Skip to content

DUPLICATES.md

Latest commit

 

History

History
234 lines (159 loc) · 12.6 KB

DUPLICATES.md

File metadata and controls

234 lines (159 loc) · 12.6 KB

DUPLICATES.md — Definitive Register of Duplicated Files and Directories

Last updated: 2026-04-23 (v1.0 callout added 2026-05-03) Author: Védis Eikleið (Cartographer); v1.0 callout by Runa Gridweaver Scope: Every duplicate pair or set I could verify in Viking-Code-Mythic-Engineering-CLI-Vibe-Coding on branch development. Companion scrolls: MAP.md, ARCHITECTURE.md, DEPENDENCIES.md, YGGDRASIL_COMPARISON.md, IMPACT_integration.md.

v1.0.0 callout (2026-05-03). This survey catalogs duplications across the whole monorepo, including vendor mirrors and dormant islands. The active product (mythic_vibe_cli/) does NOT carry the duplicates listed here — they live in dormant islands. The boundary law in REPO_BOUNDARY.md ensures these duplications cannot leak into the v1.0 product surface. For the active-product-only architecture record, see docs/ARCHITECTURE.md.

Symbol legend (recommendation labels)

  • [IDENTICAL] — byte-for-byte same content (md5 matched).
  • [DRIFT] — similar but diverged (checksum differs; size close).
  • [SHELL] — one side is a stub / empty placeholder.
  • [PARTIAL] — one side is a strict (or near-strict) subset of the other.
  • [NAME-ONLY] — same name, unrelated content (no real duplication).

I recommend no action here. This scroll is the ledger — Eirwyn's RECOMMENDATIONS.md will advise keep/merge/drop.

1. Top-level doctrinal markdown — four-way copies

These files appear at repo root AND inside each subproject, identical.

File Copies Sizes (lines/bytes) Label
PHILOSOPHY.md root, mindspark_thoughtform/, mindspark_thoughtform/MindSpark_ThoughtForge/, WYRD-Protocol-.../ md5 67078d2d... across all 4 [IDENTICAL]
RULES.AI.md root, mindspark_thoughtform/, mindspark_thoughtform/MindSpark_ThoughtForge/, WYRD-Protocol-.../ md5 fec75d83... across all 4 [IDENTICAL]
Technical_Architecture_of_Volmarrs_AI_Ecosystem.md root, mindspark_thoughtform/, WYRD-Protocol-.../ md5 0a56e825... across all 3 [IDENTICAL]
WORLD_MODELING_SKILL.md root, mindspark_thoughtform/, WYRD-Protocol-.../ md5 505e2afd... across all 3 [IDENTICAL]

2. research_data/ — three-way full copy

Three complete copies of the 41-doc research series (00–40 plus supporting specs, pyproject.toml, src/, tests/, wyrd_runtime/):

Path Status
research_data/ canonical (has pyproject.toml, src/, tests/, config/, examples/, scripts/)
mindspark_thoughtform/research_data/ full copy of the .md corpus (no src/, no tests/)
WYRD-Protocol-World-.../research_data/ full copy of the .md corpus (no src/, no tests/)

Sample md5 checks:

File All three copies Label
00_INDEX.md md5 7fabffaa... across all 3 [IDENTICAL]
01_architecture_overview.md md5 0e365a29... across all 3 [IDENTICAL]

Recommendation label for the corpus as a whole: [IDENTICAL] for the MDs, [PARTIAL] for the tree as a whole (only root has src/, tests/, pyproject.toml, config/, examples/, scripts/).

3. The three wyrdforge trees

Two Python package trees plus a reference-only folder.

Path Completeness Has __init__.py? pyproject.toml name? Label
WYRD-Protocol-.../src/wyrdforge/ Full — bridges/, ecs/, evals/, hardening/, llm/, loaders/, models/, oracle/, persistence/, runtime/, schemas/, security/, services/ (14 sub-packages) yes wyrdforge v1.0.0 canonical
research_data/src/wyrdforge/ Strict subset — only models/, runtime/, schemas/, security/, services/ (5 sub-packages) no (no __init__.py anywhere in the tree) wyrdforge v0.1.0 [PARTIAL]
mindspark_thoughtform/research_data/src/ (not checked as full tree — appears similar to research_data/src/) — — [PARTIAL] (inherits research_data layout)

