[bugfix]fix the cut-off issue due to LLM predict token limit(256 for openai python lib default), by setting temperature to 0 and set LLM predict method from compact-refine to refine

reformat well

reformat

first edition

add dataset and change langchain directory name to langchain_manager, incase of a name consuming in python

app.py fixable to match hugging-face's app portal

pylint and py format ready

add tools

refactor directory construct