A faster way to look up for a knowledge in a user's input?

[nuttyartist]

Rather than compiling for all possible element + key variations and then trying to match it to the user input, doesn't it make more sense to cut the user input and look up at a hash table for the right knowledge data? It should be much faster.

Current way:

Currently, we're going through all possible keys of a knowledge ([websites], for example) and pushing their respective element_plus_key to words_with_knowledge.

command: open [webssites] please
input: open new york times please

[websites]:

reddit: https:://
twitter: https:://
facebook: https:://
new york times: https:://

words_with_knowledge:

open reddit please
open twitter please
open facebook please
open new york times please

This method then compares every item in words_with_knowledge with the current [NAME] element node. This is not efficient.

The suggested and more efficient method is:

• if we find brackets ([])
• Get the respective data from the user input based on the current [NAME] element node by:
• Matching the string before the start bracket '[' and after the end bracket ']' and taking the string the middle of them. In our example, we need to remove from the user input "open " from the start and " please" from the end of the string.
• Then using a hash table we check if the string in the middle exists as a key in the knowledge file. In our example, we check if "new york times" exist in the knowledge file.
• If it does, we push it to the bestMatch ElementPlusKnoeledgeData and continue regularly.

[nuttyartist]

We need to check carefully if it works well.