feat: make mnemonic loading format-first

[Naman015]

What is this PR for?

Load Mnemonic currently asks users to choose between camera, manual input, and storage before choosing the backup they recognise. This PR proposes a format-first start instead:

1. QR Code
2. Words
3. From Storage
4. Other Formats
5. Back

Other Formats is one flat menu containing Tinyseed, Tinyseed (Bits), OneKey KeyTag, Binary Grid, Word Numbers, Stackbit 1248, and Back.

The intent is to remove one translation step from the common recovery paths without changing any decoder, validation, storage, keypad, or cryptographic behaviour. The same shared loader is used by Login and Mnemonic XOR.

Design reference: Direction 3: format-first mnemonic recovery

The common routes become shorter, while specialist formats move one level deeper:

Format	Current actions	Proposed actions
QR Code	2	1
Words	3	2
From Storage	3	3
Tinyseed	3	5
Tinyseed (Bits)	5	6
OneKey KeyTag	4	7
Binary Grid	5	8
Word Numbers	4	9
Stackbit 1248	6	10

The action count starts on the initially focused row and includes every navigation and Enter action.

Changes made to:

• Code
• Tests
• Docs
• CHANGELOG

Did you build the code and tested on device?

• Yes, build and tested on TZT

Simulator acceptance completed on TZT, M5StickV, and Amigo. All 11 translation files were checked against M5StickV’s 135-pixel wrapping behaviour. Simulator captures cover every wrapping locale: es-MX, ko-KR, pt-BR, and tr-TR. All wraps are clean, with no clipping or overlap, and every menu item plus Back remains visible. Physical TZT testing confirmed QR Code, Words, From Storage, both Back levels, and button and touch navigation.

What is the purpose of this pull request?

• Bug fix
• New feature
• Docs update
• Other

Verification

• 67 focused Login, Mnemonic XOR, and loader tests
• 1,125 full-suite tests passed, with one expected failure and 97% coverage
• Black formatting
• Pylint 10.00/10
• translation validation
• Vulture

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 97.47%. Comparing base (be5eda2) to head (53d96d8).
⚠️ Report is 5 commits behind head on develop.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #922      +/-   ##
===========================================
+ Coverage    97.42%   97.47%   +0.05%     
===========================================
  Files           83       84       +1     
  Lines        10880    10952      +72     
===========================================
+ Hits         10600    10676      +76     
+ Misses         280      276       -4     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[qlrd]

cACK. A doubt: the translation stress occured only in pt-BR and es-MX?

[Naman015]

the translation stress occured only in pt-BR and es-MX?

I checked all 11 translations on M5StickV. Wrapping occurs in es-MX, ko-KR, pt-BR, and tr-TR. es-MX wraps From Storage, Binary Grid, and Word Numbers; ko-KR wraps From Storage; pt-BR and tr-TR wrap Word Numbers. They all fit cleanly with Back still visible. TZT and Amigo don't wrap. Thanks for pointing it, I updated the PR body now.

I also missed that Other Formats was still hard-coded in English, so I fixed that and refreshed the M5StickV captures.

[qlrd]

First round docs nits (no blockable):

[Naman015]

First round docs nits (no blockable):

Thank you, fixed the weird wrapping and punctuation nits in docs.

[qlrd]

overall LGTM. Tested with:

gh pr checkout 922
git submodule update --recursive
uv sync --frozen --reinstall-package=uUR
uv run poe pre-commit
./krux build-release
./build/ktool-linux -B {goE,dan}-b 1500000 ./krux-v28.06.0/maixpy_{amigo,wonder_mv}/kboot.kfpkg

so i would give a tACK 993acc6 with some minor nits (non-blockable, maybe worth of discussion)

[odudex]

tACK - tested on TZT

I like the rearrangement. Would only change 2 things:

• To differentiate manual input from scans, add "scan" term on respective items of "Other Formats". "Scan Tinyseed", "Scan OneKey". Or "Tinyseed (scan)".
• Take the opportunity to rename "Tinyseed (Bits)" to a more generic "Binary Grid", or even "Binary Grid (manual)"

[Naman015]

tACK - tested on TZT

I like the rearrangement. Would only change 2 things:

• To differentiate manual input from scans, add "scan" term on respective items of "Other Formats". "Scan Tinyseed", "Scan OneKey". Or "Tinyseed (scan)".
• Take the opportunity to rename "Tinyseed (Bits)" to a more generic "Binary Grid", or even "Binary Grid (manual)"

Thanks for testing. Agreed, Binary Grid (manual/scan) makes more sense to me here and also keeps the name generic.

I’ve pushed the new labels:

• Tinyseed (scan)
• Binary Grid (manual)
• OneKey KeyTag (scan)
• Binary Grid (scan)

The translations, tests, docs and screenshots are updated too.