6. Advanced Features¶
OnTask is built in Python using the web framework Django in combination with some additional libraries such as Django REST Framework, Pandas, etc. The application is available as open source with MIT License. This means that the advanced users can get a copy of the source code and modify it to suite their needs using the already existing models and functions.
6.1. The API (Application Programming Interface)¶
OnTask is a platform that facilitates the connection between data and the provision of personalised learner support actions. The higher the quality of the data the higher number of possible effective support actions. This means that OnTask should facilitate the connection with already existing data sources so that it can combine data sets and create a comprehensive view of how a learning experience is evolving.
The API is documented online through the URL suffix
apidoc. The page
contains the description of every entry point available with the required
When manipulating the elements in the table there are two versions of the basic operations (create a table, update a table, merge).
Pandas Version. This version handles the encoding of data frames using the pandas pickle encoding. This encoding has the advantage that maintains the elements of the dataframe intact. In other words, when the data frame is decoded from the pandas pickle format back to a regular dataframe, the same initial dataframe is obtained.
JSON Version. This version encodes the dataframes in JSON. The problem with this format is that values such as NaN and NaT (not a time) are not allowed by JSON, and they are substituted by empty strings. This change may have significant effects on the dataframe, specially on the types of the columns. If there is a dataframe with a column of type datetime and with any element with value NaT and it is first extracted through the JSON interface, and then uploaded again, the NaT value is transformed into an empty string, and Pandas will no longer recognise that column as datetime, but instead it will render the column of type string. This my have also an effect on how rules and actions are evaluated.
If OnTask is hosted in a server like Apache, you may need to check your
configuration so that it allows the processing of the
headers. For example, the module WSGI for Apache by default filters out this