Question for multiline examples and examples with double quotes

[combokang]

Hi, I used make_ABSA_dataset() to auto-annonate my review dataset and there are some issues:

1. For multi-line example quoted with "", the model does not consider it as one example.

  " How this $T$ restaurant gets 5 stars is beyond me . Or . . rather , it ’ s paid astroturfing ,
this
Negative
The portions $T$ are small , and the prices leave much to be desired . I was extremely excited to see more Asian options in midtown , but fam , this ain ’ t it . "
portions
Negative

　　This is the major issue I got, any suggesting workarounds besides remove line breaks?

2. For example begin but not end with a ", 2 spaces were added the the front. Like the first line of issue 1.

3. For some long example quoted with "", 2 spaces were added the the front as well.

  " Good Thai $T$ food , fair New York City prices . The outdoor dining ambiance is beautiful , with greenery everywhere . The drinks we got were beautiful as well ! Service was fast although our waitress seemed to be rushing us a bit - it ’ s a popular place so I think they are trying to move tables quickly . Overall , the food was good and the experience was positive . "
Thai
Positive

　　Not sure if the 2 spaces will affect training, but I will probably remove them.

Thanks for the help! 謝謝

[combokang]

And also, make_ABSA_dataset() didn't remove the aspect from an example, not sure if it's a correct format.

[combokang]

And also, make_ABSA_dataset() didn't remove the aspect from an example, not sure if it's a correct format.

Find same issue here, guess I'm not using the latest version.
#47 (comment)