feat: controller-owned allocator API on control-plane DB (Closes #600) #622

Merged
sysadmin merged 1 commits from feat/issue-600-controller-allocator-api into master 2026-07-10 02:12:00 -05:00
5 changed files with 1305 additions and 1 deletions
+610
View File
@@ -0,0 +1,610 @@
"""Controller-owned work allocator policy (#600).
Builds on the #613 control-plane DB substrate (``ControlPlaneDB.assign_and_lease``).
Workers must not self-select exclusive work under the standard multi-LLM
workflow. They call ``gitea_allocate_next_work`` which:
1. Inspects candidate Gitea issues/PRs (never raw monitoring incidents).
2. Applies ADR routing policy (role, terminal path, leases, blocked, deps).
3. Atomically assigns + leases the selected item in one DB transaction.
This module is pure selection + substrate orchestration. Gitea I/O for live
inventory lives in the MCP tool wrapper so tests can inject candidates.
#612 remains downstream: bridge-created Gitea issues become candidates only
after they exist as normal issues; this module never assigns incidents.
"""
from __future__ import annotations
import os
import uuid
from dataclasses import dataclass, field
from typing import Any, Sequence
from control_plane_db import (
ControlPlaneDB,
ControlPlaneError,
InvalidWorkKindError,
LeaseRequiredError,
WORK_KINDS,
)
# Outcomes required by #600 / ADR §5.
OUTCOME_ASSIGNED = "assigned_work"
OUTCOME_WAIT = "wait"
OUTCOME_BLOCKED_TERMINAL = "blocked_by_terminal_path"
OUTCOME_BLOCKED_LEASE = "blocked_by_active_lease"
OUTCOME_NEEDS_CONTROLLER = "needs_controller"
OUTCOME_NO_SAFE = "no_safe_work"
OUTCOME_ROLE_INELIGIBLE = "role_ineligible"
OUTCOME_PREVIEW = "preview" # dry-run only (apply=false)
ROLE_AUTHOR = "author"
ROLE_REVIEWER = "reviewer"
ROLE_MERGER = "merger"
ROLE_RECONCILER = "reconciler"
ROLE_CONTROLLER = "controller"
VALID_ROLES = frozenset(
{ROLE_AUTHOR, ROLE_REVIEWER, ROLE_MERGER, ROLE_RECONCILER, ROLE_CONTROLLER}
)
# Default action matrices by role (mutation gate will re-check).
ROLE_ACTIONS: dict[str, tuple[tuple[str, ...], tuple[str, ...]]] = {
ROLE_AUTHOR: (
("implement", "comment", "push", "create_pr"),
("approve", "merge", "request_changes", "self_select_without_assignment"),
),
ROLE_REVIEWER: (
("review", "comment", "approve", "request_changes"),
("merge", "push", "create_pr", "self_select_without_assignment"),
),
ROLE_MERGER: (
("merge", "comment"),
("approve", "request_changes", "push", "create_pr", "self_select_without_assignment"),
),
ROLE_RECONCILER: (
("comment", "diagnose", "cleanup"),
("approve", "merge", "push", "create_pr", "self_select_without_assignment"),
),
ROLE_CONTROLLER: (
("comment", "diagnose", "allocate"),
("approve", "merge", "push", "create_pr"),
),
}
@dataclass
class WorkCandidate:
"""One assignable Gitea issue or PR presented to the allocator."""
kind: str # issue | pr
number: int
state: str = "open"
labels: tuple[str, ...] = ()
title: str = ""
priority: int = 0
head_sha: str | None = None
# Routing signals (callers derive from Gitea / review feedback).
request_changes_current_head: bool = False
approval_on_current_head: bool = False
approval_stale: bool = False
approval_contaminated: bool = False
mergeable: bool = False
blocked: bool = False
dependency_unmet: bool = False
dependency_reason: str | None = None
already_claimed_elsewhere: bool = False
def __post_init__(self) -> None:
self.kind = (self.kind or "").strip().lower()
self.state = (self.state or "open").strip().lower()
self.labels = tuple(
str(x).strip().lower() for x in (self.labels or ()) if str(x).strip()
)
if self.kind not in WORK_KINDS:
raise InvalidWorkKindError(
f"candidate kind '{self.kind}' is not assignable; only "
f"{sorted(WORK_KINDS)} (never raw incidents)"
)
def as_dict(self) -> dict[str, Any]:
return {
"kind": self.kind,
"number": self.number,
"state": self.state,
"labels": list(self.labels),
"title": self.title,
"priority": self.priority,
"head_sha": self.head_sha,
"request_changes_current_head": self.request_changes_current_head,
"approval_on_current_head": self.approval_on_current_head,
"approval_stale": self.approval_stale,
"approval_contaminated": self.approval_contaminated,
"mergeable": self.mergeable,
"blocked": self.blocked,
"dependency_unmet": self.dependency_unmet,
"dependency_reason": self.dependency_reason,
}
@dataclass
class SkipRecord:
kind: str
number: int
reason: str
def as_dict(self) -> dict[str, Any]:
return {"kind": self.kind, "number": self.number, "reason": self.reason}
def normalize_role(role: str | None, *, profile_name: str | None = None) -> str:
"""Map profile/role strings to a canonical allocator role."""
