"""coreci/executor-config/app/main.py

Allowlist classifier (v4.1) — see coreci/executor-config/tools/coreci-executor/SPEC.md §11.
"""
import re

# v4 — pre-hoc classifier gate. Any pack containing a write verb against
#      read-only targets is NEEDS_APPROVAL.
WRITE_VERB_RE = re.compile(
    r"(?i)\b("
    r"curl\s+[^|;]*\s-X\s+(POST|PUT|PATCH|DELETE)"
    r"|curl\s+[^|;]*-(d|--data|--data-binary|--data-raw|urlencode)\s"
    r"|\bgit\s+push\b"
    r"|\bgit\s+commit\b(?![^|;\n]*--amend\b)"
    r"|\bgh\s+(pr|issue|release|repo)\s+(create|edit|close|merge|delete)\b"
    r"|\b(rm|mv|cp)\s+-[^\n]*\s+/\w"
    r")"
)

# v4.1 — mailbox-path exemption. Recognises the per-home mailbox envelope
#        from coreci/company/standards/communications.md §"Mailbox" (decisions 1–9).
#        Path shape: <home-repo>/mailbox/(outbox|inbox)/<file-or-dir>
#        where <home-repo> = coreci-<agent>/home or coreci/<agent>/home.
#        Does NOT exempt paths in scripts/ or anywhere outside mailbox/.
MAILBOX_PATH_RE = re.compile(
    r"^(?:coreci[-/][\w-]+/home|"
    r"coreci[-/]\w+)/mailbox/(?:outbox|inbox)(?:/[^\s]*)?$"
)


def current_allowlist():
    import os, yaml
    with open(os.environ.get(
            "ALLOWLIST_PATH", "allowlist.yaml"), "r") as f:
        return yaml.safe_load(f)


def _is_own_home(agent, target):
    """Return True if `target` is <agent>'s own home repo.

    Naming convention v2: agent field = Gitea username = allowlist key.
    Format: coreci-<role>. Two accepted shapes for the home repo path:
        coreci-<role>/home         (standard)
        coreci/<role>/home         (forward-compatible)
    """
    role = agent.replace("coreci-", "", 1)
    return target in {f"coreci-{role}/home", f"coreci/{role}/home"}


def _mailbox_only_targets(targets):
    """Return True iff every target is an own-home mailbox path."""
    for t in targets:
        if not MAILBOX_PATH_RE.match(t):
            return False
    return True


def classify(agent, targets, content, kind):
    al = current_allowlist()
    defaults = al["defaults"]
    agent_cfg = al["agents"].get(agent)
    if agent_cfg is None:
        return defaults["decision"]

    agent_kind_cfg = (
        agent_cfg.get("kinds", {}).get(kind, {})
        if isinstance(agent_cfg, dict) else {}
    )
    auto_run_targets = (
        set(defaults.get("auto_run_targets", []))
        | set(agent_kind_cfg.get(
                "auto_run_targets",
                agent_cfg.get("auto_run_targets", [])))
    )
    read_targets = (
        set(defaults.get("read_targets", []))
        | set(agent_kind_cfg.get(
                "read_targets",
                agent_cfg.get("read_targets", [])))
    )

    # v4.1 — Decision 6 mailbox-path exemption [1].
    # Mailbox writes are ALWAYS AUTO_RUN when:
    #   (a) every target is the sender's own home mailbox path, AND
    #   (b) every target is on the sender's own home (not a cross-agent mirror
    #   to someone else's <n>/home/mailbox/), AND
    #   (c) the content does not contain any non-mailbox write verb.
    # Cross-agent mailbox INBOX mirror (CAIRO writes to coreci-coo/home/mailbox/inbox/)
    # is NOT covered by this exemption — that needs explicit cross-agent grant.
    # That's the gap Decision 6 left; COO needs to add mailbox mirror grants in v4.2.
    if (_mailbox_only_targets(targets)
            and all(_is_own_home(agent, _home_part(t)) for t in targets)
            and not WRITE_VERB_RE.search(content)):
        return "AUTO_RUN"

    # v4 — pre-existing classifier, unchanged.
    if all(t in auto_run_targets for t in targets):
        return "AUTO_RUN"
    allowed_for_read = auto_run_targets | read_targets
    if all(t in allowed_for_read for t in targets):
        if not WRITE_VERB_RE.search(content):
            return "AUTO_RUN"
    return defaults["decision"]


def _home_part(target):
    """Return the <home-repo> portion of a mailbox target.

    Example: 'coreci-coo/home/mailbox/outbox/x.md' -> 'coreci-coo/home'
    Used to gate the Decision 6 exemption to own-home only.
    """
    parts = target.split("/mailbox/", 1)
    return parts[0] if parts else target