Python: Use Whoosh to indexing wordnet and search fields

gistfile1.txt

# %%
import nltk
from nltk.corpus import wordnet2021 as wn

from whoosh.index import create_in, open_dir
from whoosh.fields import *

from tqdm.notebook import tqdm
from whoosh.qparser import QueryParser

from IPython.display import display

ix = None

# %% [markdown]
# ## Create Index

# %% [markdown]
# ### Prepare schema and index dir

# %%
schema = Schema(
    sskey=ID(stored=True),
    lemma_names=ID(stored=True),
    cmn_lemma_names=NGRAMWORDS(1, 4, stored=True),
    definitions=TEXT(stored=True),
    pos=ID(stored=True)
)

ix = create_in('wn_index', schema)

# %%
all_ss = list(wn.all_synsets())

# %% [markdown]
# ### Write index

# %%
writer = ix.writer()

for ss in tqdm(all_ss):
    sskey = ss.name()
    lemma_names = ss.lemma_names()
    cmn_lemma_names = ' '.join(ss.lemma_names('cmn'))
    defi = ss.definition()
    pos = ss.pos()
    writer.add_document(sskey=sskey, lemma_names=lemma_names, cmn_lemma_names=cmn_lemma_names, definitions=defi, pos=pos)
writer.commit()

# %% [markdown]
# ## Query

# %%
if ix is None:
    ix = open_dir('wn_index', readonly=True)
   
with ix.searcher() as searcher:
    query = QueryParser("cmn_lemma_names", ix.schema).parse("氯化")
    results = searcher.search(query)
    for r in results:
        display(r)

# %% [markdown]
# ## Close

# %%
if ix is not None:
    ix.close()
    ix = None