MongoTs

A fast python API for storing and querying time series in MongoDb.

• Free software: MIT license
• Documentation: https://mongots.readthedocs.io.

Requirements

• python >= 3.3
• MongoDb >= 3.4

Install

Install package:

pip install mongots

Manual install:

1. Clone this repo: git clone [email protected]:AntoineToubhans/MongoTs.git
2. install dependencies: pip install -r requirements.txt

Usage

You can instanciate client, database, collection just like you would do with pymongo:

import mongots
client = mongots.MongoTSClient()
db = client.MyDatabase
collection = db.temperatures

Insert

collection.insert_one(value, datetime, tags=tags)

Arguments:

• value (float): the value to be inserted
• timestamp (ddatemie.datetime): the timestamp for the value
• tags (dict, default=None): tags for the value; can be use the search/filter the value later on.

Return (bool): True if the insertion succeeded, False otherwise.

Query

collection.query(start, end, tags=tags, aggregateby=aggregateby, groupby=groupby)

Arguments:

• start (datetime): filters values after the start datetime
• end (datetime): filters values before the end datetime
• aggregateby (str): aggregates value per interval:
  • years: '1y', '2y', ...
  • month '1M', '2M', ...
  • days: '1d', '2d', ...
  • hours: '1h', '2h', ...
• tags (dict, default=None): similar to a mongo filter
• groupby (array): return statistics grouped by a list of tags (string)

Return (pandas.DataFrame): dataframe containing the statistics and indexed by datetimes.

Get tags

collection.get_tags()

Return (dict): The tags contained in the collection.

Get timerange

collection.get_timerange():

Return (None | tuple):

None if the collection is empty. Otherwise, a tuple containing two datetime object representing the minimum and maximum timestamps in the collection.

Run tests

Integration test requires a MongoDb to be up (run docker-compose up).

Launch all tests:

pytest

Launch only unit test:

python -m unittest -v test