inference_engine.py

# this is the inference engine for knowledge bases

import json

class InferenceEngine:
    def __init__(self, knowledge_base):
        self.knowledge_base = knowledge_base
        with open(self.knowledge_base) as json_:
            self.kb = json.load(json_)
    def start_inference_session(self):
        # prepare inference output
        
        sub_topics = kb['1']['topics'].keys()
        opt = ""
        for i in sub_topics:
            opt += i+". "+kb['1']['topics'][i]['title']+"\n"
        r = kb['1']['title']+"\n"+opt
        return (r, "1", "CON")

    def is_inference_input_valid(self, previous_state, input):
        split_state = previous_state.split('.')
        obj = self.kb
        for i in split_state: 
            obj = obj[i]['topics']
            if obj[input]['node_type'] == 'leaf':
                return True
            elif obj[input]['node_type'] == 'decision':
               return True
        return False

    def load_node(self, previous_state, key):
        # state looks like 1.1 with sub chapters in the kb
        split_state = previous_state.split('.')
        obj = self.kb
        for i in split_state: 
            obj = obj[i]['topics']
            if obj[key]['node_type'] == 'leaf':
                # directly get advice
                return (obj[key]['content'], previous_state, "END")
            elif obj[key]['node_type'] == 'decision':
                sub_topics = obj[key]['topics'].keys()
                opt = ""
                for i in sub_topics:
                    opt += i+". "+obj[key]['topics'][i]['title']+"\n"
                r = obj[key]['title']+"\n"+opt
                new_state = previous_state+"."+key
                # update state
                return (r, new_state, "CON")
        return ('Error', 0, 'END')