notes for future reference

ACMAS/app/ACMAS_Web/future_reference.txt

@@ -0,0 +1,36 @@
+NOTES FOR FUTURE REFERENCE
+
+The current model for searching by keyword is to use yake to extract all keywords from the question. 
+After doing this, the code then stores all the keywords of the question in a database so that we dont need to 
+re-extract the keywords from files we already parsed since extracting keywords is a costly operation. 
+Although extraction is a costly operation, it is fast enough for the low amount of traffic ACMAS gets. 
+One edge case that is possible is when two different context questions have the same keywords. 
+Currently one solution would be to extract and match the keywords in the order they appear in the original 
+question.
+
+
+
+
+One solution to solve the issue of overloading the server with extraction is to have the client extract the 
+keywords in their questions instead. In this case, the clients would run the yake program in their browser and
+send their results to our servers. This would help offset the load to the clients.
+
+
+
+
+A solution to solve the edge case would be the keyword sorting solution
+
+Q1 Keywords: integrate lambda function C
+Q2 Keywords: integrate C lambda function 
+
+These two questions have the same keywords so they would both show up when being queried even though they are two unrelated questions.
+
+If our search query was "How do I integrate lambda in C" the keywords of this query would be
+Query keywords: integrate lambda C
+
+If we were to match the key words in in order with the questions we would get
+
+Q1: integrate lambda C
+Q2: integrate lambda
+
+C gets skipped over because in our search query, lambda is the second keyword.