Announce: A prototype SQL <--> AtomSpace bridge

[Linas Vepstas]

I invite everyone to take a look at a new demo prototype for an SQL-to-AtomSpace bridge, here: https://github.com/opencog/atomspace-automap

The goal of the bridge is to automatically map existing SQL databases into the AtomSpace,  both for reading, and for writing. Also, instead of loading/importing the *whole* dataset, it only accesses the parts that are touched. So it's not a traditional bulk import.

The idea of "automatic" means that it can work with "any" SQL database, without requiring custom hand-written code or or hand-written mappings or templates. Of course, automatic mapping can result in AtomSpace structures that are a bit ugly, because SQL tables tend to be ugly. The mapping is relatively transparent; it doesn't hide any of the SQL dirt.

This is a demo. Version 0.2.1 -- So far, read-only, no write. It works. I tried it with the flybase genetic dataset. I even created a mini-browser, so that you can hop from gene to gene, row to row, table to table in the flybase dataset.  The mini-browser should work with any SQL database.  But it is .. .version 0.2.1

The work was inspired by Mike Duncan, who has been asking for this for years. Sorry it took so long.

-- Linas

--

Patrick: Are they laughing at us?

Sponge Bob: No, Patrick, they are laughing next to us.