changes mainly to the evaluation script and the chat.py files, increased score from 0.66 to 0.89. Also added some evaluation methods for hallucinations and stuff but could only run it once due to lack of credits. Recorded the Response Quality which was 0.882. Disclaimer: longer periods of wait for answer

fix huggingface

fix bug

Load facilities from HF Dataset when FACILITIES_DATASET is set

add google maps API