3.1 File-level comparison — WYRD canonical vs research_data/src/wyrdforge

Spot-checked md5:

File WYRD research_data WYRD lines RD lines Label
models/bond.py 1c9769d8... 1c9769d8... — — [IDENTICAL]
models/memory.py c3b17e10... c3b17e10... — — [IDENTICAL]
security/permission_guard.py 6b56c39f... 6b56c39f... — — [IDENTICAL]
services/bond_graph_service.py ba0b64bd... 744693b1... 61 60 [DRIFT] (1-line delta)
services/memory_store.py 20d3538e... 8fe6f3ff... 69 68 [DRIFT] (1-line delta)
runtime/demo_seed.py 869580920c... 6d4e14bd... 33 32 [DRIFT] (1-line delta)

3.2 Sub-packages missing entirely on the research_data side

Present in WYRD, absent in research_data/src/wyrdforge/:

  • bridges/ (9 files)
  • ecs/ (entire ECS core)
  • evals/
  • hardening/
  • llm/
  • loaders/
  • oracle/
  • persistence/

Present in both:

  • models/ — full bond.py, common.py, evals.py, memory.py, micro_rag.py, persona.py
  • runtime/ — RD has only demo_seed.py; WYRD adds character_context.py, turn_loop.py
  • schemas/ — all 14 .schema.json files identical
  • security/ — 2 files identical
  • services/ — RD has 5 of 11 (bond_graph_service.py, memory_store.py, micro_rag_pipeline.py, persona_compiler.py, truth_calibrator.py); WYRD adds contradiction_detector.py, memory_promoter.py, memory_to_rag.py, runic_engine.py, writeback_engine.py, __init__.py

3.3 Crucial namespace note

Both WYRD-Protocol-.../pyproject.toml AND research_data/pyproject.toml declare name = "wyrdforge". If both were installed, they would collide on package name (pip/setuptools would refuse, or the second install overwrites the first). Absent __init__.py in research_data/src/wyrdforge/, it is not currently importable as a package anyway.

Recommendation label for the full comparison: [PARTIAL] + [DRIFT] — research_data/ is a partial, slightly-older snapshot; WYRD is canonical.

4. NSE systems — root copy vs imports/norsesaga/ copy

File systems/ (root) imports/norsesaga/systems/ Size (lines) Label
event_dispatcher.py present, md5 8fd2f248... present, md5 8fd2f248... 145 / 145 [IDENTICAL]
world_dreams.py absent present, 140 lines — unique to imports/
world_will.py absent present, 182 lines — unique to imports/

imports/norsesaga/systems/ is not a duplicate subset — it is a tiny companion directory adding two files (world_dreams.py, world_will.py) that the root systems/ does not carry, plus one that is identical.

Recommendation label: event_dispatcher.py = [IDENTICAL]; the two world_*.py files = neither duplicate nor shell — they are additions.

First-pass note H-4 (in MAP.md) framed this as "two copies of NSE systems, each incomplete"; the more precise reading is: the root systems/ has 28 modules (the full engine), and imports/norsesaga/systems/ has 3 modules where 1 is identical and 2 are additive.

5. NSE engine name duplication — wyrd_system.py

Two unrelated files share this name:

Path Purpose Label
systems/wyrd_system.py NSE "Three Wells / Norns" fate system (~33 KB per INVENTORY.md) —
yggdrasil/core/wyrd_system.py Yggdrasil's own wyrd subsystem —

These are [NAME-ONLY] — different implementations, different packages, not duplicates in the byte sense. Flag only because reader confusion is guaranteed.

6. Yggdrasil name duplication — already mapped

See YGGDRASIL_COMPARISON.md for full side-by-side.

Path Kind Label
yggdrasil/ (top-level package) NSE cognitive router [NAME-ONLY] (vs WYRD)
WYRD-Protocol-.../src/wyrdforge/ecs/yggdrasil.py WYRD spatial hierarchy service [NAME-ONLY] (vs NSE)

7. MindSpark internal shell

Path Contents Label
mindspark_thoughtform/ (outer) Full project (src/thoughtforge/, tests/, configs/, data/, research_data/, etc.) canonical
mindspark_thoughtform/MindSpark_ThoughtForge/ (inner) Only PHILOSOPHY.md, README.md, RULES.AI.md [SHELL]

The inner folder has no Python, no tests, no config — only three doctrinal MDs that are themselves identical to the ones at the root (see section 1). Visible artifact of a renaming or nested-clone operation.