raw = (role or "").strip().lower()
if raw in VALID_ROLES:
return raw
prof = (profile_name or "").strip().lower()
for token in VALID_ROLES:
if token in prof or prof.endswith(f"-{token}"):
return token
if "author" in raw:
return ROLE_AUTHOR
if "review" in raw:
return ROLE_REVIEWER
if "merg" in raw:
return ROLE_MERGER
if "reconcil" in raw:
return ROLE_RECONCILER
if "control" in raw:
return ROLE_CONTROLLER
raise ControlPlaneError(
f"unknown allocator role '{role}' (profile={profile_name!r}); "
f"expected one of {sorted(VALID_ROLES)}"
)
def expected_role_for_candidate(c: WorkCandidate) -> str:
"""ADR §5.3 routing: which role should take this work next."""
if c.kind == "pr":
if c.approval_contaminated:
return ROLE_RECONCILER
if c.request_changes_current_head:
return ROLE_AUTHOR
if c.approval_stale:
return ROLE_REVIEWER
if c.approval_on_current_head and c.mergeable:
return ROLE_MERGER
# Open PR without terminal verdict → reviewer
return ROLE_REVIEWER
# Issues: ready work → author by default; blocked stays controller/none
labels = set(c.labels)
if "status:blocked" in labels or c.blocked:
return ROLE_CONTROLLER
return ROLE_AUTHOR
def role_actions(role: str) -> tuple[tuple[str, ...], tuple[str, ...]]:
return ROLE_ACTIONS.get(role, ROLE_ACTIONS[ROLE_AUTHOR])
def classify_skip(
c: WorkCandidate,
*,
role: str,
terminal_pr: int | None,
) -> str | None:
"""Return skip reason, or None if candidate is selectable for *role*."""
if c.state in ("merged", "closed"):
return f"{c.kind}#{c.number} is {c.state}; never assign"
if c.blocked or "status:blocked" in c.labels:
return f"{c.kind}#{c.number} is blocked"
if c.dependency_unmet:
return (
c.dependency_reason
or f"{c.kind}#{c.number} has unmet dependencies"
)
if c.already_claimed_elsewhere:
return f"{c.kind}#{c.number} already claimed elsewhere"
if c.kind == "pr" and not (c.head_sha or "").strip():
return f"pr#{c.number} missing head_sha pin"
# Terminal path first: when an active terminal PR exists, only that PR
# (or controller diagnosis) is assignable for review-path roles.
if terminal_pr is not None and c.kind == "pr" and c.number != terminal_pr:
if role in (ROLE_REVIEWER, ROLE_MERGER):
return (
f"pr#{c.number} skipped: active terminal-review lock on "
f"PR #{terminal_pr} must be resolved first"
)
expected = expected_role_for_candidate(c)
if role == ROLE_CONTROLLER:
# Controller may inspect anything but only assigns diagnosis targets
# when contaminated / blocked.
if expected == ROLE_RECONCILER or c.blocked:
return None
return f"{c.kind}#{c.number} does not require controller (expected {expected})"
if role != expected:
return (
f"{c.kind}#{c.number} expects role '{expected}', active role is '{role}'"
)
# Ready-gate for issues: prefer status:ready when labels present.
if c.kind == "issue" and c.labels:
if "status:ready" not in c.labels and "status:in-progress" not in c.labels:
# Allow unlabeled open issues; only skip explicit non-ready states.
if any(l.startswith("status:") for l in c.labels):
return f"issue#{c.number} not status:ready ({','.join(c.labels)})"
return None
def sort_candidates(candidates: Sequence[WorkCandidate]) -> list[WorkCandidate]:
"""Higher priority first; then lower number (older issues) for stability."""
return sorted(
candidates,
key=lambda c: (-int(c.priority), c.kind != "pr", int(c.number)),
)
def allocate_next_work(
db: ControlPlaneDB,
*,
session_id: str,
role: str,
remote: str,
org: str,
repo: str,
candidates: Sequence[WorkCandidate],
apply: bool = False,
profile_name: str | None = None,
username: str | None = None,
lease_ttl_seconds: int | None = None,
) -> dict[str, Any]:
"""Select and optionally reserve the next work unit via control-plane DB.
*apply=False* (default): dry-run selection only — no lease/assignment.
*apply=True*: atomic ``assign_and_lease`` for the selected candidate.
Never uses file locks or comment-only leases as the assignment source.
