Is there a way to save and reload the model?

[lingvisa]

@znwang25 Nice to see this package. I haven't found another fuzzy matching library that works well for Chinese. Can the model generated by fit() be saved and reloaded? Also ,do you have evaluation of the overall quality of this library? Thanks for the info.

[znwang25]

Hi @lingvisa ,
Thanks for using this package.

Can the model generated by fit() be saved and reloaded?

Currently, there is no direct interface for saving and loading a model. But it's a good point, I may add it in the next update.
In the meantime, I think you can just use pickle to save the FCM object.

# fcm = FuzzyChineseMatch(ngram_range=(3, 3), analyzer='stroke')
# fcm.fit(test_dict)

# Saving the model
with open('model.pkl', 'wb') as outp:
    pickle.dump(fcm, outp)

# Loading the saved model
with open('model.pkl', 'rb') as inp:
    fcm1 = pickle.load(inp)

Do you have evaluation of the overall quality of this library.

No. But I plan to add this in the future.

[sujianwei1]

Character '1' is not Chinese.
2024-10-11 20:23:17,008 - WARNING - Character '1' is not Chinese.
Character '/' is not Chinese.
2024-10-11 20:23:17,008 - WARNING - Character '/' is not Chinese.
Character '2' is not Chinese.
2024-10-11 20:23:17,008 - WARNING - Character '2' is not Chinese.
Character '【' is not Chinese.
2024-10-11 20:23:17,008 - WARNING - Character '【' is not Chinese.
Character '、' is not Chinese.
2024-10-11 20:23:17,011 - WARNING - Character '、' is not Chinese.
Character '、' is not Chinese.
2024-10-11 20:23:17,011 - WARNING - Character '、' is not Chinese.
Character '（' is not Chinese.
2024-10-11 20:23:17,011 - WARNING - Character '（' is not Chinese.
Character '）' is not Chinese.
2024-10-11 20:23:17,011 - WARNING - Character '）' is not Chinese.
Character '、' is not Chinese.
2024-10-11 20:23:17,011 - WARNING - Character '、' is not Chinese.
Character '/' is not Chinese.
2024-10-11 20:23:17,012 - WARNING - Character '/' is not Chinese.
Character ' ' is not Chinese.
2024-10-11 20:23:17,012 - WARNING - Character ' ' is not Chinese.
Character '】' is not Chinese.
2024-10-11 20:23:17,012 - WARNING - Character '】' is not Chinese.
Character ' ' is not Chinese.
2024-10-11 20:23:17,012 - WARNING - Character ' ' is not Chinese.
Character '+' is not Chinese.
2024-10-11 20:23:17,012 - WARNING - Character '+' is not Chinese.