Two elements with strange pos and info classification

[udoprog]

Hi,

Sorry in advance since I'm not very well versed with data policy on JMDict. If something below is wrong, please advise!

I'm writing a machine conjugator based on JMdict. To improve coverage I've started reporting what seems to be classification mismatches, or words which have a pos that seems to contradict their structure.

This is the pseudocode for my inclusion criteria for words which can be conjugated:

for reading in reading_elements:
    if reading.is_search_only(): # sK tag
        continue;

    if reading.is_no_kanji || kanji_elements.is_empty():
        yield (None, reading)

    for kanji in kanji_elements:
        if kanji.is_search_only():
            continue;

        if reading.applies_to(kanji): # checks if reading string applies
            yield (Some(kanji), reading) # yields kanji with associated reading

With this, there's exactly two elements across verbs and adjectives which as far as I can tell seem to be misclassified:

id	kanji	pos	description
2222710	弱っちぃ	adj-i	Should this kanji element be marked as search only sK?
2858769	買い増す	v1, vt	This is marked as an ichidan verb.

Thank you

[JMdictProject]

Thanks for this bit of feedback.

• 買い増す (2858769). The POS was wrong. It's been changed to "v5s".
• 弱っちぃ (2222710), That "ぃ" is certainly inconsistent with a 形容詞, but it's also common enough in this case to record. Hiding it behind an [sK] would probably work for your purposes, but maybe allowing "ぃ" would be better.

As these anomalies involve individual entries, it would probably be best to raise them as comments in the entries themselves rather than use this (general) issue forum (https://www.edrdg.org/jmwsgi/entr.py?svc=jmdict&sid=&q=2858769)

I'll close this issue.