Framework for automatic composition of videos.

In this Python code, we are taking Input Script ( Love Letters) and processing this file into meaningful sentences/chunks so that a computational scene can be constructed and a well-formatted meta- JSON file is generated. The JSON file acts as a structure for building the sequence of video composition and at the same time it conducts a Timeline analysis.

The python app is made in the Flask library and is used for evaluating the quality of SQL queries generated. These SQL queries are generated at runtime and gets their input parameter/ search arguments from the LLM. Basically, the LLM gives the extracted keywords from the input Love Letters script. The DAR system helps us evaluate and compute its performance parameters automatically as we go through the evaluation process.