Delete dead modules: openfang, paperclip, brain, swarm, cascade_adapter #23

Closed
opened 2026-03-14 13:33:34 +00:00 by Rockachopa · 0 comments
Owner

What

Multiple modules exist as stubs, scaffolding, or integrations to systems that don't exist yet. They add complexity, confuse agents reading the codebase, and inflate test/coverage overhead.

Modules to evaluate for deletion:

src/infrastructure/openfang/ — "Rust-compiled Agent OS" that doesn't exist. client.py (78 lines) + tools.py (41 lines, 0% coverage). Nothing calls it.

src/integrations/paperclip/ — Remote Paperclip AI API client. 4 files (~400 lines). Requires network access to a Paperclip server. Violates local-first principle unless it's a deliberate opt-in integration you still want.

src/brain/ — Distributed brain with rqlite. schema.py (0% coverage), worker.py (47% coverage), client.py, memory.py. Uses rqlite which isn't in the stack. If the goal is local-first SQLite, this whole module is dead weight.

src/swarm/ — Task queue and event log. Referenced by 10+ files but the task queue was explicitly removed. The remaining code is stubs returning "not available" messages.

src/timmy/cascade_adapter.py — Only referenced by dashboard/routes/router.py. 40 lines wrapping the cascade router. May be unnecessary indirection.

src/timmy/memory_migrate.py — One-shot migration script, 0% coverage, 8 print statements. Run it once if needed, then delete.

Why

~1500+ lines of dead code. Every line is cognitive overhead for you and any agent working on the codebase. Less code = faster understanding = better PRs.

What to do

  1. For each module, grep for imports to confirm nothing critical depends on it
  2. Delete the module directory/file
  3. Remove any imports, route registrations, and config settings that reference it
  4. Run make test, fix any broken imports in tests (likely just need to remove test files too)

Estimated effort: 1-2 hours

## What Multiple modules exist as stubs, scaffolding, or integrations to systems that don't exist yet. They add complexity, confuse agents reading the codebase, and inflate test/coverage overhead. ### Modules to evaluate for deletion: **`src/infrastructure/openfang/`** — "Rust-compiled Agent OS" that doesn't exist. `client.py` (78 lines) + `tools.py` (41 lines, 0% coverage). Nothing calls it. **`src/integrations/paperclip/`** — Remote Paperclip AI API client. 4 files (~400 lines). Requires network access to a Paperclip server. Violates local-first principle unless it's a deliberate opt-in integration you still want. **`src/brain/`** — Distributed brain with rqlite. `schema.py` (0% coverage), `worker.py` (47% coverage), `client.py`, `memory.py`. Uses `rqlite` which isn't in the stack. If the goal is local-first SQLite, this whole module is dead weight. **`src/swarm/`** — Task queue and event log. Referenced by 10+ files but the task queue was explicitly removed. The remaining code is stubs returning "not available" messages. **`src/timmy/cascade_adapter.py`** — Only referenced by `dashboard/routes/router.py`. 40 lines wrapping the cascade router. May be unnecessary indirection. **`src/timmy/memory_migrate.py`** — One-shot migration script, 0% coverage, 8 print statements. Run it once if needed, then delete. ## Why ~1500+ lines of dead code. Every line is cognitive overhead for you and any agent working on the codebase. Less code = faster understanding = better PRs. ## What to do 1. For each module, grep for imports to confirm nothing critical depends on it 2. Delete the module directory/file 3. Remove any imports, route registrations, and config settings that reference it 4. Run `make test`, fix any broken imports in tests (likely just need to remove test files too) ## Estimated effort: 1-2 hours
Sign in to join this conversation.
No Label
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Rockachopa/Timmy-time-dashboard#23