Re: Remembrance Agents

[Jean Louis <bugs@gnu.support>]

* Eric S Fraga <e.fraga@ucl.ac.uk> [2020-11-30 13:11]:
> On Monday, 30 Nov 2020 at 11:37, Gerardo Moro wrote:
> > Ok, thanks. How do you specify the location and is the formatting ok when
> > showing you the relevance pieces of text from the local databases?
> 
> The agent needs to be told which files to index (check the man page for
> ra-index) and, in my case, I point it to all .org files and all text
> files under ~/Mail.  The formatting is okay as far as I am concerned but
> this is a subjective aspect that will differ for everybody!

SMART (String Matching Algorithm Research Tool)
https://github.com/smart-tool/smart

Does it use this algorithm? I am cloning that one to see how useful it
could be in relevance searches.

,----
| However, I can also see that the PostgreSQL database has already
| built-in features for relevance searches:
| 
| https://www.postgresql.org/docs/current/textsearch-intro.html
`----

This makes then many external software pieces redundant, it becomes
again trivial to work with a database as underlying algorithm and
functions have already been programmed. Making a new remembrance agent
based on various structured information is making few higher level
functions that integrate the underlying capability of PostgreSQL
database and Emacs users' editing environment:

0. Create appropriate table in the database.

1. Make automatic or semi-automatic list of Org files or other files.

2. Make simple Emacs Lisp function to index all those files in the
   database. This is probably few lines function.

3. Make few Emacs Lisp functions that observes what users is doing and
   display to user relevant results with the possibility to construct
   hyperlinks from the result buffer straight to Org buffer, or
   possibility to open up those relevant files for more references.

Jean