Basic Utility module for the Python programming language

pip install pbu==0.6.11


Python Basic Utilities pbu

Available on PyPi

Table of Contents

  1. Installation
  2. Usage
  3. Classes
    1. JSON - a JavaScript-like dictionary access helper
    2. Logger - a wrapper around the Python logging framework
    3. TimeSeries - powerful helper class to organise time series
    4. AbstractMongoStore - helper and wrapper class for MongoDB access
    5. AbstractMysqlStore - helper and wrapper class for MySQL access
    6. BasicMonitor - monitor class orchestrating regular operations
    7. ConstantListing - a parent class allowing to fetch attribute values from a constant class
  4. Functions
    1. list_to_json
    2. default_options
    3. default_value


Install via pip:

pip install pbu


Optional: If you have a requirement.txt file, you can add pbu:


Then, simply import the class / module you need:

from pbu import JSON

# and start using it
obj = JSON({"my": {"obj": "content"}})



This is an adaptation of the native dict class, providing Javascript-like dictionary access using the "dot-notation" (e.g. person.relations[0].address.street) rather than the Python-native bracket notation (e.g. person["relations"][0]["address"]["street"]). It overrides the basic __getattr__ and __setattr__ methods as a shortcut to manage the dictionary content.


from pbu import JSON
my_obj = JSON({"initial": "content"})
# prints out "content"

my_obj.initial = {"a": 5, "b": 3}
print(my_obj.initial.a + my_obj.initial.b)
# prints out 8
my_obj.initial.b = 13
print(my_obj.initial.a + my_obj.initial.b)
# prints out 18

my_obj.extension = 10
# prints out 10


This is a basic logger allowing to write log files, for it writes a debug.log and for logger.error or logger.exception it writes an error.log file.


from pbu import Logger

logger = Logger(name="logger-name")
logger.debug("Some debug message goes here")
logger.error("Error executing something")

logger = Logger(name="logger-name", log_folder="./logs")
logger.debug("This will create the debug.log and error.log in the ./logs folder")


The time series class is a helper utility, that allows to compile complex time-series, offering functionality to add time series, remove time series and most importantly align time series with timestamps to a previously defined resolution by interpolating missing values and re-aligning measurements within the tolerance of the provided time series.

It supports 2 different structures:

List of Dictionary Items

from datetime import datetime, timedelta

list_of_dict = [
    { "date_time":, "measurement_1": 12, "measurement_2": 15 },
    { "date_time": + timedelta(hours=1), "measurement_1": 10, "measurement_2": 16 },
    { "date_time": + timedelta(hours=2), "measurement_1": 9, "measurement_2": 12 },

Dictionary of Lists

from datetime import datetime, timedelta

dict_of_list = {
    "date_time": [, + timedelta(hours=1), datetime + timedelta(hours=2)],
    "measurement_1": [12, 10, 16],
    "measurement_2": [15, 16, 12],


from pbu import TimeSeries
from datetime import datetime, timedelta

# initial time series base data (you can add measurements as well or provide as list of dictionaries
dict_of_list = {
    "date_time": TimeSeries.create_date_range(, + timedelta(days=1), timedelta(hours=3)),

# init time series
ts = TimeSeries(input_data=dict_of_list, date_time_key="date_time")
# add values (ensure same length as date_time series)
ts.add_values("measurement_1", [12, 10, 16, 10, 5, 8, 12, 9])  

# you can translate into a list of dictionary items (keys are maintained)
list_of_dict = ts.translate_to_list_of_dicts()

# extract data series from the time series
measurement_1 = ts.get_values("measurement_1")

# create new series that provides same value for all timestamps
ts.fill_values("constant_series", 5)

# remove a series from the total data structure

# re-sample data to 5 minute resolution, interpolating values, also pre-pending another day in front of the time series 
ts.align_to_resolution(resolution=timedelta(minutes=5), - timedelta(days=1))
# this will result in "interpolated" values for the first day, using the first value (12) to fill missing values
print(len(ts.translate_to_list_of_dicts()))  # 12 an hour, 2 days, 48 * 12 = ~576 items

# the same can also be achieved by:
# no need to provide resolution now
ts.align_to_resolution( - timedelta(days=1))


Database store with helper functions for accessing MongoDB. Each store instance represents a single collection. This comes with an AbstractMongoDocument class, which can be used to model the document types you store within a MongoDB collection.


from pbu import AbstractMongoStore, AbstractMongoDocument

# this is the object type stored in the mongo store
class MyObjectType(AbstractMongoDocument):
    def __init__(self, val1, val2):
        # optional: provide id and data model version 
        self.attribute = val1
        self.attribute2 = val2,
    def to_json(self):
        # init with version and id
        result = super().to_json()
        # add attributes to dictionary and return
        result["attribute"] = self.attribute
        result["attribute2"] = self.attribute2
        return result
    def from_json(json):
        result = MyObjectType(json["attribute1"], json["attribute2"])
        # get _id and version attributes
        return result

class MyObjectStore(AbstractMongoStore):
    def __init__(self, mongo_url, db_name, collection_name, data_model_version):
        # provide object type class as de-serialisation class (providing from_json and to_json)
        super.__init__(mongo_url, db_name, collection_name, MyObjectType, data_model_version)