"""
if db is None:
return {
"success": False,
"outcome": OUTCOME_NO_SAFE,
"reasons": [
"control-plane DB substrate unavailable (fail closed, #600/#613)"
],
"skipped": [],
"assignment": None,
"substrate": "control_plane_db",
"file_lock_only": False,
"comment_lease_only": False,
}
try:
role_norm = normalize_role(role, profile_name=profile_name)
except ControlPlaneError as exc:
return {
"success": False,
"outcome": OUTCOME_ROLE_INELIGIBLE,
"reasons": [str(exc)],
"skipped": [],
"assignment": None,
"substrate": "control_plane_db",
}
session_id = (session_id or "").strip() or f"alloc-{uuid.uuid4().hex[:12]}"
try:
db.upsert_session(
session_id=session_id,
role=role_norm,
profile=profile_name,
pid=os.getpid(),
)
except Exception as exc: # noqa: BLE001 — surface structured
return {
"success": False,
"outcome": OUTCOME_NO_SAFE,
"reasons": [
f"failed to register session in control-plane DB: {exc} "
"(fail closed, #613)"
],
"skipped": [],
"assignment": None,
"substrate": "control_plane_db",
}
# Expire stale leases globally before selection.
try:
db.expire_stale_leases()
except Exception as exc: # noqa: BLE001
return {
"success": False,
"outcome": OUTCOME_NO_SAFE,
"reasons": [f"lease expiry failed: {exc} (fail closed)"],
"skipped": [],
"assignment": None,
"substrate": "control_plane_db",
}
terminal = None
try:
terminal = db.get_active_terminal_lock(remote=remote, org=org, repo=repo)
except Exception as exc: # noqa: BLE001
return {
"success": False,
"outcome": OUTCOME_NO_SAFE,
"reasons": [f"terminal lock lookup failed: {exc} (fail closed)"],
"skipped": [],
"assignment": None,
"substrate": "control_plane_db",
}
terminal_pr = int(terminal["terminal_pr"]) if terminal else None
skipped: list[SkipRecord] = []
ordered = sort_candidates(list(candidates))
selected: WorkCandidate | None = None
for c in ordered:
reason = classify_skip(c, role=role_norm, terminal_pr=terminal_pr)
if reason:
skipped.append(SkipRecord(c.kind, c.number, reason))
continue
selected = c
break
if selected is None:
# If terminal lock blocks all review work, surface that explicitly.
if terminal_pr is not None and role_norm in (ROLE_REVIEWER, ROLE_MERGER):
outcome = OUTCOME_BLOCKED_TERMINAL
reasons = [
f"no safe work for role '{role_norm}': active terminal-review "
f"lock on PR #{terminal_pr} (resolve terminal path first, #332/#600)"
]
else:
outcome = OUTCOME_NO_SAFE
reasons = [
f"no safe assignable work for role '{role_norm}' "
f"among {len(ordered)} candidates"
]
return {
"success": True,
"outcome": outcome,
"apply": bool(apply),
"role": role_norm,
"profile_name": profile_name,
"username": username,
"session_id": session_id,
"remote": remote,
"org": org,
"repo": repo,
"selected": None,
"expected_role_next": None,
"reasons": reasons,
"skipped": [s.as_dict() for s in skipped],
"terminal_pr": terminal_pr,
"assignment": None,
"substrate": "control_plane_db",
"file_lock_only": False,
"comment_lease_only": False,
"downstream_note": (
"#612 incident bridge remains downstream of #600; "
"allocator never assigns raw monitoring incidents"
),
}
expected_role = expected_role_for_candidate(selected)
allowed, forbidden = role_actions(role_norm)
selection = {
"kind": selected.kind,
"number": selected.number,
"title": selected.title,
"labels": list(selected.labels),
"head_sha": selected.head_sha,
"priority": selected.priority,
"expected_role_next": expected_role,
"reason_selected": (
f"highest-priority candidate for role '{role_norm}' "
f"(expected_role={expected_role})"
),
}
if not apply:
return {
"success": True,
"outcome": OUTCOME_PREVIEW,
"apply": False,
"role": role_norm,
"profile_name": profile_name,
"username": username,
"session_id": session_id,
"remote": remote,
"org": org,
"repo": repo,
"selected": selection,
"expected_role_next": expected_role,
"reasons": [
"dry-run only (apply=false); no assignment/lease created — "
"call again with apply=true to reserve via control-plane DB"
],
"skipped": [s.as_dict() for s in skipped],
"terminal_pr": terminal_pr,
"assignment": None,
"substrate": "control_plane_db",
"file_lock_only": False,
"comment_lease_only": False,
"downstream_note": (
"#612 incident bridge remains downstream of #600; "
"allocator never assigns raw monitoring incidents"
