The intent of this version of the Pointrel system was to have a quick-and-dirty experimental command line version of a triple store inspired by the MH (mail handling) concept, where you could use standard shell operations alongside command to create and find Pointrel triples. I made an initial version in Node.js. So, with the proper aliasing, you could issue shell commands like: "$ pointrel add a b c".
I rapidly realized it would be nice to also have a server version so you access the content via a web browser. So, I added that also using Node.js to the same JavaScript file (adding a "pointrel server" command).
Then I added a test file to add triples from the browser, which interacted with extra server functions to add a triple or to list things.
License: MIT license
For a shorthand command in bash to use "p" instead of "pointrel", add this to a "p" file in ~/bin and make it executable (adjust for where you copied Pointrel20150110 or where node is installed):
#!/bin/bash
/usr/bin/node ~/Pointrel20150110/pointrel.js "$@"
You can list the first item of all triples with:
$ pointrel list
You can list the second (b, or relationship) part of triples for a first (a) part with:
$ pointrel list $ITEM
You can list all the third (c, or value) part of triples for the a and b parts with:
$ pointrel list $ITEM $RELATIONSHIP
You can also "find" the last value for an item with (note the underscore "_", which is optional at the end, and can be used in any of the three fields):
$ pointrel find $ITEM $RELATIONSHIP _
You can find all the matches for something, which lists the file ID something is in with:
$ pointrel findall $ITEM $RELATIONSHIP _
$ pointrel findall _ $RELATIONSHIP $VALUE
You can delete previously added relationships if you know their timestamp ID (from the previous command, or from inspecting the files under the resources directory) with the delete command:
$ pointrel delete $SOME_TIMESTAMP_ID
To get the webserver running in the background to use the pages:
$ pointrel server &
To add web pages for display, create a triple like so (replace $PAGE_NAME with your page name):
$ pointrel add page:$PAGE_NAME content 'Your page content...'
You can also specify a content-type if it is not the default "text/plain" like so:
$ pointrel add page:$PAGE_NAME content-type 'text/html'
$ pointrel add page:$PAGE_NAME content-type 'application/javascript'
As an example, to add the content for a bootstrap editor running as "http://localhost:8000/editor.html" using bash in the project directory:
$ pointrel add page:editor.html content "$(< demos/editor.html)"
$ pointrel add page:editor.html content-type "text/html"
$ pointrel add page:editor.js content "$(< demos/editor.js)"
$ pointrel add page:editor.js content-type "application/javascript"
Setting the content-type for those two demos files is actually optional, as URLs ending in .html and .js will have their content-type guessed correctly if not set. The default content-type otherwise if it is not set is "text/plain".
That bootstrap file supports defining new triples, so you can define new pages or even upgrade the bootstrap page itself. You can add web content files with any extension or none at all, but you need to set the content type appropriately for non-text files. You can specify arbitrary paths with slashes in them. You can add more JavaScript files to make complex apps (but you should set the contentType for them). You can store data back to the server as a new triple using a POST command from JavaScript (see add_test.html for an example). There may be a short delay (currently up to one second) before the webserver picks up changes store via the command line.
There is also a more complex IBIS demo. You can load the files by running "upload-ibis.sh" after cd-ing to the demos directory. You can then run that demo by opening the URL of "http://localhost:8000/ibis/ibis.html".
Screenshots from updating the README.txt file (before it became README.md to include the screenshots and before further improvements):
