Merge branch 'tft_nbeats' into 'morphemic-rc2.0'

Ensembler - influx integration

See merge request !276

deployment/ensembler/.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
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.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/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/

deployment/ensembler/Dockerfile

+FROM python:3.8-slim-buster
+
+# Install Python dependencies.
+WORKDIR /wd
+
+COPY deployment/ensembler/poetry.lock deployment/ensembler/pyproject.toml /wd/
+
+RUN apt-get update && apt-get install -y build-essential g++ libgl1-mesa-glx libx11-6 \ 
+    && pip install --no-cache-dir --upgrade pip \
+    && pip install --no-cache-dir poetry \
+    \
+    && poetry install --no-dev && mkdir models
+
+
+ADD https://gitlab.ow2.org/melodic/morphemic-preprocessor/-/archive/morphemic-rc1.5/morphemic-preprocessor-morphemic-rc1.5.tar.gz /var/lib/morphemic/
+
+# Copy the rest of the codebase into the image
+COPY deployment/ensembler/ ./
+
+RUN cd /var/lib/morphemic/ \
+    && tar -zxf morphemic-preprocessor-morphemic-rc1.5.tar.gz \
+    && cd morphemic-preprocessor-morphemic-rc1.5 \
+    && cd morphemic-datasetmaker && python3 setup.py install \ 
+    && cd ../.. \
+    && cp -R /var/lib/morphemic/morphemic-preprocessor-morphemic-rc1.5/amq-message-python-library /wd/amq_message_python_library \
+    && rm -rf /var/lib/morphemic \
+    && mkdir -p /wd/logs
+
+
+CMD ["poetry", "run" , "uvicorn", "app:app", "--host", "0.0.0.0", "--port", "8000"]
+
+# For local tests:
+# CMD ["poetry", "run" , "uvicorn", "app:app", "--reload"]
+
+

deployment/ensembler/README.md

+# Ensembler 
+
+To test locally:
+
+    cd ../../
+    docker build -t ensembler -f ./deployment/ensembler/Dockerfile .
+    docker run -t --env-file=deployment/ensembler/env -v <LOGS PATH>:/wd/logs --network=host ensembler
+
+Go to local amw web cosnole: http://localhost:8161/admin
+
+send example start_ensembler message:
+
+    {
+    "metrics":[
+        {
+            "metric":"MaxCPULoad",
+            "level":3,
+            "publish_rate":60000
+        },
+        {
+            "metric":"MinCPULoad",
+            "level":3,
+            "publish_rate":50000
+        }
+    ],
+    "models":[
+        "tft",
+        "nbeats",
+        "gluon"
+    ]
+    }
+
+
+send ensemble request:
+
+    curl -i http://127.0.0.1:8000/ensemble -X POST -H 'Content-Type: application/json' -d '{"method":"BestSubset", "metric": "MaxCPULoad", "predictionTime": 1234567, "predictionsToEnsemble": {"tft": 0, "nbeats": null, "gluon": 9}}' -w '\n'
+
+
+
+

deployment/ensembler/app.py

+"""main script for ensembler module"""
+
+import logging
+from datetime import datetime
+
+import setproctitle
+from fastapi import FastAPI
+from pytz import timezone
+
+from ensembler.env_config import create_env_config
+from ensembler.messages_schemas import EnsembleResponse, Prediction
+from ensembler.services import AMQService
+
+env_config = create_env_config()
+setproctitle.setproctitle("Ensembler")
+logging.basicConfig(
+    filename=f"logs/{env_config['LOGING_FILE_NAME']}.out",
+    level=logging.INFO,
+    datefmt="%Y-%m-%d %H:%M:%S",
+    format="%(asctime)s.%(msecs)03d %(levelname)s %(module)s - %(funcName)s: %(message)s",
+)
+logging.Formatter.converter = lambda *args: datetime.now(
+    tz=timezone(env_config["TZ"])
+).timetuple()
+
+amq_service = AMQService(env_config)
+
+app = FastAPI()
+
+
+@app.post("/ensemble", response_model=EnsembleResponse)
+async def add_country(prediction: Prediction):
+    """Function for returning ensembled value on request"""
+    return amq_service.ens.on_ensemble(prediction.dict())
+
+
+log = logging.getLogger()
+log.info(f"Ensebler service started for application: {env_config['APP_NAME']}")

