Hi all, I have some problems with duckling extractor as below :
In this case, Duckling extractor is working 2 times and get the wrong resuilt. I got the wrong date (2021-07-15T19:02:00.000+07:00).
Any ideal for it?
Simply sending it to Duckling gets me this, I am assuming this is vietnamese. I have no idea what this means but google translate tells me you have written july 12th and since you don’t specify the year, duckling gives you 12th july for next 3 years.
the scond bit duckling tokenises is 07 2 which it is assuming as time for some reason, and returns you 07h2minutes of today’s thus the 15th July. I think the error is here.
third token it sees is 2m which is recognised as 2 million as a dimension number.
Normally i would check the corpus of vietnamese for duckling and how it is done for the time dimension. sometimes wrong corpus can lead to mistakes.
[{"body":"ngày 12/07","start":10,"value":{"values":[{"value":"2021-07-12T00:00:00.000-07:00","grain":"day","type":"value"},{"value":"2022-07-12T00:00:00.000-07:00","grain":"day","type":"value"},{"value":"2023-07-12T00:00:00.000-07:00","grain":"day","type":"value"}],"value":"2021-07-12T00:00:00.000-07:00","grain":"day","type":"value"},"end":20,"dim":"time","latent":false},{"body":"07 2","start":18,"value":{"values":[{"value":"2021-07-15T07:02:00.000-07:00","grain":"minute","type":"value"},{"value":"2021-07-15T19:02:00.000-07:00","grain":"minute","type":"value"},{"value":"2021-07-16T07:02:00.000-07:00","grain":"minute","type":"value"}],"value":"2021-07-15T07:02:00.000-07:00","grain":"minute","type":"value"},"end":22,"dim":"time","latent":false},{"body":"2m","start":21,"value":{"value":2,"type":"value","minute":2,"unit":"minute","normalized":{"value":120,"unit":"second"}},"end":23,"dim":"duration","latent":false},{"body":"2m","start":21,"value":{"value":2000000,"type":"value"},"end":23,"dim":"number","latent":false}]I think like you too. I found some regex method of time which are working at this case like dd/mm and hh mm. Thank you for let me sure about it.