term frequency chart in ocellus

[punkish]

built on the new term frequency offering of the zenodeo api, Ocellus now offers a chart of how many times a term occurs per year in our treatment texts along with how many of those have images. Please see https://ocellus.info/?page=1&size=30&resource=images&q=linnaeus for an example

if there are issues with the chart, please open an issue here and let me know. Thanks

cc @myrmoteral @gsautter @flsimoes @tcatapano

[myrmoteras]

sorry, I did not get it (there is a typo in my alias).

This looks interesting.

What do the results now represent: treatments and figures about a serach taxon, or figures in which captions and treatments the search term occurs, irrespective of whether the figure is about the searched taxon or not?

[punkish]

in this specific case, you get the images from the treatments that have the term 'linnaeus' anywhere in the text

[myrmoteras]

If I search for Merula merula, then I want images about the black bird, not other things.

If I am a bit more familiar, I want to have others stuff that might be linked to it.

[myrmoteras]

in this specific case, you get the images from the treatments that have the term 'linnaeus' anywhere in the text

All the efforts we make in Plazi is to go from full text search to enable specific search. This is our KPI, what makes us outstanding.

[punkish]

I don't know what is a KPI, but nothing stops a user from doing specific searches. You can do full text searches and you can go specific searches

[myrmoteras]

Sorry puneet, the user wants first to get images about the taxon searched, not related taxa mentioned in the text somewhere.

When you open an encyclopedia your goal is not to find every instance with your search term, but the section about your search term.,

it just doesn't look right, and especially for a taxonomist.

[punkish]

So search for the taxon. As I said, any kind of search is possible.

The user has to somehow convey to the program what the user wants to search. The program can't guess that. If the user doesn't provide any "instructions" to the program, then the program carries out a full text search.

I am not sure what it is that you envision in a search program. Perhaps if you could write it down how you would like a search program to function then we could take a look at whether it is possible or not.

Of course, theoretically anything is possible given enough time and resources. There is always a trade off

[punkish]

Ocellus is not s specialized search tool. Well, it is specialized in that it searches and shows only images. But it makes trade offs for simplicity. Maybe some other tool could be developed that wouldn't make this trade off and would allow the taxonomist to specify all kinds of things.