deployment/ensembler/ensembler/__init__.py

+__version__ = '0.1.0'

deployment/ensembler/ensembler/dataset/__init__.py

+__version__ = '0.1.0'

deployment/ensembler/ensembler/dataset/data.py

+"""Script for data frame with predictions class
+"""
+
+import time
+
+
+class PredictionsDF:
+    """Predictions data frame class
+    df: data frame with predictions and real value,
+    target_column: target column name"""
+
+    def __init__(self, data_frame, target_column="y"):
+        """Init method"""
+        self.df = data_frame
+        self.last_update_time = int(time.time())
+        self.target_column = target_column
+
+    def update(self, data_frame):
+        """Upade method, change df, and updates last_update_time"""
+        self.data_frame = data_frame
+        self.last_update_time = int(time.time())

deployment/ensembler/ensembler/dataset/download_data.py

+"""Script for downloading data from influx
+"""
+
+import os
+
+import pandas as pd
+from influxdb import DataFrameClient
+
+
+class InfluxDataDownloader:
+    def __init__(
+        self,
+    ) -> None:
+        """class for downloading data from inlux,
+        necessary are columns with predictions and
+        real values
+        """
+
+        self.influx_client = DataFrameClient(
+            host=os.environ.get("INFLUXDB_HOSTNAME", "localhost"),
+            port=int(os.environ.get("INFLUXDB_PORT", "8086")),
+            username=os.environ.get("INFLUXDB_USERNAME", "morphemic"),
+            password=os.environ.get("INFLUXDB_PASSWORD", "password"),
+        )
+        self.influx_client.switch_database(
+            os.environ.get("INFLUXDB_DBNAME", "morphemic")
+        )
+
+    @staticmethod
+    def convert_timestamp(data_frame: pd.DataFrame) -> pd.DataFrame:
+        """convert date index to desired format
+
+        Args:
+        -------
+            data_frame (pd.DataFrame): data frame with
+            time index (pandas time index)
+
+        Returns:
+        -------
+            pd.DataFrame: data frame with date index
+            with desired format
+        """
+        return pd.to_datetime(data_frame.index, unit="s").tz_convert(None)
+
+    def download_predictions(self, metric_name: str) -> pd.DataFrame:
+        """Download predicted values from influx
+
+        Returns:
+        -------
+            pd.DataFrame: pandas data
+             frame with predictions values
+        """
+        return self.influx_client.query(
+            f'SELECT * FROM "{metric_name}Predictions" WHERE time > now() - {os.environ.get("MAX_PAST_DAYS", 100)}d'
+        )[f"{metric_name}Predictions"]
+
+    def download_real(self, start_time: pd.DatetimeIndex) -> pd.DataFrame:
+        """Download real values from influx
+
+        Args:
+        -------
+            start_time (pd.DatetimeIndex): first
+            date with predictions,
+
+        Returns:
+        -------
+            pd.DataFrame: pandas data
+             frame with real values from PS
+        """
+        return self.influx_client.query(
+            f'SELECT * FROM "{os.environ.get("APP_NAME", "default_application")}" WHERE time > {start_time}'
+        )[os.environ.get("APP_NAME", "default_application")]
+
+    def download_data(self, metric_name: str, predictions_freq: int) -> pd.DataFrame:
+        """
+        Download data from inlux
+        (2 tables one with predictions, second with
+        real values from PS), merge data and save them to csv
+
+        Args:
+        ------
+            metric_name (str): metric name
+            predictions_freq (int): predictions
+            frequency (in seconds)
+
+        Returns:
+        -------
+            pd.DataFrame: pandas data
+             frame with real and predicted values
+        """
+        predictions = self.download_predictions(metric_name)
+        predictions = self.convert_timestamp(predictions)
+        start_time = predictions.index.values[0]
+
+        real = self.download_data(start_time)
+        real.index = real["ems_time"]
+        real = self.convert_timestamp(real)
+
+        predictions = predictions.resample(
+            f"{predictions_freq}S", origin=start_time
+        ).mean()
+        real = (
+            real.resample(f"{predictions_freq}S", origin=start_time)
+            .mean()
+            .rename({metric_name: "y"}, axis=1)
+        )["y"]
+
+        merged = pd.merge(
+            predictions,
+            real,
+            how="left",
+            left_index=True,
+            right_index=True,
+        ).dropna(subset=["y"])
+
+        return merged

