Add integration code for restricted algorithm implementations (#50)

* feat: add integration code for restricted algorithm implementations

* fix: base Dockerimage refs

Author: Sebastian Schmidl

arima/.gitignore

@@ -0,0 +1,151 @@
+# ingore protected source code
+ptsa/
+setup.py
+
+########################################
+# Python.gitignore from github/gitignore
+########################################
+
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+########################################
+
+
+**.png

arima/Dockerfile

@@ -0,0 +1,11 @@
+FROM ghcr.io/timeeval/python3-base:0.3.0
+
+LABEL maintainer="thorsten.papenbrock@hpi.de"
+
+ENV ALGORITHM_MAIN="/app/algorithm.py"
+
+COPY requirements.txt /app/
+RUN pip install -r /app/requirements.txt;
+
+COPY ptsa /app/ptsa
+COPY algorithm.py /app/

arima/README.md

@@ -8,9 +8,10 @@
 | Source Code | https://github.com/johnpaparrizos/AnomalyDetection/tree/master/code/ptsa |
 | Learning type | unsupervised |
 | Input dimensionality | univariate |
-
 |||
 
+After receiving the original source code from the authors, place the directory `ptsa` into this folder. 
+
 ## Notes
 
 The ptsa algorithms require sklearn in version 19 to 23. This is checked in the utility.py. Our python image, however, uses a newer sklearn version, which is 24.1 or higher. Hence we removed the check:

arima/algorithm.py

@@ -0,0 +1,128 @@
+#!/usr/bin/env python3
+
+import json
+import sys
+import argparse
+import numpy as np
+
+from dataclasses import dataclass
+
+from ptsa.models.arima import ARIMA
+from ptsa.models.distance import Euclidean
+from ptsa.models.distance import Mahalanobis
+from ptsa.models.distance import Garch
+from ptsa.models.distance import SSA
+from ptsa.models.distance import Fourier
+from ptsa.models.distance import DTW
+from ptsa.models.distance import EDRS
+from ptsa.models.distance import TWED
+
+
+@dataclass
+class CustomParameters:
+    window_size: int = 20
+    max_lag: int = 30000
+    p_start: int = 1
+    q_start: int = 1
+    max_p: int = 5
+    max_q: int = 5
+    differencing_degree: int = 0
+    distance_metric: str = "Euclidean"
+    random_state: int = 42  # seed for randomness
+
+
+class AlgorithmArgs(argparse.Namespace):
+    @staticmethod
+    def from_sys_args() -> 'AlgorithmArgs':
+        if len(sys.argv) != 2:
+            raise ValueError("Wrong number of arguments specified! Single JSON-string pos. argument expected.")
+        args: dict = json.loads(sys.argv[1])
+        custom_parameter_keys = dir(CustomParameters())
+        filtered_parameters = dict(filter(lambda x: x[0] in custom_parameter_keys, args.get("customParameters", {}).items()))
+        args["customParameters"] = CustomParameters(**filtered_parameters)
+        return AlgorithmArgs(**args)
+
+
+def set_random_state(config: AlgorithmArgs) -> None:
+    seed = config.customParameters.random_state
+    import random
+    random.seed(seed)
+    np.random.seed(seed)
+
+
+def distance_to_measure(distance_metric):
+    switcher = {
+        "euclidean": Euclidean(),
+        "mahalanobis": Mahalanobis(),
+        "garch": Garch(),
+        "ssa": SSA(),
+        "fourier": Fourier(),
+        "dtw": DTW(),
+        "edrs": EDRS(),
+        "twed": TWED()
+    }
+    return switcher.get(distance_metric.lower(), "missing")
+
+
+def main():
+    config = AlgorithmArgs.from_sys_args()
+    ts_filename = config.dataInput  # "/data/dataset.csv"
+    score_filename = config.dataOutput  # "/results/anomaly_window_scores.ts"
+
+    print(f"Configuration: {config}")
+
+    if config.executionType == "train":
+        print("No training required!")
+        exit(0)
+
+    if config.executionType != "execute":
+        raise ValueError("Unknown executionType specified!")
+
+    set_random_state(config)
+
+    # read only single "value" column from dataset
+    print(f"Reading data from {ts_filename}")
+    da = np.genfromtxt(ts_filename, skip_header=1, delimiter=",")
+    data = da[:, 1]
+    labels = da[:, -1]
+    length = len(data)
+    contamination = labels.sum() / length
+    # Use smallest positive float as contamination if there are no anomalies in dataset
+    contamination = np.nextafter(0, 1) if contamination == 0. else contamination
+
+    # run ARIMA
+    print("Executing ARIMA ...")
+    model = ARIMA(
+       window=config.customParameters.window_size,
+       max_lag=config.customParameters.max_lag,
+       p_start=config.customParameters.p_start,
+       q_start=config.customParameters.q_start,
+       max_p=config.customParameters.max_p,
+       max_q=config.customParameters.max_q,
+       d=config.customParameters.differencing_degree,
+       contamination=contamination,
+       neighborhood="all")
+    model.fit(data)
+
+    # get outlier scores
+    measure = distance_to_measure(config.customParameters.distance_metric)
+    if measure == "missing":
+        raise ValueError(f"Distance measure '{config.customParameters.distance_metric}' not supported!")
+    measure.detector = model
+    measure.set_param()
+    model.decision_function(measure=measure)
+    scores = model.decision_scores_
+
+    #from ptsa.utils.metrics import metricor
+    #grader = metricor()
+    #preds = grader.scale(scores, 0.1)
+
+    print(f"Input size: {len(data)}\nOutput size: {len(scores)}")
+    print("ARIMA result:", scores)
+
+    print(f"Writing results to {score_filename}")
+    np.savetxt(score_filename, scores, delimiter=",")
+
+
+if __name__ == "__main__":
+    main()

arima/requirements.txt

@@ -0,0 +1,15 @@
+combo
+joblib
+matplotlib
+numpy>=1.13
+numba>=0.35
+pandas>=0.25
+scipy>=0.19.1
+scikit_learn>=0.19.1
+six
+statsmodels
+suod
+pmdarima
+arch
+tsfresh
+hurst