),
}
# Atomic reserve via #613 substrate.
ttl = lease_ttl_seconds if lease_ttl_seconds is not None else None
try:
kwargs: dict[str, Any] = {
"session_id": session_id,
"role": role_norm,
"remote": remote,
"org": org,
"repo": repo,
"kind": selected.kind,
"number": selected.number,
"expected_head_sha": selected.head_sha,
"allowed_actions": allowed,
"forbidden_actions": forbidden,
"phase": "allocated",
}
if ttl is not None:
kwargs["lease_ttl_seconds"] = int(ttl)
result = db.assign_and_lease(**kwargs)
except (InvalidWorkKindError, LeaseRequiredError, ControlPlaneError) as exc:
return {
"success": False,
"outcome": OUTCOME_NO_SAFE,
"apply": True,
"role": role_norm,
"session_id": session_id,
"remote": remote,
"org": org,
"repo": repo,
"selected": selection,
"expected_role_next": expected_role,
"reasons": [f"atomic assign+lease failed: {exc} (fail closed, #613)"],
"skipped": [s.as_dict() for s in skipped],
"terminal_pr": terminal_pr,
"assignment": None,
"substrate": "control_plane_db",
"file_lock_only": False,
"comment_lease_only": False,
}
if result.outcome == "wait":
return {
"success": True,
"outcome": OUTCOME_WAIT,
"apply": True,
"role": role_norm,
"session_id": session_id,
"remote": remote,
"org": org,
"repo": repo,
"selected": selection,
"expected_role_next": expected_role,
"reasons": [result.reason or "foreign active lease"],
"skipped": [s.as_dict() for s in skipped],
"terminal_pr": terminal_pr,
"assignment": result.as_dict(),
"owner_session_id": result.owner_session_id,
"substrate": "control_plane_db",
"file_lock_only": False,
"comment_lease_only": False,
}
if result.outcome == "no_safe_work":
return {
"success": True,
"outcome": OUTCOME_NO_SAFE,
"apply": True,
"role": role_norm,
"session_id": session_id,
"remote": remote,
"org": org,
"repo": repo,
"selected": selection,
"expected_role_next": expected_role,
"reasons": [result.reason or "no_safe_work"],
"skipped": [s.as_dict() for s in skipped],
"terminal_pr": terminal_pr,
"assignment": result.as_dict(),
"substrate": "control_plane_db",
"file_lock_only": False,
"comment_lease_only": False,
}
# assigned
return {
"success": True,
"outcome": OUTCOME_ASSIGNED,
"apply": True,
"role": role_norm,
"profile_name": profile_name,
"username": username,
"session_id": session_id,
"remote": remote,
"org": org,
"repo": repo,
"selected": selection,
"expected_role_next": expected_role,
"reasons": [
selection["reason_selected"],
result.reason or "atomic assign+lease created",
],
"skipped": [s.as_dict() for s in skipped],
"terminal_pr": terminal_pr,
"assignment": result.as_dict(),
"lease_proof": {
"assignment_id": result.assignment_id,
"lease_id": result.lease_id,
"expires_at": result.expires_at,
"expected_head_sha": result.expected_head_sha,
"allowed_actions": list(result.allowed_actions),
"forbidden_actions": list(result.forbidden_actions),
"source": "control_plane_db.assign_and_lease",
},
"next_valid_command": _next_command(role_norm, selected),
"substrate": "control_plane_db",
"file_lock_only": False,
"comment_lease_only": False,
"downstream_note": (
"#612 incident bridge remains downstream of #600; "
"allocator never assigns raw monitoring incidents"
),
}
def _next_command(role: str, c: WorkCandidate) -> str:
if role == ROLE_AUTHOR and c.kind == "issue":
return f"implement issue #{c.number} under a branches/ worktree; open PR when ready"
if role == ROLE_AUTHOR and c.kind == "pr":
return (
f"address REQUEST_CHANGES on PR #{c.number} at head "
f"{(c.head_sha or '')[:12]} and push fixes"
)
if role == ROLE_REVIEWER:
return (
f"review PR #{c.number} pinned at head {(c.head_sha or '')[:12]} "
"via full reviewer workflow"
)
if role == ROLE_MERGER:
return (
f"merge PR #{c.number} only with explicit operator MERGE "
f"authorization at head {(c.head_sha or '')[:12]}"
)
if role == ROLE_RECONCILER:
return f"diagnose contested state for {c.kind}#{c.number}"
return f"proceed on {c.kind}#{c.number} under role {role}"
def candidate_from_dict(data: dict[str, Any]) -> WorkCandidate:
"""Build a WorkCandidate from a plain dict (tests / MCP inventory)."""
return WorkCandidate(
kind=str(data.get("kind") or "issue"),
number=int(data["number"]),
state=str(data.get("state") or "open"),
labels=tuple(data.get("labels") or ()),
title=str(data.get("title") or ""),
priority=int(data.get("priority") or 0),
head_sha=data.get("head_sha"),
request_changes_current_head=bool(data.get("request_changes_current_head")),
approval_on_current_head=bool(data.get("approval_on_current_head")),
approval_stale=bool(data.get("approval_stale")),
approval_contaminated=bool(data.get("approval_contaminated")),
mergeable=bool(data.get("mergeable")),
blocked=bool(data.get("blocked")),
dependency_unmet=bool(data.get("dependency_unmet")),
dependency_reason=data.get("dependency_reason"),
already_claimed_elsewhere=bool(data.get("already_claimed_elsewhere")),
)
@@ -40,9 +40,19 @@
- **#600** must call this substrate (not file locks / comment-only leases alone) for completion.
- **#612** must write `incident_links` here and create **Gitea issues**; the allocator assigns those issues, not raw incidents.
## Allocator API (#600)
Module: `allocator_service.py` · MCP tool: `gitea_allocate_next_work`
- Workers call `gitea_allocate_next_work(apply=false|true)` instead of self-selecting work.
- `apply=false` returns a dry-run selection (`outcome=preview`) with skip reasons.
- `apply=true` reserves via `ControlPlaneDB.assign_and_lease` (atomic assignment+lease).
- Coordination source is **always** the control-plane DB — not file locks or comment-only leases.
- Routing follows ADR §5.3 (REQUEST_CHANGES → author, terminal path first, foreign lease → wait).
- Raw monitoring incidents are never candidates (#612 remains downstream).