deployment/ensembler/ensembler/dataset/preprocess_dataset.py

+import logging
+import time
+
+import numpy as np
+import pandas as pd
+from pytorch_forecasting import TimeSeriesDataSet
+from pytorch_forecasting.data import NaNLabelEncoder
+
+pd.options.mode.chained_assignment = None
+
+"""Script for preparing time series dataset from pythorch-forecasting package 
+TODO: add checking whether data consists of multiple series, handle nans values"""
+
+
+class Dataset(object):
+    def __init__(
+        self,
+        dataset,
+        target_column="value",
+        time_column="ems_time",
+        tv_unknown_reals=[],
+        known_reals=[],
+        tv_unknown_cat=[],
+        static_reals=[],
+        classification=0,
+        context_length=40,
+        prediction_length=5,
+        publish_rate=10000,
+    ):
+
+        self.max_missing_values = (
+            20  # max consecutive missing values allowed per series
+        )
+        self.target_column = target_column
+        self.time_column = time_column
+        self.tv_unknown_cat = tv_unknown_cat
+        self.known_reals = known_reals
+        self.tv_unknown_reals = tv_unknown_reals
+        self.static_reals = static_reals
+        self.classification = classification
+        self.context_length = context_length
+        self.prediction_length = prediction_length
+        self.publish_rate = publish_rate
+        self.dataset = dataset
+        self.dropped_recent_series = True  # default set to be true
+        if self.dataset.shape[0] > 0:
+            self.check_gap()
+        self.n = dataset.shape[0]
+        if self.dataset.shape[0] > 0:
+            self.ts_dataset = self.create_time_series_dataset()
+
+    def cut_nan_start(self, dataset):
+        dataset.index = range(dataset.shape[0])
+        first_not_nan_index = dataset[self.target_column].first_valid_index()
+        if first_not_nan_index == first_not_nan_index:  # check is if it;s not np.nan
+            if first_not_nan_index is not None:
+                if first_not_nan_index > -1:
+                    return dataset[dataset.index > first_not_nan_index]
+        else:
+            return dataset.dropna()
+
+    def fill_na(self, dataset):
+        dataset = dataset.replace(np.inf, np.nan)
+        dataset = dataset.ffill(axis="rows")
+        return dataset
+
+    def convert_formats(self, dataset):
+        if not self.classification:
+            dataset[self.target_column] = dataset[self.target_column].astype(float)
+        else:
+            dataset[self.target_column] = dataset[self.target_column].astype(int)
+
+        for name in self.tv_unknown_cat:
+            dataset[name] = dataset[name].astype(str)
+        return dataset
+
+    def convert_time_to_ms(self):
+        if self.dataset.shape[0] > 0:
+            digit_len = len(str(int(self.dataset[self.time_column].values[0])))
+            if digit_len >= 13:
+                self.dataset[self.time_column] = self.dataset[self.time_column].apply(
+                    lambda x: int(str(int(x))[:13])
+                )
+            else:
+                self.dataset[self.time_column] = self.dataset[self.time_column].apply(
+                    lambda x: int(int(str(int(x))[:digit_len]) * 10 ** (13 - digit_len))
+                )
+            self.dataset[self.time_column] = self.dataset[self.time_column].apply(
+                lambda x: int(x // 1e4 * 1e4)
+            )
+
+    def add_obligatory_columns(self, dataset):
+        n = dataset.shape[0]
+        dataset["time_idx"] = range(n)  # TODO check time gaps
+        return dataset
+
+    def get_time_difference_current(self):
+        if self.dataset.shape[0] > 0:
+            last_timestamp_database = self.dataset[self.time_column].values[-1]
+            current_time = int(time.time())
+            print(
+                f"Time difference between last timestamp and current time: {current_time - last_timestamp_database}"
+            )
+            logging.info(
