Calvin and Hobbes Search Engine | Hacker News Reader

Calvin and Hobbes Search Engine

michaelyingling.com

96 pointsArubis29 comments

Calvin & Hobbes & Dune (Calvin & Muad'Dib):

https://calvinanddune.tumblr.com/

This is incredible; thank you!

This page is great and I've used it for years. One thing to note is the underlying data (from gocomics) is not always accurate, and sometimes you will end up on a comic that does not match the search. Sometimes you can find the one you're looking for by browsing about a week out in either direction. If you email Yingling about it with the correct date or URL, he'll update the database.

> Your search returned no results.Remember that it searches for the EXACT phrase you typed! (And dates must be of the format MM/DD/YYYY)

Requiring an exact search is fairly prohibitive. For example I searched for "feeling frustrated" and got the above error, despite 100% certainty there are strips about feeling frustrated. You (or someone) could update this to search by semantics and not by keyword, using a combination of SBERT (or similar model) + a vector database: https://www.pinecone.io/docs/examples/semantic-text-search/

I think this is serious overkill, like you can get a way with simple stemming.

Or lemmatization and BM25. Often better than SBERT in my experience when using it out of the box.

I believe there is built in stemming in this search.

I searched the key word "commit" and some results were returned with the word "committed"

My guess is this guy just set up a simple DB query on the search form and left it at that.

That's what I've done in the past for simple text search.

I know that at least mongodb comes with built in full text search and stemming.

What he really should do is set up elastic search but that's a big pita to learn all that

Felt the same way. Some off the shelf algorithms would do wonders here. More sophisticated ML to tag (image and themes) would be awesomely overkill.

Or allow people to tag strips, etc.

shudders at the thought of what "people" would contribute

Calvin and Hobbes Search Engine - https://news.ycombinator.com/item?id=26119380 - Feb 2021 (163 comments)

Calvin & Hobbes Search Engine - https://news.ycombinator.com/item?id=1600211 - Aug 2010 (27 comments)

Calvin & Hobbes "Search Engine" - https://news.ycombinator.com/item?id=1600151 - Aug 2010 (5 comments)

I remember this from a while back - good to bring back memories. This may be a bit excessive, but using CLIP (https://towhee.io/image-text-embedding/clip) or some other multimodal learning model in conjunction with Milvus (https://milvus.io) could make for a search engine that takes the graphics into consideration as well. Not sure how well it would work for comics, but it could make for a pretty unique weekend hack.

I searched for "Somehow, it's always right now until it's later." and it found nothing, I checked on http://www.s-anand.net/comic.calvin.jsz and it's "Somehone, it's always right now until it's later." – it's not very forgiving.