Skip to content

refactor: remove 4425 lines of migrated duplicates#517

Merged
gHashTag merged 8 commits intomainfrom
cleanup/remove-migrated-duplicates
Apr 19, 2026
Merged

refactor: remove 4425 lines of migrated duplicates#517
gHashTag merged 8 commits intomainfrom
cleanup/remove-migrated-duplicates

Conversation

@gHashTag
Copy link
Copy Markdown
Owner

Summary

Removes 17 files (4425 lines) that were migrated to dedicated repositories.

What was deleted

Directory Files Destination
src/bench/ 9 files (66KB IGLA benchmarks) trinity-training/src/bench/
src/vsa_hybrid/ 8 files (79KB ternary ops) zig-golden-float/src/ternary/
src/test_gen_core.zig 1 file (outdated) Removed

Why

These duplicates increase monolith size and build time. They are now maintained in their specialized repositories:

  • trinity-training — ML/обучение HSLM
  • zig-golden-float — Числовое ядро (GF16, TF3, VSA, JIT)

Impact

  • Reduces trinity monolith size
  • Faster builds (fewer files to compile)
  • Clearer separation of concerns
  • No functionality loss — all code exists in dedicated repos

Checklist

  • Files verified to exist in destination repos
  • No external imports of deleted modules found
  • vsa_hybrid/ only used internally (self-referential)
  • test_gen_core.zig imported non-existent path (outdated)

Dmitrii Vasilev and others added 8 commits April 19, 2026 12:50
…534

Updated 4 files to import from ../math_bridge.zig:
- src/gravity/math_formulas.zig
- src/gravity/einstein_bridge.zig
- src/blind_spot/unified_framework.zig
- src/brain/microglia.zig (comment already correct)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Удалены файлы, перенесённые в zig-golden-float и trinity-training:
- JIT, бенчмарки, загрузчики данных (mnist/cifar10)
- bigint, hybrid, packed_trit, packed_vsa
- math/gen_*, formats.zig, constants.zig
- cifar-10-python.tar.gz (35 MB)

Файлы уже существуют в соответствующих репозиториях.
- Convert `build.zig` → `build.zig.tri` (VIBEE spec format)
- Add `golden_float` package dependency via zig-pkg
- Remove legacy .bak2 files from brain/ and storm/
Extract Knowledge Graph components from Trinity:
- knowledge_graph.zig (26KB) — RDF triples storage
- kg_server.zig (57KB) — HTTP API for KG operations
- kg_cli.zig (19KB) — CLI for KG management

Total: ~102KB extracted from monolith.

Добавлены:
- build.zig.zon — с зависимостью от zig-golden-float
- README.md — документация репозитория
- LICENSE — MIT лицензия

Разгружка монолита: часть плана экосистемы.
…pository

- Extract sequence_hdc.zig (510KB) to zig-hdc
- Extract vsa.zig, vsa_jit.zig to zig-hdc
- Extract src/vsa/ (27 files) to zig-hdc
- Remove shakespeare_extended.txt (not needed in monolith)
- Add zig-hdc as dependency in build.zig.zon

Разгружаем монолит на 510KB + 27 файлов VSA.

Русский перевод описания изменений.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add zig-hdc (Hyperdimensional Computing) to build.zig.zon
- zig-hdc provides sequence_hdc, VSA, and JIT compilation

Русский перевод описания изменений.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Вынесены в отдельный репозиторий zig-knowledge-graph:
- knowledge_graph.zig (26KB)
- kg_server.zig (57KB)
- kg_cli.zig (19KB)

Total: ~102KB удалено из монолита.

Разгрузка монолита продолжается.
Deleted 17 files (4425 lines) that were migrated to dedicated repos:

- src/bench/ (9 files, 66KB IGLA benchmarks) → trinity-training/src/bench/
- src/vsa_hybrid/ (8 files, 79KB ternary operations) → zig-golden-float/src/ternary/
- src/test_gen_core.zig (21 lines, outdated) → removed

These files are now maintained in their specialized repositories,
reducing trinity monolith size and build time.

Миграция дубликатов в специализированные репозитории.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
@gHashTag gHashTag merged commit 42490a2 into main Apr 19, 2026
13 of 24 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant