+License: MIT
+Project-URL: Repository, https://github.com/justthzz/mlecho
+Project-URL: Issues, https://github.com/justthzz/mlecho/issues
+Keywords: machine-learning,mlops,data-poisoning,anomaly-detection,mlflow,vae,reinforcement-learning
+Classifier: Development Status :: 2 - Pre-Alpha
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Science/Research
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Security
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Dynamic: license-file
+
+# MLEcho
+
+**A Real-Time CNN-VAE and Reinforcement Learning Framework for Detecting and Responding to Training Data Poisoning in MLOps Pipelines.**
+
+> ⚠️ **Work in progress.** MLEcho is under active development as part of an ongoing Final Year Project (FYP). Core detection, classification, and response modules are not yet functional — see [Status](#status) below.
+
+## What is MLEcho?
+
+MLEcho monitors training batches as they stream through an MLOps pipeline, flagging poisoned data during active training rather than before or after it. It combines:
+
+- A **CNN feature extractor** and **VAE anomaly scorer** that flag anomalous batches by reconstruction error — clean batches reconstruct faithfully (a true "echo"), poisoned batches produce a distorted one.
+- An **attack-type classifier** that identifies the likely poisoning strategy (backdoor, label-flip, gradient-based, accumulative).
+- A **deterministic rule-based response module** (quarantine, rollback, retrain, escalate).
+- *(Research enhancement, in progress)* A **PPO-based reinforcement-learning response agent**, evaluated against the rule-based baseline.
+
+MLEcho integrates with **MLflow**, **Weights & Biases**, and **Docker** for experiment tracking, live monitoring, and reproducible deployment.
+
+## Installation
+
+```bash
+pip install mlecho
+```
+
+## Status
+
+This package currently ships a placeholder API only. Calling `mlecho.detect()` raises `NotImplementedError`. Follow the repository for progress as the core CNN-VAE detector, classifier, and response modules are implemented.
+
+## License
+
+MIT — see [LICENSE](LICENSE).

mlecho-0.0.1/README.md

@@ -0,0 +1,30 @@
+# MLEcho
+
+**A Real-Time CNN-VAE and Reinforcement Learning Framework for Detecting and Responding to Training Data Poisoning in MLOps Pipelines.**
+
+> ⚠️ **Work in progress.** MLEcho is under active development as part of an ongoing Final Year Project (FYP). Core detection, classification, and response modules are not yet functional — see [Status](#status) below.
+
+## What is MLEcho?
+
+MLEcho monitors training batches as they stream through an MLOps pipeline, flagging poisoned data during active training rather than before or after it. It combines:
+
+- A **CNN feature extractor** and **VAE anomaly scorer** that flag anomalous batches by reconstruction error — clean batches reconstruct faithfully (a true "echo"), poisoned batches produce a distorted one.
+- An **attack-type classifier** that identifies the likely poisoning strategy (backdoor, label-flip, gradient-based, accumulative).
+- A **deterministic rule-based response module** (quarantine, rollback, retrain, escalate).
+- *(Research enhancement, in progress)* A **PPO-based reinforcement-learning response agent**, evaluated against the rule-based baseline.
+
+MLEcho integrates with **MLflow**, **Weights & Biases**, and **Docker** for experiment tracking, live monitoring, and reproducible deployment.
+
+## Installation
+
+```bash
+pip install mlecho
+```
+
+## Status
+
+This package currently ships a placeholder API only. Calling `mlecho.detect()` raises `NotImplementedError`. Follow the repository for progress as the core CNN-VAE detector, classifier, and response modules are implemented.
+
+## License
+
+MIT — see [LICENSE](LICENSE).

mlecho-0.0.1/pyproject.toml

@@ -0,0 +1,43 @@
+[build-system]
+requires = ["setuptools>=61.0"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "mlecho"
+version = "0.0.1"
+description = "Real-time CNN-VAE poisoning detection for MLOps pipelines"
+readme = "README.md"
+requires-python = ">=3.9"
+license = { text = "MIT" }
+authors = [
+    { name = "Thanuja Liyanage" }
+]
+keywords = [
+    "machine-learning",
+    "mlops",
+    "data-poisoning",
+    "anomaly-detection",
+    "mlflow",
+    "vae",
+    "reinforcement-learning",
+]
+classifiers = [
+    "Development Status :: 2 - Pre-Alpha",
+    "Intended Audience :: Developers",
+    "Intended Audience :: Science/Research",
+    "License :: OSI Approved :: MIT License",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.9",
+    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Topic :: Scientific/Engineering :: Artificial Intelligence",
+    "Topic :: Security",
+]
+
+[project.urls]
+Repository = "https://github.com/justthzz/mlecho"
+Issues = "https://github.com/justthzz/mlecho/issues"
+
+[tool.setuptools.packages.find]
+where = ["src"]

mlecho-0.0.1/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

mlecho-0.0.1/src/mlecho/__init__.py

@@ -0,0 +1,35 @@
+"""MLEcho: Real-Time CNN-VAE and Reinforcement Learning Framework for 
+Detecting and Responding to Training Data Poisoning in MLOps Pipelines.
+
+MLEcho monitors training batches as they stream through an MLOps
+pipeline. A CNN feature extractor and VAE anomaly scorer flag batches
+whose reconstruction ("echo") deviates significantly from clean,
+calibrated training data. Flagged batches are classified by attack
+type and handled by a deterministic rule-based response module
+(quarantine, rollback, retrain, escalate), with a PPO-based
+reinforcement-learning response agent under investigation as a
+research enhancement.
+
+This is an early-stage release. Core detection, classification, and
+response modules are under active development as part of an ongoing
+Final Year Project (FYP). See the project README and repository for
+current status.
+"""
+
+__version__ = "0.0.1"
+__all__ = ["detect", "__version__"]
+
+
+def detect(*args, **kwargs):
+    """Run poisoning detection on a training batch.
+
+    Not yet implemented. This is a placeholder for the CNN-VAE
+    detection pipeline, which is under active development.
+
+    Raises:
+        NotImplementedError: Always, until the core detector ships.
+    """
+    raise NotImplementedError(
+        "MLEcho's core detection module is still in development. "
+        "Track progress at the project repository."
+    )

mlecho-0.0.1/src/mlecho.egg-info/PKG-INFO

@@ -0,0 +1,55 @@
+Metadata-Version: 2.4
+Name: mlecho
+Version: 0.0.1
+Summary: Real-time CNN-VAE poisoning detection for MLOps pipelines
+Author: Thanuja Liyanage
+License: MIT
+Project-URL: Repository, https://github.com/justthzz/mlecho
+Project-URL: Issues, https://github.com/justthzz/mlecho/issues
+Keywords: machine-learning,mlops,data-poisoning,anomaly-detection,mlflow,vae,reinforcement-learning
+Classifier: Development Status :: 2 - Pre-Alpha
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Science/Research
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Security
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Dynamic: license-file
+
+# MLEcho
+
+**A Real-Time CNN-VAE and Reinforcement Learning Framework for Detecting and Responding to Training Data Poisoning in MLOps Pipelines.**
+
+> ⚠️ **Work in progress.** MLEcho is under active development as part of an ongoing Final Year Project (FYP). Core detection, classification, and response modules are not yet functional — see [Status](#status) below.
+
+## What is MLEcho?
+
+MLEcho monitors training batches as they stream through an MLOps pipeline, flagging poisoned data during active training rather than before or after it. It combines:
+
+- A **CNN feature extractor** and **VAE anomaly scorer** that flag anomalous batches by reconstruction error — clean batches reconstruct faithfully (a true "echo"), poisoned batches produce a distorted one.
+- An **attack-type classifier** that identifies the likely poisoning strategy (backdoor, label-flip, gradient-based, accumulative).
+- A **deterministic rule-based response module** (quarantine, rollback, retrain, escalate).
+- *(Research enhancement, in progress)* A **PPO-based reinforcement-learning response agent**, evaluated against the rule-based baseline.
+
+MLEcho integrates with **MLflow**, **Weights & Biases**, and **Docker** for experiment tracking, live monitoring, and reproducible deployment.
+
+## Installation
+
+```bash
+pip install mlecho
+```
+
+## Status
+
+This package currently ships a placeholder API only. Calling `mlecho.detect()` raises `NotImplementedError`. Follow the repository for progress as the core CNN-VAE detector, classifier, and response modules are implemented.
+
+## License
+
+MIT — see [LICENSE](LICENSE).

mlecho-0.0.1/src/mlecho.egg-info/SOURCES.txt

@@ -0,0 +1,8 @@
+LICENSE
+README.md
+pyproject.toml
+src/mlecho/__init__.py
+src/mlecho.egg-info/PKG-INFO
+src/mlecho.egg-info/SOURCES.txt
+src/mlecho.egg-info/dependency_links.txt
+src/mlecho.egg-info/top_level.txt

mlecho-0.0.1/src/mlecho.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

mlecho-0.0.1/src/mlecho.egg-info/top_level.txt

@@ -0,0 +1 @@
+mlecho