feat(migrations): update schema version to 3 and retire chat-v1 model

[M3gA-Mind]

Summary

• Retires chat-v1 as the default model: the backend removed it from its strict model registry, causing new sub-agent spawns (new inference threads) to receive 400 errors
• Updates DEFAULT_MODEL constant to reasoning-quick-v1 (Kimi K2.6 Turbo on Fireworks — same underlying model chat-v1 previously aliased to)
• Adds schema migration 2 → 3 (retire_chat_v1_model) that remaps any persisted config.default_model = "chat-v1" to "reasoning-quick-v1" on next launch
• Remaps hint:chat in the OpenHuman backend provider factory to reasoning-quick-v1 as well

Test plan

• Unit tests in retire_chat_v1_model.rs: remap when chat-v1, idempotent when already reasoning-quick-v1, no-op on None and other model values
• Migration integration tests in mod_tests.rs: fresh install and upgrade from v0 all advance to schema_version=3
• cargo test passes
• Existing workspace with default_model = "chat-v1" in config.toml is migrated to reasoning-quick-v1 on launch

Summary by CodeRabbit

• Chores
  • Default AI model changed from chat-v1 to reasoning-quick-v1
  • Existing configurations using chat-v1 are automatically remapped to reasoning-quick-v1 on next launch
  • Model selection logic updated to reflect the new default
  • Full backward compatibility ensured—no user action required

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: f9918a97-781e-4dc9-b085-971fd199ae75

📥 Commits

Reviewing files that changed from the base of the PR and between ebd6457 and 086218b.

📒 Files selected for processing (5)

• src/openhuman/config/schema/types.rs
• src/openhuman/inference/provider/factory.rs
• src/openhuman/migrations/mod.rs
• src/openhuman/migrations/mod_tests.rs
• src/openhuman/migrations/retire_chat_v1_model.rs

---

📝 Walkthrough

Walkthrough

This PR retires the chat-v1 model by updating the default model selection and hint routing to use reasoning-quick-v1, implements a versioned configuration migration that persists the change for existing configs, and verifies the entire migration lifecycle through test coverage.

Changes

Chat-v1 Model Retirement and Migration

Layer / File(s)	Summary
Model default and hint routing updates src/openhuman/config/schema/types.rs, src/openhuman/inference/provider/factory.rs	DEFAULT_MODEL constant updated from MODEL_CHAT_V1 to MODEL_REASONING_QUICK_V1 with retired model documentation; hint:chat routing remapped to MODEL_REASONING_QUICK_V1 in the backend factory.
Chat-v1 retirement migration module src/openhuman/migrations/retire_chat_v1_model.rs	New migration module introduces MigrationStats struct and run function to remap default_model from chat-v1 to reasoning-quick-v1 in-memory with info/debug logging; includes unit tests for remapping, preservation of other model values, None handling, and idempotence.
Migration system versioning and wiring src/openhuman/migrations/mod.rs	CURRENT_SCHEMA_VERSION bumped from 2 to 3; retire_chat_v1_model module registered; run_pending startup runner extended with schema_version==2 gate that executes the migration, persists the config, and rolls back the in-memory version on save failure with retry-on-next-launch semantics.
Migration lifecycle test coverage src/openhuman/migrations/mod_tests.rs	Migration integration tests updated to expect schema_version 3, verifying version advancement from older schemas, successful versioning on fresh installs, and idempotent behavior on repeated invocations.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~12 minutes

Possibly related PRs

• tinyhumansai/openhuman#1801: Overlapping client-side routing translation normalizing hint:chat to reasoning-quick-v1.
• tinyhumansai/openhuman#1761: Companion provider/router wiring for reasoning-quick-v1 model tier selection.

Suggested labels

agent, working

Poem

🐰 whiskers twitch with glee
Chat-v1 hops away with grace,
Quick reasoning takes its place—
Configs migrate, smooth as tea,
Schema v3, wild and free! 🚀

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}