8. docs/specs/ — root vs mindspark_thoughtform/docs/specs/

All 27 spec files in docs/specs/ are byte-identical copies of those in mindspark_thoughtform/docs/specs/. Sample md5 checks:

File root mindspark Label
Algorithms_and_Pseudocode_Spec.md 9364a488... 9364a488... [IDENTICAL]
ThoughtForge_Complete_Implementation_Guide.md afc5d1f4... afc5d1f4... [IDENTICAL]
GALDRABOK_PREFACE.md 21df822f... 21df822f... [IDENTICAL]

Recommendation label for the full set: [IDENTICAL] — either the root has been "lifted" from MindSpark, or MindSpark and root share a sync.

9. Emotional Engine Integration Plan — space-vs-underscore drift

File md5 Lines Label
Emotional Engine Integration Plan for Norse Saga Engine.md 2611fc29... 299 older/shorter
Emotional_Engine_Integration_Plan_for_Norse_Saga_Engine.md e5e8f776... 397 newer/longer

Same title, different filenames (space vs underscore). Not identical — the underscored version is ~33% longer. Label: [DRIFT].

10. Mythic Engineering CLI doctrine — filename family

These live at repo root and share concept but not bytes (unchecked for md5 here; filenames alone show drift):

  • Mystic_Engineering_Protocals1.0.md (170 KB, note: typo Mystic/Protocals)
  • Mythic_Engineers_Codex.md (88 KB)
  • Mythic_Engineering_CLI_Design_Ideas_7373y4yj.md (47 KB)
  • MYTHIC_ENGINEERING.md (generated by cli.cmd_init)
  • practical_mythic_engineering_step_by_step.md
  • Quick_Guide_to_Mythic_Engineering_Vibe_Coding.md
  • Ada_Lovelace_Explains_Mythic_Engineering.md

Label: [NAME-ONLY] — different docs on a shared theme; treat as one corpus, not as duplicates.

11. arXiv data dumps

File Size md5 Consumed by? Label
arxiv_results.json 44 KB — scripts/parse_arxiv_and_generate.py (live) canonical
arxiv_all_papers.json 77 KB — nothing redundant
arxiv_papers.json 9 KB — nothing redundant
relevant_papers.json 67 KB — nothing redundant

Four separate dumps, only one live consumer. I have not md5-compared for drift (they differ in size, so they are not identical). Label: [NAME-ONLY] — same corpus family, different snapshots / subsets.

12. Ephemeral runtime artifacts (noted for completeness — not duplicates)

Path Kind Note
diagnostics/turn_trace.jsonl 46 MB JSONL Only one copy; a captured NSE run log. Not a duplicate. Included here only because its size (46 MB) is unusual for a source repo.

13. Summary table

Duplication class Count Worst case Lowest-risk action
Identical doctrinal MDs across 3–4 roots 4 files × 3–4 locations — Decide on a single canonical root; keep others as symlinks or generate-on-install.
Identical docs/specs across root + mindspark 27 files × 2 locations — One of these should be the source.
Identical research_data corpus across 3 locations ~41 MDs × 3 — Root research_data/ should be canonical; inner copies can be pruned.
wyrdforge partial duplicate (root vs WYRD) ~16 files present in both, 3 drift partial + drift See IMPACT_integration.md section on WYRD.
NSE systems vs imports/norsesaga/systems 1 identical + 2 unique — Identify whether world_dreams.py / world_will.py are needed; absorb or drop.
MindSpark internal [SHELL] 1 dir — Remove or document.
Emotional Engine plan drift 2 files, different sizes — Pick the longer (newer) as canonical; archive the shorter.
Name-only collisions Yggdrasil (2), wyrd_system (2) — Document the disambiguation in the future integration README.

Nothing here is broken by the duplication itself — Python would not crash, the CLI runs regardless. These are maintenance and clarity costs, not runtime costs, with one exception: the research_data/pyproject.toml declaring name = "wyrdforge" is a latent install-time collision with the real WYRD package. That one is a ticking watch, not a present fire.