Chapter 1, section 10-11: "ORTH" vs "TEXT" in pattern matching about spacy-course

[ines]

Thanks! 🙏

And sorry, that's definitely a mistake – I must have used an older version of the slides that didn't yet include this change. Already fixed this!

The change to TEXT in spaCy v2.1 was actually directly inspired by this course when I realised that ORTH was just really difficult to explain, because it kinda made no sense. The obvious question would always be, why is it not TEXT? The ORTH is a reference to "orthography" and the Token.orth (hash ID of the token) and Token.orth_ (string value of the token and the same as Token.text – only really exists for consistency and isn't normall used).

I'm not sure if it is meant to also explain the pattern matching keys.

spacy.explain currently only explains labels of the label schemes and not attributes etc.

Once when I had to install the en_core_web_sm myself

Which exercise are you referring to here? And you mean load the model with spacy.load, right?

[thorbjornwolf]

Thanks for the swift response!

spacy.explain currently only explains labels of the label schemes and not attributes etc.

All right, thanks!

Which exercise are you referring to here? And you mean load the model with spacy.load, right?

When I tried to load the model with spacy.load, I had to install it first. Sorry for the unclarity 😄 I have derived issue #5 to delve into this part.

I can validate that the ORTH->TEXT transformation is complete - at least in my browser - so I think we can close this issue 👌