Kokoro tts

[ashum9]

Adds a local OpenAI-compatible TTS server using Kokoro ONNX.
Runs offline on CPU, port 8421. Handles Chinese text chunking (25 char limit for phoneme constraints) and auto-maps OpenAI voice names via af_ prefix. Includes mofa_hybrid.toml production config with Ollama + Kokoro providers.

Enables the existing mofa-fm pipeline to run fully offline no API key, no internet.

[Copilot]

Pull request overview

This PR introduces a local, offline Kokoro-based OpenAI-compatible TTS bridge and adds a new mofa-observability workspace crate that defines an event schema, in-memory metrics collection, and Prometheus text rendering, plus a mock harness and Grafana/Prometheus docker setup for validation.

Changes:

• Add kokoro_tts_server.py and a production mofa_hybrid.toml config wiring an OpenAI-compatible “kokoro” provider at http://127.0.0.1:8421/v1.
• Add mofa-observability crate with event types, metrics collector, and Prometheus text exposition renderer (with tests).
• Add docker-compose + Grafana dashboards + Prometheus config to exercise and visualize the new metrics via a mock harness.

Reviewed changes

Copilot reviewed 17 out of 19 changed files in this pull request and generated 8 comments.

Show a summary per file

File	Description
mofa-observability/src/prometheus.rs	Implements Prometheus text exposition rendering + tests.
mofa-observability/src/lib.rs	Exposes observability crate modules.
mofa-observability/src/events.rs	Defines the engine event schema used for telemetry.
mofa-observability/src/collector.rs	In-memory metrics state + collector loop and channel abstraction.
mofa-observability/examples/mock_harness.rs	Axum-based mock event generator + /metrics endpoint.
mofa-observability/docker/prometheus/prometheus.yml	Prometheus scrape config for the mock harness container.
mofa-observability/docker/grafana/provisioning/datasources/prometheus.yml	Grafana datasource provisioning for Prometheus.
mofa-observability/docker/grafana/provisioning/dashboards/dashboards.yml	Grafana dashboard provisioning configuration.
mofa-observability/docker/Dockerfile	Builds and packages the mock harness into a container image.
mofa-observability/docker/docker-compose.yml	Spins up mock harness + Prometheus + Grafana locally.
mofa-observability/dashboards/preflight_routing.json	Grafana dashboard for preflight/routing-related metrics.
mofa-observability/dashboards/memory_lifecycle.json	Grafana dashboard for memory/model lifecycle metrics.
mofa-observability/dashboards/engine_overview.json	Grafana dashboard for high-level engine health/latency views.
mofa-observability/Cargo.toml	Declares the new observability crate’s dependencies.
mofa_hybrid.toml	Adds a hybrid production config including the Kokoro provider.
kokoro_tts_server.py	Local OpenAI-compatible TTS HTTP server bridging to Kokoro ONNX.
Cargo.toml	Adds mofa-observability to the workspace members.
Cargo.lock	Locks new crate/dependency additions (incl. mofa-observability, rand).
.gitignore	Ignores .kokoro-models/ directory.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.