## Non-goals (intentionally deferred)
- Full `gitea_allocate_next_work` routing policy (REQUEST_CHANGES → author, etc.) — **#600**
- Sentry/GlitchTip provider adapters and auto-watchdog — **#612**
- Gitea comment/label mirror writers (may be added by allocator tooling later)
+307
View File
@@ -21,6 +21,7 @@ import json
import functools
import contextlib
import subprocess
import uuid
from datetime import datetime, timedelta, timezone
@@ -832,6 +833,8 @@ import review_workflow_boundary # noqa: E402
import review_workflow_load # noqa: E402
import mcp_session_state # noqa: E402
import stale_review_decision_lock # noqa: E402
import allocator_service # noqa: E402
import control_plane_db # noqa: E402
import agent_temp_artifacts
import issue_lock_worktree # noqa: E402
import issue_lock_provenance # noqa: E402
@@ -11109,6 +11112,310 @@ def gitea_capability_stop_terminal_report() -> dict:
})
def _control_plane_db_or_error() -> tuple[Any | None, list[str]]:
"""Open the #613 control-plane DB substrate; fail closed on errors."""
try:
db = control_plane_db.ControlPlaneDB()
return db, []
except Exception as exc: # noqa: BLE001
return None, [
f"control-plane DB substrate unavailable: {_redact(str(exc))} "
"(fail closed, #613/#600)"
]
def _allocator_candidates_from_gitea(
*,
remote: str,
host: str | None,
org: str,
repo: str,
include_issues: bool = True,
include_prs: bool = True,
limit: int = 50,
) -> tuple[list[Any], list[str]]:
"""Build allocator candidates from live Gitea open issues/PRs."""
reasons: list[str] = []
candidates: list[Any] = []
try:
h, o, r = _resolve(remote, host, org, repo)
auth = _auth(h)
except Exception as exc: # noqa: BLE001
return [], [f"failed to resolve Gitea target: {_redact(str(exc))}"]
if include_prs:
try:
prs = api_get_all(
f"{repo_api_url(h, o, r)}/pulls?state=open", auth
) or []
except Exception as exc: # noqa: BLE001
reasons.append(f"failed to list open PRs: {_redact(str(exc))}")
prs = []
for pr in prs[: max(1, int(limit))]:
if not isinstance(pr, dict):
continue
number = pr.get("number")
if number is None:
continue
head = pr.get("head") if isinstance(pr.get("head"), dict) else {}
head_sha = head.get("sha") or pr.get("head_sha")
labels = []
for lab in pr.get("labels") or []:
if isinstance(lab, dict) and lab.get("name"):
labels.append(str(lab["name"]))
elif isinstance(lab, str):
labels.append(lab)
# Lightweight review signals (best-effort; fail soft into reviewer path).
rc_current = False
approval_current = False
approval_stale = False
try:
feedback = gitea_get_pr_review_feedback(
int(number), remote=remote, org=o, repo=r
)
if feedback.get("success"):
rc_current = bool(
feedback.get("has_blocking_change_requests")
and not feedback.get("review_feedback_stale")
)
# Stale RC means author pushed; reviewer still next.
if feedback.get("has_blocking_change_requests") and feedback.get(
"review_feedback_stale"
):
approval_stale = False
except Exception:
pass
mergeable = bool(pr.get("mergeable"))
try:
candidates.append(
allocator_service.WorkCandidate(
kind="pr",
number=int(number),
state="open",
labels=tuple(labels),
title=str(pr.get("title") or ""),
priority=10 if rc_current else 5,
head_sha=head_sha,
request_changes_current_head=rc_current,
approval_on_current_head=approval_current,
approval_stale=approval_stale,
mergeable=mergeable,
)
)
except Exception as exc: # noqa: BLE001
reasons.append(
f"skipped invalid PR candidate #{number}: {_redact(str(exc))}"
)
if include_issues:
try:
issues = api_get_all(
f"{repo_api_url(h, o, r)}/issues?state=open&type=issues", auth
) or []
except Exception as exc: # noqa: BLE001
# Fallback without type filter
try:
issues = api_get_all(
f"{repo_api_url(h, o, r)}/issues?state=open", auth
) or []
except Exception as exc2: # noqa: BLE001
reasons.append(
f"failed to list open issues: {_redact(str(exc2))}"
)
issues = []
for issue in issues[: max(1, int(limit))]:
if not isinstance(issue, dict):
continue
# Pull requests also appear in /issues on Gitea — skip them.
if issue.get("pull_request") is not None:
continue
number = issue.get("number")
if number is None:
continue
labels = []
for lab in issue.get("labels") or []:
if isinstance(lab, dict) and lab.get("name"):
labels.append(str(lab["name"]).lower())
elif isinstance(lab, str):
labels.append(lab.lower())
body = str(issue.get("body") or "")
title = str(issue.get("title") or "")
blocked = "status:blocked" in labels
# Explicit downstream dependency: #612 waits on #600 allocator.
dep_unmet = False
dep_reason = None
if int(number) == 612:
dep_unmet = True
dep_reason = (
"issue#612 (incident bridge) remains downstream of #600 "
"allocator and must not be allocated until #600 is complete"
)
