Fleksja is a web application created with a ReactJS front end and a Django-Rest-Framework backend. It is designed to provide Polish language exercises generated programmatically.

1. Source of High-Quality Polish Text: Polish literature in the public domain was used as a source of high-quality Polish text.

2. Word Form Generation: Morfeusz2 was utilized to generate various word forms of any given word.

3. Conversion to Morphology Exercise: Polish passages were transformed into grammar exercises by automatically converting a percentage of words into dropdown multiple-choice exercises. This was accomplished by creating a ReactJS component called drop, which takes "word" and "choices" as props. Bootstrap was used to implement the dropdown functionality, and React's useState hook was employed to dynamically change the appearance of buttons based on user responses.

4. Script for Passage Transformation: A Python script was written to take a passage as input, split it into words, and create a serialized dictionary containing these words. A for loop was then used to iterate a number of times equal to 70% of the total number of words in the passage. In each iteration, a key value was generated by adding a randomly generated number between 2 and 7. Using Morfeusz2, word forms were generated, and if at least 2 word forms were successfully generated, the BeautifulSoup library was used to add the necessary tags for the drop component, along with the required props. Finally, the dictionary was converted back into a paragraph, and a POST request was made to the backend at the /passages endpoint to add the passages.

5. Backend Development: A Django model was created to represent passages, and Django Rest Framework's Model Serializer class, generic class-based views, and API views were used to quickly set up the API for CRUD operations. Additionally, a custom permission class was implemented to restrict editing and deletion of passage objects to their respective owners.

The frontend design of the application was also a significant focus. The useEffect hook was utilized to fetch data asynchronously from a JSON file when the component mounts. The fetched data was then shuffled and stored in the passages state. A loop was used to wrap each paragraph in a div element, and React-Jsx-Parser was employed to dynamically render the parsed passages. Careful attention was paid to ensure responsiveness across different screens.