nope, it really isn't, all we are asking for is "these are the words, find them in titles".
lower case everything and remove shit like emojis and not in the subset that search uses anything in quotes add a where clause split by spaces for keywords remaining remove words that aren't used look up each word and normalize groups of words (plural to singular, "c.i.a." to "cia", "shirtless goth" to "styxhexen", etc.) sort keywords based on popularity for optimization decrease to subset of hit with cached results (if "hitler" is a keyword then have a cache of hitler results ready to go) and so on.
These aren't all in order, like normalization might happen 2x (a phrase one earlier). This is also not something I've done (my sets are small enough I can just query each term) and i can use exact matches since it is for me. But I could easily pump this out in a few days.
It is similar to voat and how shitty search was that someone rolled out searchvoat. searchvoat worked correctly, and was just a guy who was sick of not being able to search.
GoldenAgeWhen 0 points 2.7 years ago
nope, it really isn't, all we are asking for is "these are the words, find them in titles".
lower case everything and remove shit like emojis and not in the subset that search uses
anything in quotes add a where clause
split by spaces for keywords remaining
remove words that aren't used
look up each word and normalize groups of words (plural to singular, "c.i.a." to "cia", "shirtless goth" to "styxhexen", etc.)
sort keywords based on popularity for optimization
decrease to subset of hit with cached results (if "hitler" is a keyword then have a cache of hitler results ready to go)
and so on.
These aren't all in order, like normalization might happen 2x (a phrase one earlier). This is also not something I've done (my sets are small enough I can just query each term) and i can use exact matches since it is for me. But I could easily pump this out in a few days.
It is similar to voat and how shitty search was that someone rolled out searchvoat. searchvoat worked correctly, and was just a guy who was sick of not being able to search.