If a new issue is posted
We can ask AI models to extract a word list of the most important words.
We search the repository for all issues with the important words
We go from highest issue number and read the specification.
If confidence is high, stop the search and link back to it with a warning saying that it's likely to be a duplicate.
This approach might be a little more brittle but we won't have to worry about the database which should make implementation and maintenance much easier.
If a new issue is posted We can ask AI models to extract a word list of the most important words. We search the repository for all issues with the important words We go from highest issue number and read the specification. If confidence is high, stop the search and link back to it with a warning saying that it's likely to be a duplicate. This approach might be a little more brittle but we won't have to worry about the database which should make implementation and maintenance much easier.