Fix OCR user dictionary add-actions

[HellbringerOnline]

Summary

• use the displayed or confirmed OCR word when adding unknown words to the user dictionary
• refresh the in-memory OCR dictionary immediately after popup add-actions so the current OCR run sees the new user word right away
• keep the side-panel unknown word add-action UI-only and avoid mutating the current unknown word list there

Problem

Some OCR add-actions used the raw OCR token instead of the displayed or corrected word. In addition, popup flows did not refresh the current OCR dictionary state immediately after adding a user word, so the just-added word could still be treated as unknown until a later refresh.

Changes

• add AddUserWord(string word) to the OCR fix engine and update the runtime dictionary state immediately
• add DisplayWord and ResolveSubmittedWord(...) so add-actions consistently submit the visible or confirmed word
• keep popup traversal cleanup where it belongs, but remove unknown-word list cleanup from the side-panel add-action
• cover the behavior with OCR-focused tests

Validation

• dotnet test tests/UI/UITests.csproj --filter "FullyQualifiedName~UnknownWordItemTests|FullyQualifiedName~OcrFixEngineUserDictionaryTests"
• dotnet build src/UI/UI.csproj -c Debug --no-restore

[niksedk]

Looks good :)

RemoveUnknownWordsFromCurrentState shouldn’t be called here, as the UI unknown word list actions are only related to the UI.

Also, it would be great to keep PRs focused on a single concern. Could you move the Unicode normalization changes into a separate PR?

[HellbringerOnline]

Looks good :)

RemoveUnknownWordsFromCurrentState shouldn’t be called here, as the UI unknown word list actions are only related to the UI.

Also, it would be great to keep PRs focused on a single concern. Could you move the Unicode normalization changes into a separate PR?

Ok, I'll try.

[niksedk]

Looks good :)
RemoveUnknownWordsFromCurrentState shouldn’t be called here, as the UI unknown word list actions are only related to the UI.
Also, it would be great to keep PRs focused on a single concern. Could you move the Unicode normalization changes into a separate PR?

Ok, I'll try.

thx :)

[niksedk]

I think this does too much - and the remove stuff is not wanted.