# Generic body markers for blocked-on unfinished deps.
lower_body = body.lower()
if "blocked on #600" in lower_body or "downstream of #600" in lower_body:
dep_unmet = True
dep_reason = dep_reason or (
f"issue#{number} body marks dependency on #600"
)
try:
candidates.append(
allocator_service.WorkCandidate(
kind="issue",
number=int(number),
state="open",
labels=tuple(labels),
title=title,
priority=20 if "status:ready" in labels else 1,
blocked=blocked,
dependency_unmet=dep_unmet,
dependency_reason=dep_reason,
)
)
except Exception as exc: # noqa: BLE001
reasons.append(
f"skipped invalid issue candidate #{number}: {_redact(str(exc))}"
)
return candidates, reasons
@mcp.tool()
def gitea_allocate_next_work(
apply: bool = False,
role: str | None = None,
session_id: str | None = None,
remote: str = "dadeschools",
host: str | None = None,
org: str | None = None,
repo: str | None = None,
include_issues: bool = True,
include_prs: bool = True,
candidates_json: str | None = None,
limit: int = 50,
) -> dict:
"""Controller-owned next-work allocator using the #613 control-plane DB (#600).
Workers must not self-select exclusive work under the standard multi-LLM
workflow. Call this tool instead.
*apply=false* (default): dry-run selection only no assignment/lease.
*apply=true*: atomically assign + lease the selected Gitea issue/PR via
``ControlPlaneDB.assign_and_lease`` (never file locks or comment-only
leases as the coordination source).
Outcomes include: ``assigned_work``, ``preview``, ``wait``,
``blocked_by_terminal_path``, ``no_safe_work``, ``role_ineligible``.
*candidates_json* may inject a JSON list of candidate dicts (tests /
controller overrides). When omitted, open issues/PRs are loaded from Gitea.
#612 remains downstream: raw monitoring incidents are never candidates.
"""
read_block = _profile_operation_gate("gitea.read")
if read_block:
return {
"success": False,
"outcome": allocator_service.OUTCOME_NO_SAFE,
"apply": bool(apply),
"reasons": read_block,
"permission_report": _permission_block_report("gitea.read"),
"assignment": None,
"substrate": "control_plane_db",
"file_lock_only": False,
"comment_lease_only": False,
}
try:
h, o, r = _resolve(remote, host, org, repo)
except ValueError as exc:
return {
"success": False,
"outcome": allocator_service.OUTCOME_NO_SAFE,
"reasons": [str(exc)],
"assignment": None,
"substrate": "control_plane_db",
}
profile = get_profile()
profile_name = (profile.get("profile_name") or "").strip() or None
active_role = _profile_role_kind(profile)
role_in = (role or active_role or "").strip() or "author"
try:
username = _authenticated_username(h)
except Exception:
username = None
db, db_errs = _control_plane_db_or_error()
if db is None:
return {
"success": False,
"outcome": allocator_service.OUTCOME_NO_SAFE,
"apply": bool(apply),
"reasons": db_errs,
"assignment": None,
"substrate": "control_plane_db",
"file_lock_only": False,
"comment_lease_only": False,
}
inv_reasons: list[str] = []
candidates: list[Any] = []
if candidates_json:
try:
raw = json.loads(candidates_json)
if not isinstance(raw, list):
raise ValueError("candidates_json must be a JSON list")
for item in raw:
if not isinstance(item, dict):
continue
candidates.append(allocator_service.candidate_from_dict(item))
except Exception as exc: # noqa: BLE001
return {
"success": False,
"outcome": allocator_service.OUTCOME_NO_SAFE,
"apply": bool(apply),
"reasons": [
f"invalid candidates_json: {_redact(str(exc))} (fail closed)"
],
"assignment": None,
"substrate": "control_plane_db",
}
else:
candidates, inv_reasons = _allocator_candidates_from_gitea(
remote=remote,
host=host,
org=o,
repo=r,
include_issues=include_issues,
include_prs=include_prs,
limit=limit,
)
sid = (session_id or "").strip() or (
f"{profile_name or 'session'}-{os.getpid()}-{uuid.uuid4().hex[:8]}"
)
result = allocator_service.allocate_next_work(
db,
session_id=sid,
role=role_in,
remote=remote if remote in REMOTES else remote,
org=o,
repo=r,
candidates=candidates,
apply=bool(apply),
profile_name=profile_name,
username=username,
)
if inv_reasons:
result.setdefault("inventory_warnings", inv_reasons)
result["candidate_count"] = len(candidates)
result["inventory_source"] = (
"candidates_json" if candidates_json else "gitea_live"
)
return result
# ── Entry point ───────────────────────────────────────────────────────────────
if __name__ == "__main__":
+11
View File
@@ -139,6 +139,17 @@ TASK_CAPABILITY_MAP: dict[str, dict[str, str]] = {
"permission": "gitea.pr.create",
"role": "author",
},
# #600: controller-owned allocator — any authenticated profile may call;
# routing enforces role match to selected work. Uses control-plane DB (#613).
"allocate_next_work": {
"permission": "gitea.read",
"role": "author",
},
"gitea_allocate_next_work": {
"permission": "gitea.read",
"role": "author",
},
"reconcile_landed_pr": {
"permission": "gitea.read",
"role": "author",
+366
View File
@@ -0,0 +1,366 @@
"""Tests for controller-owned allocator (#600) on control-plane DB (#613)."""
