Explore and document how to use coli-ana for retrieval

[nichtich]

Explore and document methods to use decomposed DDC numbers for indexing and/or query expansion in retrieval systems such as Solr/Elasticsearch/...

To give an example, a document with DDC number 700.90440747471 should be indexed with "Modern arts", "1940-1949", "Museums, collections, exhibits", and "New York Metropolitan Area" (plus synonyms for each of these classes) and (probably ranked lower) with all labels of classes in the hierarchy.

The use case could be split into two steps:

1. Analyze DDC numbers and split them into their main components
2. Expand index and/or query for each component based on class labels and hierarchy

Only the first is task of coli-ana but the use case should be documented as part of coli-ana still.

[nichtich]

Some notations with titles in K10plus (found by iterating related notations)

• 641.5 - cooking

• 641.509 - historic cooking

• 641.50901 - prehistoric cooking

• 641.50902 - medieval cooking

• 641.509024 - cooking like at Martin Luther's time

• 641.50904 - 20th century cooking (no titles)

• 641.509041 - cooking during WW1

• 641.509047 - cooking in the 1970s

• 37 education (no titles)

• 370 - education

• 370.09 - historic education

• 370.0902 - medieval education (no titles)

• 370.940902 - medieval education in Europa

It should be possible to jump from 641.509 - historic cooking to

• 641.5 cooking
• T1--09 history (see Support analysis of pure table numbers #61)
• 641.50902 and other sub(-sub) classes with titles
• 370.09 and other classes that use T1--09 (if there are titles in these classes)

This requires a database with the full DDC hierarchy, all decompositions and the number of titles for each class.