+                f"Time difference between last timestamp and current time: {current_time - last_timestamp_database}"
+            )
+
+    def check_gap(self):
+        if (self.dataset.shape[0] > 0) and (self.target_column in self.dataset.columns):
+            self.dataset = self.dataset.groupby(by=[self.time_column]).min()
+            self.dataset[self.time_column] = self.dataset.index
+            self.dataset.index = range(self.dataset.shape[0])
+            self.convert_time_to_ms()
+            self.dataset[self.target_column] = pd.to_numeric(
+                self.dataset[self.target_column], errors="coerce"
+            ).fillna(np.nan)
+            self.dataset = self.dataset.replace(np.inf, np.nan)
+            self.dataset = self.dataset.dropna(subset=[self.target_column])
+            if self.dataset.shape[0] > 0:
+                max_gap = self.dataset[self.time_column].diff().abs().max()
+                logging.info(
+                    f"Metric: {self.target_column} Max time gap in series {max_gap}"
+                )
+                print(f" Metric: {self.target_column} Max time gap in series {max_gap}")
+                series_freq = (
+                    (self.dataset[self.time_column])
+                    .diff()
+                    .fillna(0)
+                    .value_counts()
+                    .index.values[0]
+                )
+
+                logging.info(
+                    f"Metric: {self.target_column} Detected series with {series_freq} frequency"
+                )
+                print(
+                    f"Metric: {self.target_column} Detected series with {series_freq} frequency"
+                )
+                if series_freq != self.publish_rate:
+                    logging.info(
+                        f"Metric: {self.target_column} Detected series with {series_freq} frequency, but the frequency should be: {self.publish_rate}!"
+                    )
+                    print(
+                        f"Metric: {self.target_column} Detected series with {series_freq} frequency, but the frequency should be: {self.publish_rate}!"
+                    )
+
+                # check series length
+                series = np.split(
+                    self.dataset,
+                    *np.where(
+                        self.dataset[self.time_column]
+                        .diff()
+                        .abs()
+                        .fillna(0)
+                        .astype(int)
+                        >= np.abs(self.max_missing_values * self.publish_rate)
+                    ),
+                )
+                logging.info(f"Metric: {self.target_column} {len(series)} series found")
+                print(f"{len(series)} series found")
+                preprocessed_series = []
+                for i, s in enumerate(series):
+                    s = self.fill_na(s)
+                    s = self.cut_nan_start(s)
+                    s = self.add_obligatory_columns(s)
+                    s["split"] = "train"
+                    s = self.convert_formats(s)
+                    logging.info(
+                        f"Metric: {self.target_column} Found series {i} of length: {s.shape[0]}, required data rows: {self.prediction_length * 2 + self.context_length}"
+                    )
+                    if s.shape[0] > self.prediction_length * 2 + self.context_length:
+                        s["series"] = i
+                        preprocessed_series.append(s)
+                    if i == len(series) - 1:
+                        logging.info(
+                            f"Metric: {self.target_column} Fresh data rows: {s.shape[0]}, required fresh data rows: {self.prediction_length * 2 + self.context_length}"
+                        )
+
+                logging.info(
+                    f"Metric: {self.target_column} {len(preprocessed_series)} long enough series found"
+                )
+                print(f"{len(preprocessed_series)} long enough series found")