0.1.5

What's Changed

New

• Additional function setup_gpu_modules to install Python libraries which are related to GPU libraries

• Added rag function for retrieval-augmented generation using various transformer models.

• Added sentence_similarity function for computing similarity scores between sentences using transformer models.

Modified

• Python library versions in setup_modules are frozen to prevent dependency/compatibility issues #17

• Switched from pytube to pytubefix Python library to avoid HTTP Error 400: Bad Request errors

---

Full Changelog: 0.1.4...0.1.5

transforEmotion 0.1.4

Version 0.1.4 introduces two new major functions: image_scores and video_scores. These functions perform Facial Expression Recognition (FER) and provide emotion scores based on the detected facial expressions in images or videos.

Added

• Added image_scores function for FER on images. (Issue #7)
• Added video_scores function, along with related functions to download videos from YouTube, cut the video from X to Y seconds, and save frames and the video itself in a specified directory. (#9, #8)

Modified

• Simplified the setup_miniconda function and added additional checks for environment setup. (#5, #6, #7, #8)
• Updated the package vignette to include examples of image_scores and video_scores.

transforEmotion 0.1.3

Added

• simulate_video: simulates emotions in a video using the DLO model implemented as a continuous time state space model. In other words, this function simulates data that we would expect to see when using emotion detection methods for the analysis of emotional dynamics in videos.

Modified

• The description of the package has been modified to reflect the new plan for future development, including support for emotion detection in images and videos.

Misc

• Unit testing has been implemented using the testthat framework.