Many of the links in README.md are broken.

[russelljjarvis]

I have a paper that I want to run through the scibot scigraph and scicrunch pipelines but none of the links seem to work. Actually I have a journal club, and I want to make interesting annotations and graphs for every paper I need to read for the journal club.

Is there an online portal with this stuff working, or should I just compile everything from source and run it on my local machine?

I noticed that there is a vagrant image for scigraph, is there any particular reason why vagrant was used over docker (I only ask because I want to avoid the larger memory footprint)? Will this vagrant image solve dependency issues for the whole scicrunch stack?

[tgbugs]

I updated the README to be clear about pages that only work if you are logged in.

If you want to use SciGraph loaded with the NIF-Ontology we have a vocabulary service which is currently http://trinity.neuinfo.org:9000/scigraph/docs but that will be changing soon and I will leave an update here. EDIT: the updated docs for the scigraph endpoint (and the rest of the scicrunch api) live here (need to be logged in to see and access them).

You do not have to use vagrant and the vagrant image will not solve the dependency issues. I have a python program (which functions as a shell script) that can be used to build SciGraph from scratch, load the ontology you want, and produce a zip of the graph and the web services that you can deploy wherever you want. The only prerequsites are >=python3.5, java8 and maven installed (see the pyontutils README for details). I also have upstart, openrc, and systemd startup scripts that I can make available if needs be.

[russelljjarvis]

AOK. I installed the https://github.com/tgbugs/pyontutils program as per its REAMDE.md. I also created an account at hypothesis. My main interest is in machine reading journal articles in order infer patterns, and to enrich documents with helpful machine generated annotations in a way that is difficult for a regular human to achieve.

I believe some of this functionality may be embedded in various sci-crunch related programs, but I am having trouble figuring out how all of the parts work together.

[tgbugs]

SciBot is currently set up to mine RRIDs from the literature. If you want to use it to provide more annotations then this line of code is what you need to change, specifically you need to provide a function that conforms to the type defined here (for and example see find_rrids which is the current implementation).

There are a number of SciCrunch related things you could write a finder for. If you do you will also need to update the resolver function if you you want to provide structured metadata for annotations from a different source.

We don't have anything implemented right now except find_rrids but if you are interested in writing your own I'm more than happy to do a bit of refactoring so that we can support modularized finders, checkers, resolvers, and submitters.

A simple example might be to feed the scigraph lexical endpoint (ctrl-f for /scigraph/lexical/entities and you will need to be logged in to scicrunch to see that) to submit raw text and use the tag positions returned (e.g. from a query like curl -X GET "https://scicrunch.org/api/1/scigraph/lexical/entities?text=Is%20there%20a%20brain%20in%20here%20that%20has%20a%20hippocampus%20that%20you%20can%20find%3F" -H "accept: application/json", EDIT: this now works in our python scigraph client where you can pass in a key at class initialization.