In case folks are interested, my Ocellus tool (after a brief sabbatical) is back online. For example, see images tagged with the keyword bioacoustics.

Ocellus allows searching for and viewing more than 200K images that we have extracted from scientific articles as part of my TDM project. The images (and a whole lot of other data) are stored on Zenodo that is hosted at CERN in Geneva.

I wrote Ocellus to primarily view the images. It might turn into something more, but I am also determined to keep it very simple and fast. Ocellus is also very mobile friendly. I continue to work on it to make it even faster.

Ocellus itself is a 100% client-based app in that, there is absolutely no magic happening on the server. However, Ocellus leverages another tool I wrote called Zenodeo that does all the heavy lifting of querying data from various locations including Zenodo/CERN.

Zenodeo is a REST-compliant API written completely in JavaScript. It uses a file-based cache, and runs on a tiny VPS with 2GB ram and 50GB hard disk. (Ocellus also runs on the same server!)

You too can write your own application using the Zenodeo API. Check out the Zenodeo documentation

@G_N Nagarjuna G

Great to know that this is online and rocking!

When we don’t find something, is there any way we can update the database?


unfortunately no, as this is not a typical citizen-science db/app like, for example, iNaturarist. The route for the data here is

┌───────────���     ┌───────────┐    ┌───────────┐
│  article  ├─────▶extraction ├────▶  zenodo   │
└───────────┘     └───────────┘    └─────┬─────┘
                  ┌───────────┐    ┌─────▼─────┐
                  │  ocellus  ◀────┤  zenodeo  │
                  └─────┬─────┘    └───────────┘
                  │   users   │                 

At some point I do hope to add a way for the user to provide feedback on an image or its associated metadata.

@VirenVaz Viren

Sweet stuff punkish. The app is really quick.
I couldn’t follow the image to the paper though. Is that not a feature? I also feel that the title of the paper be linked to the image.


of course you can get to the paper from the image. That is the whole point of this project, to direct users to the original literature. I have tweaked the app to make the affordances clearer. See the image below

When you click on the rec ID, a panel slides up to show the image caption, and therein a link takes you to the original record (and the paper) on Zenodo.