wiki/2020-04-15-20-41-51.org

#+title: Add a desktop search engine for your digital library
#+author: "Gabriel Arazas"
#+email: "foo.dogsquared@gmail.com"
#+date: "2020-04-15 20:41:51 +08:00"
#+date_modified: "2020-09-09 05:09:41 +08:00"
#+language: en
#+options: toc:t
#+property: header-args  :exports both


Desktop search engines adds a crucial thing for [[file:2020-04-14-18-28-55.org][Maintaining your own digital library]]: retrievability.
With the ability to search and retrieve for information scattered throughout our files, we can then quickly connect and add information without much burden.

An ideal desktop search engine should have the following things:

- Features both a graphics user interface (GUI) and a command-line interface (CLI).
- Accepts not only plain-text files but other file formats such as images, videos, and audio.
- Quickly searches through text and even [[file:2020-04-13-17-32-27.org][File metadata]] for binary files.
- Invites integration (and extension) with an API allowing for more control and features.
- Highlights searching with a query language that is both simple and powerful.

With a cursory search, I found [[https://www.lesbonscomptes.com/recoll/][Recoll]] that fits the bill with the added bonus of being a cross-platform tool and easily configurable.

A desktop search engine like Recoll and DocFetcher uses index-based searches which enables them to search quite fast.
In order to achieve Google Search-level of search speed, you have to build an index based from the contents of the directory.
Most desktop search engines index the entire disk by default (and it is not recommended) but they should let you configure what directories to include (both Recoll and DocFetcher do).
Indices should also be continually updated as your filesystem changes.
Update the files for convention Apparently, the convention (at least starting from 2018) is to make the keywords and block names to be in lowercase as stated from one of the following discussions at https://orgmode.org/list/87tuuw3n15.fsf@nicolasgoaziou.fr/. The files was updated with a one liner of shell. However, this is Emacs and org-mode does have an API to let you do stuff in your config and interact with the documents internally so it is not an elegant solution in any way. 2021-04-01 16:08:15 +00:00			`#+title: Add a desktop search engine for your digital library`
			`#+author: "Gabriel Arazas"`
			`#+email: "foo.dogsquared@gmail.com"`
			`#+date: "2020-04-15 20:41:51 +08:00"`
			`#+date_modified: "2020-09-09 05:09:41 +08:00"`
			`#+language: en`
			`#+options: toc:t`
			`#+property: header-args :exports both`
Init commit of my wiki 2020-06-23 18:29:05 +00:00

Reformat the wiki Create a unified format for my wiki. Unfortunately, since most of my notes are created on a whim and I don't care much about metadata at the time, the timestamps are only approximations from my memory. 2020-09-08 21:28:33 +00:00			`Desktop search engines adds a crucial thing for [[file:2020-04-14-18-28-55.org][Maintaining your own digital library]]: retrievability.`
Update wiki as of 2020-07-06 2020-07-06 15:32:10 +00:00			`With the ability to search and retrieve for information scattered throughout our files, we can then quickly connect and add information without much burden.`
Init commit of my wiki 2020-06-23 18:29:05 +00:00
Update wiki as of 2020-07-06 2020-07-06 15:32:10 +00:00			`An ideal desktop search engine should have the following things:`
Init commit of my wiki 2020-06-23 18:29:05 +00:00
			`- Features both a graphics user interface (GUI) and a command-line interface (CLI).`
Update wiki as of 2020-07-06 2020-07-06 15:32:10 +00:00			`- Accepts not only plain-text files but other file formats such as images, videos, and audio.`
Reformat the wiki Create a unified format for my wiki. Unfortunately, since most of my notes are created on a whim and I don't care much about metadata at the time, the timestamps are only approximations from my memory. 2020-09-08 21:28:33 +00:00			`- Quickly searches through text and even [[file:2020-04-13-17-32-27.org][File metadata]] for binary files.`
Init commit of my wiki 2020-06-23 18:29:05 +00:00			`- Invites integration (and extension) with an API allowing for more control and features.`
			`- Highlights searching with a query language that is both simple and powerful.`

Update wiki as of 2020-07-06 2020-07-06 15:32:10 +00:00			`With a cursory search, I found [[https://www.lesbonscomptes.com/recoll/][Recoll]] that fits the bill with the added bonus of being a cross-platform tool and easily configurable.`
Init commit of my wiki 2020-06-23 18:29:05 +00:00
			`A desktop search engine like Recoll and DocFetcher uses index-based searches which enables them to search quite fast.`
			`In order to achieve Google Search-level of search speed, you have to build an index based from the contents of the directory.`
			`Most desktop search engines index the entire disk by default (and it is not recommended) but they should let you configure what directories to include (both Recoll and DocFetcher do).`
			`Indices should also be continually updated as your filesystem changes.`