Search index and topic improvements

[sammeltassen]

Search index

• Metadata values are now added to the plaintext field to be indexed by Typesense.

This currently may include html tags if present in the source manifest.

Topics

• Added function to capitalize the first character of topics

This can be disabled as follows:

    config:
      extract-topics:
        upperCase: false

• Added an option to create date ranges

This can be enabled as follows:

    config:
      extract-topics:
        dateRange: ["date"]
        topicTypes:
          date: ['Date', 'Datering']

This will process metadata values such as 1900 - 1925 and create collections within the date topic for all years in between: 1900, 1901, ... 1924, 1925.

Todo: add an integer field to the Typesense index with the year values.

[stephenwf]

We can look at this again in the future, it was a way to have scripts that extract plaintext (into the keywords.txt) for indexing - but it made the output 2-3x larger!

[stephenwf]

Really nice addition with the dates too. I'll try and get this running locally to test soon 👍