misleading results

[myrmoteras]

@punkish this result looks nice, but is only secondarily helpful.

https://ocellus.info/?page=1&size=30&resource=images&q=Noctuidae#

Most of the images are NOT> noctuid moth, but parasitic wasps, that might parasitize them. If I search images, then I want to have images about moth (Noctuids) not other stuff. This especially, if I select figures, rather than treatments.

I also do not understand why we do all the semantic annoations and attributes that allow specific searches and then use full text search. The full text results, for those that know more, mean anything about Noctuids, that is Noctuids proper but also anything that is mentioned in the respective figure, treatment.

I think we really need to be able to be able to provide in the search a result that is specific, and may be in the example have the option to do full text search.

[punkish]

if you want to search for noctuidae, then you can do a specific search https://ocellus.info/?family=Noctuidae&resource=images&page=1&size=30

specific searches are possible. Fulltext searches are also possible. Sometimes you may want one, sometimes the other. There is a choice. The termFrequency chart comes up only when a fulltext search is done. That is because the way the termFreq index is created

[myrmoteras]

Ok. But we need to discuss what the preset is. I think it should be a specific search. Right now, I need each time to explain, that first this is not a moth, but something related to it. From: punkish ***@***.***> Sent: Tuesday, July 11, 2023 6:04 PM To: punkish/ocellus ***@***.***> Cc: Donat Agosti ***@***.***>; Author ***@***.***> Subject: Re: [punkish/ocellus] misleading results (Issue #45) EXTERNAL SENDER if you want to search for noctuidae, then you can do a specific search https://ocellus.info/?family=Noctuidae&resource=images&page=1&size=30 specific searches are possible. Fulltext searches are also possible. Sometimes you may want one, sometimes the other. There is a choice - Reply to this email directly, view it on GitHub<#45 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABDFPJDW4TU3ZO2QQ73M6YLXPV2PNANCNFSM6AAAAAA2GGRM6U>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[punkish]

but is only secondarily helpful.

Good! It is meant to be an additional bit of info, not the main info. The main info is still the images

[punkish]

But we need to discuss what the preset is

There is no preset. You get back images for whatever you search.

q is an Alia's for full text of the treatment. Think of it as the easiest form of search. You specify a term, any term, and it will hopefully find something. It is great for many uses and certainly a good start.

Because of the special properties of the full text index, the searches are very fast and also can return term frequency.

You can also specify any key=val pair, and the search will be more specific.

[punkish]

@myrmoteras

I have an idea that I am going to try out. If it works, it should make you a very happy taxonomist. I am going to experiment with a broad-to-specific search.

I will start with a full text search, create a new table with the results, and then, withing that much smaller table, I will search the taxonomic ranks in reverse order, from species to kingdom, stopping when and if a search produces a result. This will require some programming, and a lot of testing. Let's see if we can get more satisfying results even with very simple (simple for the end-user) searches.