Skip to content

Latest commit

 

History

History
203 lines (151 loc) · 5.77 KB

README.md

File metadata and controls

203 lines (151 loc) · 5.77 KB
Raw

Overview

This is a REPL environment for the Noctua annotation tool. This gets you some nice things:

  1. You don't have to wait until UI functionality makes it into the graph (or some other) editor--if the low-level API is capable of expressing it, you can easily do it in the REPL.
  2. For some operations, it is easier to express them succinctly with commands rather than work with a UI.
  3. An extremely easy way of prototyping new functionality.
  4. Spooky action at a distance.

Usage

Given that you have logged in to a Noctua instance, look at the URL and copy your token (the variable is "barista_token").

Know your target installation well enough to know where your Barista is hiding. Copy that URL down.

You might also need to know your Barista definition--is defaults to "minerva_local".

Should then be as simple as:

npm install
node ./bin/noctua-repl.js --token=xyz321abc987 --server http://127.0.0.1:3400

See the examples section for more information.

Examples

CLI examples

Connect to a local server with:

~/local/src/git/noctua-repl$:) reset && node ./bin/noctua-repl.js --token=01234 --server http://localhost:3400

Connection to the labs/development server with:

~/local/src/git/noctua-repl$:) reset && node ./bin/noctua-repl.js --token=01234 --server http://barista-dev.berkeleybop.org --definition minerva_public_dev

Connect to a local server and dump meta-information to a file using CLI:

~/local/src/git/noctua-repl$:) reset && node ./bin/noctua-repl.js --token=01234 --server http://localhost:3400 --definition minerva_local --command "get_meta(); show(response)" > /tmp/foo.txt

Connect to a local server local and dump meta-information to a file using a script:

~/local/src/git/noctua-repl$:) reset && node ./bin/noctua-repl.js --token=01234 --server http://localhost:3400 --definition minerva_local --file ./scripts/run-script-test.repl.js > /tmp/bar.txt

REPL examples

Get all of the meta-information for the current instance. It is suggested that you run this as your first command at the beginning of every session.

get_meta()

Assign environment to a known model, then add a couple of new individuals to it.

model_id = 'gomodel:taxon_9606-5539842e0000002'
add_individual('GO:0022008')
add_individual(intersection(['GO:0022008', 'GO:0008150']))

Add a new model, which gets the default assignment when done. Then add two new individuals as arguments to a new fact. Note that request-set operations can usually be chained together.

add_model()
var rs = new_request_set()
rs.add_fact([rs.add_individual('GO:0022008'), rs.add_individual('GO:0008150'), 'part_of'] )
request_with(rs)

Note that in this environment, while request-sets have chaining (see above), requests do not. For example, to recreate the get_model() macro:

r = new_request('model', 'get');
r.model('gomodel:5970219a00000333')
rs = new_request_set().add(r)
request_with(rs)
// Or as an insane one-liner:
request_with(new_request_set().add(r = new_request('model', 'get') && r.model('gomodel:5970219a00000333') && r))

An example of examining a request (use objectify), a request-set (use structure), and a response (just ).

r = new_request('model', 'get')
r.objectify()
r.model('gomodel:5970219a00000333')
r.objectify()
rs = new_request_set().add(r)
rs.structure()
request_with(rs)
response
response.data()
g = new_noctua_graph(response.data())
g.all_nodes()

Two ways of moving a node remotely. The first is manually, using barclient and telekinesis; gets very chatty.

get_model('gomodel:567b544200000029')
barclient.telekinesis('gomodel:567b544200000029/567b544200000126', 100, 100)

The second way uses a helper wrapper.

get_model('gomodel:5667fdd400000077')
// model.get_node('gomodel:5667fdd400000077/5667fdd400000347')
move_individual('gomodel:5667fdd400000077/5667fdd400000347', 100.0, 100.0)

Sending a general warning broadcast to all connected users. First, you must connect. While this can be done manually without a model, we'll grab a model first to make sure we have some connection. (The barista client is designed around a per-model attitude.)

get_model('gomodel:567b544200000029')
barclient.broadcast({"message_type": "warning","message":"The server will catch fire, please save!"})

Run SPARQL commands off of an endpoint of our choice.

sparql.endpoint('https://query.wikidata.org/sparql')
sparql_template('../bbop-manager-sparql/examples/template-01.yaml', {pmid: '999'})
show(sparql_response.raw())

API/Objects

This is a list of symbols that are defined or used in the REPL/scripting environment beyond what comes with node. Remember that you can do any of the normal things that you might want in this environment--it's just node!

Helpers.

  • 'bbop' the bbop-core package
  • 'us' underscore
  • 'manager' the bbop-rest-manager with bbop-reponse-minerva
  • 'show' function to display objects more intelligently

Auto-variables--these are attempted to be set after every call.

  • 'token'
  • 'model'
  • 'model_id'
  • 'request_set'
  • 'response'
  • 'query_url'

Class expressions--package class-expression.

  • 'union'
  • 'intersection'
  • 'svf'
  • 'cls'

Manager actions--these are manager functions mapped up to the top-level.

  • 'get_meta'
  • 'get_model'
  • 'add_model'
  • 'save_model'
  • 'add_individual'
  • 'new_request_set'
  • 'new_request'
  • 'request_with'

Bigger fun function macros.

  • 'show_models' show summary information for all current models; can take a single string argument of "id", "title", "deprecated", "modified-p", "contributor", "model-state", or "date" for sorting
  • 'show_response'
  • 'silent' supress/re-enable the display of action results

Limitations/TODOs

We're actually feeling pretty good about this right now. Let us know.