Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Typing issue: Number of people #689

Closed
hkiku482 opened this issue Feb 9, 2023 · 3 comments
Closed

Typing issue: Number of people #689

hkiku482 opened this issue Feb 9, 2023 · 3 comments

Comments

@hkiku482
Copy link

hkiku482 commented Feb 9, 2023 •
edited
Loading

Category of the typing issue
2. Word ranking issue (e.g. "夕日" is in the list, but ranking is lower than expected).

Issues
The following contents are out of the ranking.

input [e.g.ゆうひ] expected [e.g. 夕日] actual [e.g. ユウヒ]
ひとり 1人 (optional, write here)
ふたり 2人 (optional, write here)
... ... (optional, write here)

Version or commit-id
Mozc-2.26.4632.102+24.10.oss

Additional context
Other units may have same issue. "いっこ", "いっぽん" and so on.
@hiroyuki-komatsu
Copy link
Collaborator

Hi hkiku482,

Thank you for the feedback.
The latest version (2.28.5017) can already covert 1人, 2人, 1個.

We are investigating the case of 1本.

@hkiku482
Copy link
Author

@hiroyuki-komatsu
Thank you for your reply.
But, I noticed that this issue was a duplication with #672. Please close this if needed.

hiroyuki-komatsu added a commit that referenced this issue Mar 23, 2023
@hiroyuki-komatsu
Add 一本(いっぽん) of the same POS IDs with 一個(いっこ).
cf3e496 
* #689
* This does not solve the issue yet, because another entry of 一本 with different POS IDs is prioritized.

#typingquality

PiperOrigin-RevId: 517307936
hiroyuki-komatsu added a commit that referenced this issue Mar 23, 2023
@hiroyuki-komatsu
Update filtering rules for redundant candidates.
3c4d484 
* Without this change, candidates are filtered if another candidate has the same key and value.
* With this change, candidates are filtered only if another candidate has the same key, value, lid, and rid.
* This improves word ranking and candidate rewriting.
* Redundant candidates will be deduplicated in another layer.
* #689

#typingquality

PiperOrigin-RevId: 518537368
@hiroyuki-komatsu
Copy link
Collaborator

cf3e496 and 3c4d484 fix the case of 1本.

If the base form in Kanji (e.g. 一本) has proper POS IDs (e.g. lid: number, rid: counter suffix), it can be converted to variant forms (e.g. 1本).

In the previous implementation, if the base form has multiple entries with different POS IDs (e.g. [lid: number, rid: counter suffix], [lid: noun, rid: noun]), the variant conversion may not be performed.

Thank you for your feedback.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@hiroyuki-komatsu @hkiku482