#!/usr/bin/env python3
"""verify_sol_profile_triage.py -- re-derives the classification of
/root/.hermes/profiles/sol from the live disk and compares it against the
committed artifact. Exit code 0 = PASS, nonzero = FAIL."""
import json
import os
import sys

PROFILE = "/root/.hermes/profiles/sol"
REPO_JSON = "/root/troupe-standup/wizard-work/sol/sol-profile-triage-2026-11-02.json"
WIZ_JSON = "/root/.hermes/wizard-outputs/sol/sol-profile-triage-2026-11-02.json"

EXACT = {
    "identity": {".env", ".env.bak-20261014T165715Z"},
    "soul": {"SOUL.md"},
    "config": {".env.lock", "config.yaml", "config.yaml.bak-20261014T165715Z"},
    "state_db": {
        "state.db", "state.db-shm", "state.db-wal",
        "state.db.auto-maintenance.lock", "state.db.fts_rebuild.lock",
        "state.db.quarantine.lock",
        "auth.json", "auth.lock", ".skills_prompt_snapshot.json",
        "models_dev_cache.json", "models_dev_cache.etag",
    },
    "cache": {"cache", "audio_cache", "image_cache"},
    "sessions": {"sessions"},
    "skills": {"skills"},
    "plugins": {"plugins", "plugin-update-checks"},
    "logs": {"logs"},
    "backups": {"backups"},
    "vault": {"vault"},
}


def classify(name, is_dir):
    for cat, names in EXACT.items():
        if name in names:
            return cat
    if is_dir:
        return "extension_dir"
    if name.endswith(".py"):
        return "script"
    if name.endswith((".yaml", ".json")):
        return "config_data"
    return "other"


def expected():
    exp = {}
    for name in os.listdir(PROFILE):
        p = os.path.join(PROFILE, name)
        cat = classify(name, os.path.isdir(p))
        exp[cat] = exp.get(cat, 0) + 1
    return exp


def load(path):
    with open(path, encoding="utf-8") as fh:
        return json.load(fh)


def main():
    fails = []
    for p in (REPO_JSON, WIZ_JSON):
        if not os.path.isfile(p):
            fails.append("missing artifact: " + p)
    if not fails:
        disk_cats = expected()
        for label, jpath in (("repo", REPO_JSON), ("wizard-outputs", WIZ_JSON)):
            d = load(jpath)
            if d.get("total_entries") != len(os.listdir(PROFILE)):
                fails.append("[" + label + "] total_entries mismatch")
            if d.get("categories") != disk_cats:
                fails.append("[" + label + "] categories mismatch (got %s, want %s)"
                             % (d.get("categories"), disk_cats))
            for e in d.get("entries", []):
                if not os.path.lexists(e.get("path", "")):
                    fails.append("[" + label + "] referenced path no longer exists: "
                                 + str(e.get("path")))
            for e in d.get("entries", []):
                if e.get("category") != classify(e["name"], e.get("type") == "dir"):
                    fails.append("[" + label + "] entry category mismatch: " + e["name"])
    print("PASS" if not fails else "FAIL")
    for f in fails:
        print(" -", f)
    sys.exit(1 if fails else 0)



if __name__ == "__main__":
    main()