twiddlepy
twiddlepy is a Python library designed for end-to-end extract, transform and load pipline (or ETL for short). Using a mapper file,
and optional functions your data can be transformed into a better suited format.
Features
- Extract, Transform and Load pipelines
- Multiple datasource options for extracting data
- Multiple repository options for loading data
- Support for mapping input data
Installation
Twiddlepy is available on the PyPi repository
pip install twiddlepy
Or if you want to install directly from the repository: python setup.py install, or drop the twiddlepy directory anywhere on your PYTHONPATH.
Connectors
There are a number of data repository connectors available with Twiddlepy. Currently implemented connectors include:
Data Source (Input)
- File Based
- CSV
- Excel Document
- Support for custom file loading (e.g. HTML)
- Database
- MySQL
- MSSQL
- Oracle
- SQLite
- MongoDB
Repository (Output)
- File Based
- CSV
- Apache Solr
Usage
Create a runnable python file with the following code:
from twiddlepy.config import config
from twiddlepy.driver import TwiddleDriver
driver = TwiddleDriver(config)
driver.process_data()Example Project Structure
.
|-- mapper
| |-- mapper.csv
|-- local_functions.py
|-- run.py (File that runs Twiddle)
|-- twiddle.cfg
User Configuration
Importing config from twiddle.config will import the default configuration items for each of the processes,
and will also look for a user defined configuration file on the path where the application is being run from.
All of the configuration items, including all of the default options can be found here
Mapper File
A mapper file defined by the user is used to defined the input data that will be extracted from the data repository. The mapper file is a CSV in which there are multiple columns that can be filled in to specify the data mappings. Thw following columns must be defined in the mapper:
| Column Name | Description | Options |
|---|---|---|
| dataset | The dataset twiddlepy will use mappings for | Any name (string) |
| source_field_name | A name of a source field | Any name (string) |
| source_field_type | The data type of the source field | One of: "str", "int", "float", "double", "timestamp" |
| allow_missing | Allow the column to be missing in the dataset | One of: "y", "n" (Yes or No) |
| min | Data Validation: minimum allowed value | Any numeric value |
| max | Data Validation: maximum allowed value | Any numeric value |
| allowed_values | Data Validation: list of allowed values | Any array of values |
| unit | The unit the column is represented by | Any name (string) e.g. kg |
| repository | The repository name the column belongs to | Any name (string) |
| repository_field_name | The name the column will be renamed to for data loading | Any name (string) |
| repository_field_type | The data type that will be applied to the column when loading | One of: "string", "integer", "float", "double", "date" |
| ignore | Mark column to be ignore by mapping process (for historic datasets) | One of: "y", "n" (Yes or No) |
Contribute
As a company, we welcome any input to fix/improve the project. Whilst we don't have a style guide currently, this is something we will be working on in the future to improve the project further. We're very interested to hear what you think about Twiddlepy, and any improvements you would like to see so please raise any issues in the tracker!
Contact
Got a problem/query and want to discuss it with us personally? Contact us at info@mediaintegration.co.uk. We also have a website with more information about the company here
