Python Interface Tree¶

ChatLinux keeps the CLI thin. Substantive behavior lives in importable Python functions so future MCP tools, automation scripts, or other ChatArch packages can reuse it without shelling out.

Package Entry¶

from chatlinux import __version__

Fleet Status APIs¶

from chatlinux.fleet import (
    init_config,
    load_config,
    refresh_track,
    load_cache,
    format_table,
    state_paths,
)

Function	Purpose
`state_paths(home=None, track="cube")`	Return `home`, `config`, `cache_dir`, `runtime_dir`, `inventory`, `probe`, and `cache`; by default all live under the same ChatArch state root.
`init_config(home=None, sample="cube", force=False)`	Initialize `fleet.json`, cache, and runtime directories.
`load_config(home=None)`	Read and validate fleet config.
`refresh_track(home=None, track="cube", runner=subprocess.run)`	Run read-only Ansible refresh, normalize results, and write cache.
`load_cache(home=None, track="cube")`	Read the last cache without contacting hosts.
`format_table(snapshot)`	Render cache JSON as a compact table.
`parse_ansible_output(output, expected_hosts=...)`	Parse Ansible `-o` script output.

Module Layout¶

chatlinux
├── cli.py           # Click entry; argument parsing and output only
├── config.py        # ChatEnv provider
└── fleet.py         # fleet config, Ansible refresh, cache, and rendering logic

Output Contract¶

refresh_track returns a structured dict, writes cache/<track>-status.json, and generates runtime/fleet-inventory.ini plus runtime/fleet_probe.py under the same home.
load_cache reads local cache only and is suitable for fast shell/status checks.
Remote probe failures are preserved under each host's ok=false / error fields.
Public output should not leak tokens, cookies, internal Authorization headers, or sensitive personal data.