from __future__ import annotations
import os
import tempfile
import threading
import unittest
from concurrent.futures import ThreadPoolExecutor, as_completed
from allocator_service import (
OUTCOME_ASSIGNED,
OUTCOME_BLOCKED_TERMINAL,
OUTCOME_NO_SAFE,
OUTCOME_PREVIEW,
OUTCOME_WAIT,
WorkCandidate,
allocate_next_work,
candidate_from_dict,
classify_skip,
expected_role_for_candidate,
)
from control_plane_db import ControlPlaneDB, InvalidWorkKindError
class AllocatorServiceTest(unittest.TestCase):
def setUp(self) -> None:
self._tmp = tempfile.TemporaryDirectory()
self.db_path = os.path.join(self._tmp.name, "cp.sqlite3")
self.db = ControlPlaneDB(self.db_path)
def tearDown(self) -> None:
self._tmp.cleanup()
def _alloc(self, **kwargs):
defaults = dict(
db=self.db,
session_id="s-test",
role="author",
remote="prgs",
org="org",
repo="repo",
candidates=[],
apply=False,
profile_name="prgs-author",
username="jcwalker3",
)
defaults.update(kwargs)
return allocate_next_work(**defaults)
def test_selects_ready_issue_for_author(self) -> None:
cands = [
WorkCandidate(
kind="issue",
number=612,
labels=("status:ready",),
title="bridge",
dependency_unmet=True,
dependency_reason="downstream of #600",
),
WorkCandidate(
kind="issue",
number=600,
labels=("status:ready", "type:feature"),
title="allocator",
priority=50,
),
WorkCandidate(
kind="issue",
number=601,
labels=("status:blocked",),
title="blocked",
blocked=True,
),
]
res = self._alloc(candidates=cands, apply=False)
self.assertTrue(res["success"])
self.assertEqual(res["outcome"], OUTCOME_PREVIEW)
self.assertEqual(res["selected"]["number"], 600)
self.assertEqual(res["selected"]["kind"], "issue")
# When 600 is highest priority valid work, lower-priority blocked/dep
# candidates are not visited. Prove they are skipped when they sort first.
res2 = self._alloc(
candidates=[
WorkCandidate(
kind="issue",
number=612,
labels=("status:ready",),
priority=99,
dependency_unmet=True,
dependency_reason="downstream of #600",
),
WorkCandidate(
kind="issue",
number=601,
labels=("status:blocked",),
priority=98,
blocked=True,
),
WorkCandidate(
kind="issue",
number=600,
labels=("status:ready",),
priority=1,
),
],
apply=False,
)
skipped_nums = {s["number"] for s in res2["skipped"]}
self.assertIn(612, skipped_nums)
self.assertIn(601, skipped_nums)
self.assertEqual(res2["selected"]["number"], 600)
self.assertIsNone(res["assignment"])
self.assertFalse(res["file_lock_only"])
self.assertFalse(res["comment_lease_only"])
def test_atomic_assign_and_lease_on_apply(self) -> None:
cands = [
WorkCandidate(
kind="issue",
number=600,
labels=("status:ready",),
priority=10,
)
]
res = self._alloc(candidates=cands, apply=True, session_id="s-a")
self.assertEqual(res["outcome"], OUTCOME_ASSIGNED)
asn = res["assignment"]
self.assertEqual(asn["outcome"], "assigned")
self.assertIsNotNone(asn["assignment_id"])
self.assertIsNotNone(asn["lease_id"])
self.assertEqual(asn["work_number"], 600)
self.assertIn("implement", asn["allowed_actions"])
self.assertIn("merge", asn["forbidden_actions"])
proof = res["lease_proof"]
self.assertEqual(proof["source"], "control_plane_db.assign_and_lease")
def test_concurrent_allocators_no_double_assign(self) -> None:
cand = WorkCandidate(
kind="pr",
number=100,
head_sha="a" * 40,
priority=10,
)
# Seed work item so both race the same target
self.db.upsert_work_item(
remote="prgs",
org="org",
repo="repo",
kind="pr",
number=100,
current_head_sha="a" * 40,
)
results = []
lock = threading.Lock()
def worker(sid: str):
r = allocate_next_work(
self.db,
session_id=sid,
role="reviewer",
remote="prgs",
org="org",
repo="repo",
candidates=[cand],
apply=True,
profile_name="prgs-reviewer",
)
with lock:
results.append(r)
with ThreadPoolExecutor(max_workers=2) as pool:
futs = [pool.submit(worker, f"s-{i}") for i in range(2)]
for f in as_completed(futs):
f.result()
outcomes = [r["outcome"] for r in results]
self.assertEqual(outcomes.count(OUTCOME_ASSIGNED), 1, outcomes)
self.assertEqual(outcomes.count(OUTCOME_WAIT), 1, outcomes)
def test_blocked_and_dependency_skipped(self) -> None:
cands = [
WorkCandidate(kind="issue", number=1, blocked=True, labels=("status:blocked",)),
WorkCandidate(
kind="issue",
number=612,
labels=("status:ready",),
dependency_unmet=True,
dependency_reason="downstream of #600",
),
]
res = self._alloc(candidates=cands, apply=True, role="author")
self.assertEqual(res["outcome"], OUTCOME_NO_SAFE)
self.assertIsNone(res["selected"])
reasons = " ".join(s["reason"] for s in res["skipped"])
self.assertIn("blocked", reasons.lower())
self.assertIn("600", reasons)
def test_already_leased_returns_wait(self) -> None:
self.db.upsert_session(session_id="owner", role="author")
self.db.assign_and_lease(
session_id="owner",
role="author",
remote="prgs",
org="org",
repo="repo",
kind="issue",
number=50,
)
res = self._alloc(
session_id="other",
candidates=[
WorkCandidate(kind="issue", number=50, labels=("status:ready",))
],
apply=True,
)
self.assertEqual(res["outcome"], OUTCOME_WAIT)
self.assertEqual(res["owner_session_id"], "owner")
def test_role_ineligible_skipped(self) -> None:
# PR with current-head REQUEST_CHANGES expects author, not reviewer.
cand = WorkCandidate(
kind="pr",
number=9,
head_sha="b" * 40,
request_changes_current_head=True,
)
self.assertEqual(expected_role_for_candidate(cand), "author")
res = self._alloc(
role="reviewer",
profile_name="prgs-reviewer",
candidates=[cand],
apply=False,
)
self.assertEqual(res["outcome"], OUTCOME_NO_SAFE)
self.assertTrue(any("expects role" in s["reason"] for s in res["skipped"]))
def test_terminal_lock_blocks_downstream_reviewer_prs(self) -> None:
self.db.set_terminal_lock(
remote="prgs",
org="org",
repo="repo",
terminal_pr=10,
decision="request_changes",
status="active",
)
cands = [
WorkCandidate(kind="pr", number=11, head_sha="c" * 40, priority=5),
WorkCandidate(kind="pr", number=10, head_sha="d" * 40, priority=1),
]
res = self._alloc(
role="reviewer",
profile_name="prgs-reviewer",
candidates=cands,
apply=False,
)
# Terminal PR #10 is selectable; #11 skipped for terminal path.
self.assertEqual(res["selected"]["number"], 10)
self.assertTrue(
any(
s["number"] == 11 and "terminal-review lock" in s["reason"]
for s in res["skipped"]
)
)
def test_terminal_lock_blocks_all_when_only_downstream(self) -> None:
self.db.set_terminal_lock(
remote="prgs",
org="org",
repo="repo",
terminal_pr=10,
decision="request_changes",
status="active",
)
res = self._alloc(
role="reviewer",
profile_name="prgs-reviewer",
candidates=[
WorkCandidate(kind="pr", number=99, head_sha="e" * 40),
],
apply=False,
)
self.assertEqual(res["outcome"], OUTCOME_BLOCKED_TERMINAL)
def test_no_work_structured(self) -> None:
res = self._alloc(candidates=[], apply=True)
self.assertEqual(res["outcome"], OUTCOME_NO_SAFE)
self.assertIsNone(res["selected"])
self.assertTrue(res["reasons"])
def test_rejects_incident_kind(self) -> None:
with self.assertRaises(InvalidWorkKindError):
WorkCandidate(kind="sentry_incident", number=1)
def test_612_downstream_marker_in_result(self) -> None:
res = self._alloc(
candidates=[
WorkCandidate(kind="issue", number=600, labels=("status:ready",))
],
apply=True,
)
self.assertIn("612", res.get("downstream_note", ""))
def test_substrate_not_file_or_comment_lease(self) -> None:
res = self._alloc(
candidates=[
WorkCandidate(kind="issue", number=1, labels=("status:ready",))
],
apply=True,
)
self.assertEqual(res["substrate"], "control_plane_db")
self.assertFalse(res["file_lock_only"])
self.assertFalse(res["comment_lease_only"])
self.assertEqual(
res["lease_proof"]["source"], "control_plane_db.assign_and_lease"
)
def test_candidate_from_dict(self) -> None:
c = candidate_from_dict(
{
"kind": "pr",
"number": 7,
"head_sha": "f" * 40,
"approval_on_current_head": True,
"mergeable": True,
}
)
self.assertEqual(expected_role_for_candidate(c), "merger")
def test_merger_gets_clean_approval(self) -> None:
c = WorkCandidate(
kind="pr",
number=3,
head_sha="1" * 40,
approval_on_current_head=True,
mergeable=True,
priority=100,
)
res = self._alloc(
role="merger",
profile_name="prgs-merger",
candidates=[c],
apply=True,
session_id="merger-1",
)
self.assertEqual(res["outcome"], OUTCOME_ASSIGNED)
self.assertEqual(res["assignment"]["work_number"], 3)
self.assertIn("merge", res["assignment"]["allowed_actions"])
def test_db_unavailable_fails_closed(self) -> None:
res = allocate_next_work(
None, # type: ignore[arg-type]
session_id="x",
role="author",
remote="prgs",
org="o",
repo="r",
candidates=[],
apply=True,
)
self.assertFalse(res["success"])
self.assertIn("unavailable", res["reasons"][0].lower())
if __name__ == "__main__":
unittest.main()