Is the lookup in mthesaur all words in the line or only first?

[barskern]

When the plugin searches for synonyms it returns the first line which contains the word that is queried (see snippet).

https://github.com/Ron89/thesaurus_query.vim/blob/23d8aa3f7d1a785e31e760e1b6b3b4c18abf8332/autoload/thesaurus_query/backends/mthesaur_lookup.py#L33

The way I understood the format of mthesaur is that the "root" word is the first word of the line, so shouldn't this line then be changed to:

if word == synonym_list[0]:

The reason I'm posing the question is that I seem to get a lot of "bad" synonyms when I query a word, normally because the word appears in a synonym list of another word where the link is debatable.

[Ron89]

Hi @barskern , mthesaur organizes groups of words with similar meanings. So the target word may not be the first word in the word list.

Then again, a single word may have many meanings. So technically a word may appear in multiple rows instead of one. Looking at the logic now, this function only retrieve the synonym of a word for its first appearing meaning group.

Since this thesaurus plugin already support multiple meaning groups, let me change its logic to include all its meaning groups.

[Ron89]

@barskern , I have made the change, though... the result is for common words with a lot of meaning groups, the result can be horrendously huge.

In the plugin, we have two parameters that may reduce the pain of overly abundant result, g:tq_truncation_on_definition_num and g:tq_truncation_on_syno_list_size.

https://github.com/Ron89/thesaurus_query.vim/blob/2bb83fa7c27038baee875d42af445c20fbe09b59/doc/thesaurus_query.txt#L252-L269

For example, if you use

let g:tq_truncation_on_definition_num = 10
let g:tq_truncation_on_syno_list_size = 20

the number of returned value will be much more manageable, though there is risk that some useful results are filtered out. Unfortunately, mthesaur does not rank results by relevance.

[barskern]

@Ron89 That was quick feedback, thank you so much. I will give it a go with the changes you implemented here, though the immense amount of suggestions that you outline seem to make this backend quite difficult to use and benefit from. I might simply continue to only use the datamuse_com backend, as it gives good answer, though few to choose from.

[Ron89]

Since there is no further update for a few months, I will close this ticket for now.