Exposing relations in Wordnet class

[dchaplinsky]

This might be a bit an overkill, but it has two advantages.

First is:

Another is that you can rewrite code like this:

def path_to_top(synset):
    spo = []
    for rel in [11, 107, 171, 172, 199, 212, 213]:

with meaningful names, not numbers

[maxadamski]

Autocompletion is definitely nice, but I think a simpler approach would be better.

Since autocompletion in Jupyter works for dict keys, this is enough:

Advantages:

• deaccent and slugify are not needed
• way less fields in Wordnet objects, so one can get a quick gist of what's inside with dir(wn) without looking at source code

Drawbacks:

• a bit more verbose (can be remedied somewhat by making an alias wn.rel = wn.relation_by_name)

Do you have other suggestions? If not, I will add my approach soon.

[dchaplinsky]

My idea was to keep it illustrative yet laconic. On one hand I like the idea of having the dict of all relations rather than a bunch of fields. On the other hand, consider this code: for rel in [ wn.REL_HIPERONIMIA, wn.REL_EGZEMPLARZ, wn.REL_HYPERNYM, wn.REL_INSTANCE_HYPERNYM, wn.REL_HIPERONIMIA_MIEDZYJEZYKOWA, wn.REL_HIPER_PLWN_PWN, wn.REL_HIPER_PWN_PLWN, ]: print(rel.id) vs dict version of it for rel in [ wn.relation_by_name["hiperonimia"], ... ]: print(rel.id) But it's clearly up to you. I can do a quick workaround in my code to use it as I like, if you want to stick to the dict approach

…

On Mon, Aug 23, 2021 at 10:56 PM Max Adamski ***@***.***> wrote: Autocompletion is definitely nice, but I think a simpler approach would be better. Since autocompletion in Jupyter works for dict keys, this is enough: [image: image] <https://user-images.githubusercontent.com/14139196/130504211-3fdc2bb5-5534-42e8-8243-1b54e7765b38.png> Advantages: - deaccent and slugify are not needed - way less fields in Wordnet objects, so one can get a quick gist of what's inside with dir(wn) without looking at source code Drawbacks: - a bit more verbose (can be remedied somewhat by making an alias wn.rel = wn.relation_by_name) Do you have other suggestions? If not, I will add my approach soon. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#6 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AABAA4XKDRGLID3O6A3OM6TT6KRWDANCNFSM5CSTICRQ> .

[maxadamski]

Sorry for being so pedantic in this pull request, but I like the simple dict approach too much.

With an alias, # 1 is the same length as # 2, although I admit that it's more visually cluttered. Too bad we can't index Python dicts with a dot :)

REL = wn.relation_by_name
[REL['hiperonimia'], REL['egzemplarz'], ...] # 1
[wn.REL_HIPERONIMIA, wn.REL_EGZEMPLARZ, ...] # 2