risv3-relay/relay/config.py

"""Settings + config.yaml loader.

The relay reads two config sources:

1. ``.env`` for secrets (API key) and per-host overrides (status port,
   history cap). The ntfy topic is auto-generated on first run and
   written back to ``.env`` so it persists across restarts.

2. ``config.yaml`` for project-level settings: the registered CC
   sessions, the system prompt, the summarization prompt. First-PR
   default config is generated if the file doesn't exist, with one
   ``session-1`` registered.

Settings are read once at startup; the daemon does not hot-reload them.
"""

from __future__ import annotations

import os
import secrets
from dataclasses import dataclass, field
from pathlib import Path

import yaml
from dotenv import dotenv_values, set_key

REPO_ROOT = Path(__file__).resolve().parent.parent
ENV_FILE = REPO_ROOT / ".env"
CONFIG_FILE = REPO_ROOT / "config.yaml"

DEFAULT_SYSTEM_PROMPT = (
    "You are the chat-side counterpart to Claude Code instances working on the "
    "risv3 project. CC sends you its progress; you respond as a project lead "
    "would: check decisions, ask clarifying questions, approve or correct. "
    "When a CC turn raises a question only JC (the human owner) can answer, "
    "begin your response with the literal token [NEEDS-JC] so the relay daemon "
    "pauses and notifies JC. Otherwise reply normally and the relay forwards "
    "your reply to the originating CC session."
)

DEFAULT_SUMMARIZATION_PROMPT = (
    "Summarize the conversation so far. Preserve project context, decisions "
    "made, open work items, and any outstanding [NEEDS-JC] questions. The "
    "summary will replace earlier turns in the conversation history; the most "
    "recent turns will be retained verbatim. Be specific where specificity "
    "matters (file paths, issue numbers, decisions); be brief on routine "
    "back-and-forth."
)


@dataclass(frozen=True)
class SessionConfig:
    """One registered CC session."""

    session_id: str
    working_dir: str | None = None
    description: str | None = None


@dataclass(frozen=True)
class Settings:
    api_key: str
    model: str
    ntfy_topic: str
    status_port: int
    history_char_cap: int
    repo_root: Path
    queue_dir: Path
    dispatch_dir: Path
    state_dir: Path
    logs_dir: Path
    system_prompt: str
    summarization_prompt: str
    sessions: tuple[SessionConfig, ...] = field(default_factory=tuple)


def _ensure_runtime_dirs(repo_root: Path) -> None:
    for sub in ("queue", "dispatch", "state", "logs"):
        (repo_root / sub).mkdir(exist_ok=True)


def _generate_ntfy_topic() -> str:
    """Cryptographically random 16-character topic. Functionally a password."""
    return secrets.token_urlsafe(12)


def _load_or_init_ntfy_topic(env_values: dict[str, str | None]) -> str:
    raw = (env_values.get("NTFY_TOPIC") or "").strip()
    # Defensive: dotenv treats everything after = as the value, so an
    # inline `#` comment ends up as the topic. Treat any value that
    # starts with `#` (or contains whitespace, which a real topic never
    # does) as empty.
    if raw and not raw.startswith("#") and " " not in raw:
        return raw
    topic = _generate_ntfy_topic()
    set_key(str(ENV_FILE), "NTFY_TOPIC", topic, quote_mode="never")
    return topic


def _load_yaml_config() -> dict:
    if not CONFIG_FILE.exists():
        default = {
            "system_prompt": DEFAULT_SYSTEM_PROMPT,
            "summarization_prompt": DEFAULT_SUMMARIZATION_PROMPT,
            "sessions": [
                {
                    "session_id": "session-1",
                    "working_dir": None,
                    "description": "Default CC session",
                }
            ],
        }
        CONFIG_FILE.write_text(yaml.safe_dump(default, sort_keys=False))
        return default
    with CONFIG_FILE.open() as f:
        return yaml.safe_load(f) or {}


def load_settings() -> Settings:
    """Read .env + config.yaml. Mutates .env on first run to record the ntfy topic."""

    env_values = dotenv_values(ENV_FILE)
    api_key = (
        env_values.get("ANTHROPIC_API_KEY") or os.environ.get("ANTHROPIC_API_KEY") or ""
    ).strip()
    if not api_key:
        raise RuntimeError(f"ANTHROPIC_API_KEY missing from {ENV_FILE}")

    model = (env_values.get("ANTHROPIC_MODEL") or "claude-opus-4-7").strip()
    status_port = int((env_values.get("STATUS_PORT") or "8765").strip())
    history_char_cap = int((env_values.get("HISTORY_CHAR_CAP") or "400000").strip())
    ntfy_topic = _load_or_init_ntfy_topic(env_values)

    yaml_cfg = _load_yaml_config()
    system_prompt = str(yaml_cfg.get("system_prompt") or DEFAULT_SYSTEM_PROMPT)
    summarization_prompt = str(yaml_cfg.get("summarization_prompt") or DEFAULT_SUMMARIZATION_PROMPT)
    sessions_cfg = yaml_cfg.get("sessions") or []
    sessions = tuple(
        SessionConfig(
            session_id=str(s["session_id"]),
            working_dir=s.get("working_dir"),
            description=s.get("description"),
        )
        for s in sessions_cfg
    )

    _ensure_runtime_dirs(REPO_ROOT)

    return Settings(
        api_key=api_key,
        model=model,
        ntfy_topic=ntfy_topic,
        status_port=status_port,
        history_char_cap=history_char_cap,
        repo_root=REPO_ROOT,
        queue_dir=REPO_ROOT / "queue",
        dispatch_dir=REPO_ROOT / "dispatch",
        state_dir=REPO_ROOT / "state",
        logs_dir=REPO_ROOT / "logs",
        system_prompt=system_prompt,
        summarization_prompt=summarization_prompt,
        sessions=sessions,
    )