As has been mentioned to you in the comments above, you need to start off by breaking the big problem down into little problems. Effectively, you have two problems you need to solve here; the first is how to search the pdf content and the second is how to write the chatbot.
Let's start off with searching the pdf content. If I were tackling this, I would look to see if there was anything available that could actually read document content. Fortunately, there is; there are any number of services such as
Lucene[
^] or
Elasticsearch[
^] that provide full text-search capabilities. So, that's one part solved.
The next thing I would be looking at is what resources are available to me to help me learn how to write a Chatbot. Chatbots are not something you should want to tackle without research, or dedicated libraries if you don't want to write natural language systems from scratch yourself. Fortunately, in Python, there are toolkits such as the NLTK available which you should be able to leverage.