Dictionary import

[skakavyponik]

Can u make tutorial how to import dictionary what is the format etc.
I can see in menu where to find it but cannot import anything
For me its Slovak language

[deniserdogan]

Also need it for Turkish language, found word list (60k words) changed words.txt name to tr_base.dict and get error

[RetroZelda]

it seems to be a custom .dict format that gets generated with the scripts/preprocess-dictionaries.main.kts script. but I cant get the script to run, both by cmd and in android studio(always exits with code 0 and prints nothing)

[pzauner]

Created a Pull request where you may do this once accepted:
#95

[deniserdogan]

Created a Pull request where you may do this once accepted: #95

No turkish :(
So many turkish letters btw maybe double tap to switch characters?
O -- Ö
U -- Ü
I -- İ
S -- Ş
C -- Ç
G -- Ğ
i -- ı

[pzauner]

@deniserdogan you may set this in the variations and even insert the first variation with a longpress.

As a german user I have the similar need to paste custom characters like umlaute ä ö ü or simply insert the € when holding e. Once merged this will become available.

see this other pull request by me:
#101

[palsoftware]

it seems to be a custom .dict format that gets generated with the scripts/preprocess-dictionaries.main.kts script. but I cant get the script to run, both by cmd and in android studio(always exits with code 0 and prints nothing)

there are py scripts to convert json with word and freq to .dict serialized.

[pzauner]

Another useful case to cover in this issue: bulk import into the personal/user dictionary.

A user reported that after editing a backup ZIP (adding entries to pastiera_prefs / user dictionary JSON), the new words can appear in the dictionary UI, but they do not appear as typing suggestions immediately.

That behavior makes sense from the current architecture (persistent user dictionary storage vs in-memory suggestion indexes / SymSpell), but from a user perspective it looks like the import "worked" only partially.

Suggested improvements for a future implementation:

1. Dedicated bulk user-dictionary import (TXT/CSV/JSON)
2. Stable documented import format (instead of editing pastiera_prefs internals)
3. Import validation (w, f, u fields if JSON is used) + error reporting
4. Single rebuild/reindex after import (not per-word)
5. Immediate suggestion refresh after import completes
6. Preview + deduplication before applying large imports
7. Optional: Export user dictionary in the same stable format

This would also help users building custom dictionaries / script variants (e.g. non-Latin or transliterated dictionaries) without needing to reverse-engineer the serialized .dict format.

[pzauner]

Created a Pull request where you may do this once accepted: #95

No turkish :( So many turkish letters btw maybe double tap to switch characters? O -- Ö U -- Ü I -- İ S -- Ş C -- Ç G -- Ğ i -- ı

Implemented in:
f2b479d.

Added a Turkish QWERTY-based multi-tap layout with direct access to Turkish characters via repeated taps:

• o -> ö
• u -> ü
• i -> ı (with correct uppercase handling: İ / I)
• s -> ş
• c -> ç
• g -> ğ

Also added optional circumflex variants for a more complete Turkish setup:

• a -> â
• i -> î
• u -> û

Locale mapping now includes Turkish (tr, tr_TR, tr_CY) so the layout can be selected automatically.

will be available in next release