Trichotillomania

[zebrassimo]

Hi Duygu, on page 116:

doc = nlp("I suffered from Trichotillomania when I was in college. The doctor prescribed me psychosomatic medicine.")
pattern = [{"LENGTH": {">=":10}}]
matcher.add("longWords", [pattern])
matches = matcher(doc)
for mid, start, end in matches:
    print(start, end, doc[start:end])

I'm getting

0 1 I
3 4 Trichotillomania
5 6 I
12 13 prescribed
14 15 psychosomatic

Having "prescribed" in there is fine (page 116 however says differently), yet finding "I" is sort of unexplainable to me. I'm on spaCy version 3.1.3. Is this a bug or a feature?

Greets, rudisoft

[DuyguA]

Hi Duygu, on page 116:

doc = nlp("I suffered from Trichotillomania when I was in college. The doctor prescribed me psychosomatic medicine.")
pattern = [{"LENGTH": {">=":10}}]
matcher.add("longWords", [pattern])
matches = matcher(doc)
for mid, start, end in matches:
    print(start, end, doc[start:end])

I'm getting

0 1 I
3 4 Trichotillomania
5 6 I
12 13 prescribed
14 15 psychosomatic

Having "prescribed" in there is fine (page 116 however says differently), yet finding "I" is sort of unexplainable to me. I'm on spaCy version 3.1.3. Is this a bug or a feature?

Greets, rudisoft

Hellos back!

So, i tried the code quickly. I created a clean Matcher object from scratch. I attached a screenshot, it looked normal to me:

If you're trying the chapter code sequentially, it might be the case that some previous patterns remained within your Matcher. For instance, there's an onlyShort pattern at page 112, which matches to 1 length tokens. Would it be the case?

[zebrassimo]

Thanks Duygu, I am running each chapter in a different Jupyter notebook and you rightfully pointed to the correct issue.

[DuyguA]

Great, happy reading :wave: