Self-improving AI Employees for everyone
Python ai-agentsai-employeeai-employeesassistantautonomous-agentsco-employeescompanygrok-botinstinctloop-agentsmuseown-your-datapersonalself-hosted
"""Working dir + metadata for each ``company deploy`` deployment.
A deployment is named (``--name``, default ``opencompany``). The name is its
folder under ``<DATA_DIR>/deploy/`` and its cloud resource id, so one machine
can drive several VMs, one per name.
Pre-rebrand deployments used ``machinaos`` and may live under either
``~/.machina`` or a checkout-local ``.machina`` root. For the default name
those paths are discovered before creating new state so a rebrand can never
orphan a live VM, firewall, bucket, or Terraform state.
The selected directory is BOTH the Terraform working dir (rendered module +
``terraform.tfvars.json`` + local state) and the home of a small
``deploy-meta.json`` (provider + port + owner email) used by ``status`` /
``destroy``.
No module-level side effects -- ``user_data_dir`` (platformdirs) is only
resolved when a function is called.
"""
from __future__ import annotations
import json
import re
from pathlib import Path
from cli.platform_ import project_root, user_data_dir
#: Default deployment/resource name and the pre-rebrand compatibility name.
NAME = "opencompany"
LEGACY_NAME = "machinaos"
#: A deployment name is also a cloud resource id. GCP service-account ids set
#: the tightest rule: 6-30 lowercase letters, digits and hyphens, starting
#: with a letter and not ending with a hyphen.
NAME_PATTERN = re.compile(r"[a-z][a-z0-9-]{4,28}[a-z0-9]")
_META_FILENAME = "deploy-meta.json"
def deploy_root() -> Path:
return user_data_dir() / "deploy"
def _legacy_workdirs() -> tuple[Path, ...]:
"""All locations used by released versions before the rebrand."""
candidates = (
deploy_root() / LEGACY_NAME,
Path.home() / ".machina" / "deploy" / LEGACY_NAME,
project_root() / ".machina" / "deploy" / LEGACY_NAME,
)
# Preserve ordering while removing duplicates (DATA_DIR can point at one
# of the explicit legacy roots above).
return tuple(dict.fromkeys(candidates))
def _has_state(path: Path) -> bool:
return (path / _META_FILENAME).exists() or (path / "terraform.tfstate").exists()
def workdir(name: str = NAME) -> Path:
"""Terraform state location. The default name prefers current state, then legacy state."""
current = deploy_root() / name
if name != NAME or _has_state(current):
return current
for legacy in _legacy_workdirs():
if _has_state(legacy):
return legacy
return current
def resource_name(name: str = NAME) -> str:
"""Cloud resource id to render without replacing legacy resources."""
if name != NAME:
return name
meta = read_meta() or {}
configured = meta.get("resource_name")
if configured in {NAME, LEGACY_NAME}:
return configured
return LEGACY_NAME if workdir().name == LEGACY_NAME else NAME
def meta_file(name: str = NAME) -> Path:
return workdir(name) / _META_FILENAME
def write_meta(meta: dict, name: str = NAME) -> None:
wd = workdir(name)
wd.mkdir(parents=True, exist_ok=True)
meta_file(name).write_text(json.dumps(meta, indent=2), encoding="utf-8")
def read_meta(name: str = NAME) -> dict | None:
mf = meta_file(name)
if not mf.exists():
return None
try:
return json.loads(mf.read_text(encoding="utf-8"))
except (ValueError, OSError):
return None
def exists(name: str = NAME) -> bool:
return meta_file(name).exists()