Make max_analyzed_offset configurable and apply it on search

[andreasferber]

Elasticsearch has a per-index setting index.highlight.max_analyzed_offset that governs how much of the indexed data is analyzed for search result snippet highlighting. Searches have a corresponding max_analyzed_offset option. Without that search option if a search finds results in indexed media files that are larger than the default setting (about 1MB) you get an error.

This PR makes the index setting configurable and ensures that a matching option is set when querying the index to avoid running into abovementioned error condition.

[splitbrain]

What's the exact error you get? I would assume that elastic simply does not return a search snippet, so the error would be some undefined field warning in the plugin itself?

[andreasferber]

I don't remember the exact error message, but it was a genuine Elasticsearch error, not something on the PHP/Dokuwiki end.

See also the Elasticsearch documentation:

max_analyzed_offset
By default, the maximum number of characters analyzed for a highlight request is bounded by the value defined in the index.highlight.max_analyzed_offset setting, and when the number of characters exceeds this limit an error is returned. [...]

(from https://www.elastic.co/guide/en/elasticsearch/reference/current/highlighting.html#max-analyzed-offset)

[splitbrain]

If I understand this correctly, there are two values. A maximum set at index level and a maximum set at query level.

If no query level max is set, and the index can't find a highlight before the index.max an error is thrown.

If a query.max is set, and the index can't find a highlight before query.max, no highlight is returned and no error is thrown.

query.max should be smaller than index.max, but your change suggests that even a query.max > index.max will suppress the error.

I have not been able to reproduce this issue, yet because I haven't found the right combination of document and query.

Since the plugin creates the index, I think it should set a sensible index.max and query.max. I don't think an option would be needed.