# create instance of store
store = MyObjectStore("mongodb://localhost:27017", "mydb", "colName", 5)

# create document using a dictionary
    "version": 5,
    "attribute1": "a",
    "attribute2": 16,

# or use the type
doc = MyObjectType("a", 16)
doc.version = 5
doc_id = store.create(doc)

# update single document using helper functions
             AbstractMongoStore.set_update(["attribute1", "attribute2"], ["b", 12]))

# returns a list of MyObjectType objects matching the version
list_of_results = store.query({ "version": 5 })


An abstract class providing base-functionality for running monitors - threads that run a specific routine in a regular interval. This can be an executor waiting for new tasks to be processed (and checking every 5 seconds) or a thread that monitors some readout in a regular interval. The monitor is wrapped to re-start itself, in case of errors.


from pbu import BasicMonitor

class MyOwnMonitor(BasicMonitor):
    def __init__(self, data):
        super().__init__(monitor_id="my_id", wait_time=5)  # waits 5 seconds between each execution loop = data
    def running(self):
            # your code goes here (example):
            # result = fetch_data(
            # store_result(result)

If you want to run in a regular interval, the running method needs to be slightly modified:

from time import time
from pbu import BasicMonitor

class MyRegularOwnMonitor(BasicMonitor):
    def __init__(self, data):
        super().__init__(monitor_id="another_id", wait_time=60, run_interval=True)  # execute every 60 seconds = data
    def running(self):
            start_ts = time()  # capture start of loop
            # your code goes here (example):
            # result = do_something(
            # store_result(result)
            self.wait(exec_duration=round(time() - start_ts))  # include the execution duration

You can also pass a custom logger as custom_logger argument to the constructor. By default it will use the pbu.Logger and log major events such as start/stop/restart and errors.

Manage and run monitor

import threading

def start_monitor_thread(monitor):
    Thread function to be run by the new thread.
    :param monitor: BasicMonitor - an instance of sub-class of BasicMonitor 
    # start the monitor

# create monitor instance of your own class that implements BasicMonitor
regular_monitor = MyRegularOwnMonitor(data={"some": "data"})

# create thread with start-up function and start it
t = threading.Thread(target=start_monitor_thread, args=(regular_monitor, ), daemon=True)

# in a separate piece of code (e.g. REST handler or timer) you can stop the monitor instance

Stopping a monitor doesn't interrupt the current thread. If the monitor is for example in a wait period and you send the stop signal, the thread will still run until the wait period passes.

In an API scenario, I recommend using a dict or list to cache monitors and retrieve them via the API using the to_json() method for identification. This then allows you to signal starting / stopping of monitors by providing the monitor ID and lookup the monitor instance in the monitor cache.

BasicMonitor Methods

  • start() - starts the monitor
  • stop() - stops the monitor
  • to_json() - returns a dictionary with basic monitor technical information (id, state, wait behaviour, etc)
  • wait_till_midnight() - waits till the next midnight in your machines time zone
  • wait(exec_duration=0) - waits for the time specified in the constructor and in case of run_interval=True for the optional exec_duration, if provided.


Managing constants is good practice for avoiding typos. Imagine the following class:

class Tags:
    GEO = "GEO"

This allows you to just do: Tags.GEO allowing you to use your IDEs auto-complete, avoiding typos. But if you want to programmatically get all possible values for Tags, you can use pbu's ConstantListing class:

from pbu import ConstantListing

class Tags(ConstantListing):
    GEO = "GEO"

list_of_values = Tags().get_all()  # will return ['GEO', 'EQUIPMENT']



from pbu import list_to_json

# assuming we have `my_store` as an instance of MongoDB store or MySQL store, you can:
list_of_dictionaries = list_to_json(item_list=my_store.get_all())  # output is a list of dictionaries

This function operates on lists of objects inheriting from AbstractMongoDocument or AbstractMysqlDocument and converts them into dictionaries using the to_json() method of any object passed into the function. Objects passed into the function require the to_json() method and need to return the dictionary representation of the object. This function is just a mapping shortcut.


from pbu import default_options

    "a": 1,
    "b": 2,
    "c": 3,

result = default_options(default=DEFAULTS, override={"b": 4, "d": 5})
# result is: {"a": 1, "b": 4, "c": 3, "d": 5}

If you want to avoid additional keys other than the keys in DEFAULTS, you can provide a third argument:

from pbu import default_options

    "a": 1,
    "b": 2,

result = default_options(default=DEFAULTS, override={"b": 4, "d": 5}, allow_unknown_keys=False)
# result is: {"a": 1, "b": 4}


from pbu import default_value

result = default_value(value=None, fallback=5)  # None is by default disallowed
# result is 5

result = default_value(value=0, fallback=5, disallowed=[None, 0])  # either 0 or None would return the fallback
# result is 5

result = default_value(0, 5)  # value will be used, as it doesn't match None
# result is 0