Data-gems.txt

https://stereobooster.github.io/reference-data-as-dependency
http://lampwww.epfl.ch/papers/idealhashtrees.pdf

Task

1. compact file format for faster read from disk
  magpack? protobuf? pbf
2. traverse all file, leave only popular bits. Cache LRU
  mini mime
3. be able to unload db from memory, if not required
4. use effiecient in memory representation instead of objects
  zero copy
  http://bsonspec.org/spec.html DateTime
6. prebuild index
  pasword db
7. Read only, so no locks or transactions


+-------+
| file  | - effective format representation. Format with optional schema
+-------+
    |
+-------+
| index | - optional, depends on data type. For example for LIKE pattern% binary tree will do
+-------+
    |
+-------+
| cache | - optional, cap number of simultenious objects
+-------+
    |
+--------------+
| access layer | - effective in memory representation
+--------------+


## gems
- https://github.com/discourse/mini_mime
- https://github.com/stereobooster/micro_mime
- https://github.com/bdmac/strong_password/pull/14
- https://github.com/SixArm/sixarm_ruby_unaccent/issues/1

- https://github.com/nicotaing/yaml_record
- https://github.com/byroot/frozen_record
- https://github.com/zilkey/active_hash
- https://github.com/hexorx/countries

## npm
- https://github.com/zeke/all-the-cities/pull/3

## lisp
- https://picolisp.com/wiki/?home
- http://www.cliki.net/CommonLispPrevalence

## natural lang js

https://github.com/nhunzaker/speakeasy
https://github.com/NaturalNode/natural
https://github.com/nlp-compromise/compromise