Cache repo hash on server and reuse in updater

Data/Agent/Roles/role_DeviceAudit.py

@@ -9,6 +9,7 @@ import shutil
 import string
 import asyncio
 from pathlib import Path
+from typing import Optional
 
 try:
     import psutil  # type: ignore
@@ -134,24 +135,132 @@ def _project_root():
         return os.getcwd()
 
 
-_AGENT_HASH_CACHE = {"path": None, "mtime": None, "value": None}
+_AGENT_HASH_CACHE = {
+    "path": None,
+    "mtime": None,
+    "value": None,
+    "source": None,
+    "extra": None,
+}
+
+
+def _iter_hash_roots():
+    seen = set()
+    root = _project_root()
+    for _ in range(6):
+        if not root or root in seen:
+            break
+        yield root
+        seen.add(root)
+        parent = os.path.dirname(root)
+        if not parent or parent == root:
+            break
+        root = parent
+
+
+def _resolve_git_head_hash(root: str) -> Optional[str]:
+    git_dir = os.path.join(root, ".git")
+    head_path = os.path.join(git_dir, "HEAD")
+    if not os.path.isfile(head_path):
+        return None
+    try:
+        with open(head_path, "r", encoding="utf-8") as fh:
+            head = fh.read().strip()
+    except Exception:
+        return None
+    if not head:
+        return None
+    if head.startswith("ref:"):
+        ref = head.split(" ", 1)[1].strip() if " " in head else head.split(":", 1)[1].strip()
+        if not ref:
+            return None
+        ref_path = os.path.join(git_dir, *ref.split("/"))
+        if os.path.isfile(ref_path):
+            try:
+                with open(ref_path, "r", encoding="utf-8") as rf:
+                    commit = rf.read().strip()
+                    return commit or None
+            except Exception:
+                return None
+        packed_refs = os.path.join(git_dir, "packed-refs")
+        if os.path.isfile(packed_refs):
+            try:
+                with open(packed_refs, "r", encoding="utf-8") as pf:
+                    for line in pf:
+                        line = line.strip()
+                        if not line or line.startswith("#") or line.startswith("^"):
+                            continue
+                        try:
+                            commit, ref_name = line.split(" ", 1)
+                        except ValueError:
+                            continue
+                        if ref_name.strip() == ref:
+                            commit = commit.strip()
+                            return commit or None
+            except Exception:
+                return None
+        return None
+    # Detached head contains the commit hash directly
+    commit = head.splitlines()[0].strip()
+    return commit or None
 
 
 def _read_agent_hash():
     try:
-        root = _project_root()
-        path = os.path.join(root, 'github_repo_hash.txt')
         cache = _AGENT_HASH_CACHE
-        if not os.path.isfile(path):
-            cache.update({"path": path, "mtime": None, "value": None})
-            return None
-        mtime = os.path.getmtime(path)
-        if cache.get("path") == path and cache.get("mtime") == mtime:
+        for root in _iter_hash_roots():
+            path = os.path.join(root, 'github_repo_hash.txt')
+            if not os.path.isfile(path):
+                continue
+            mtime = os.path.getmtime(path)
+            if (
+                cache.get("source") == "file"
+                and cache.get("path") == path
+                and cache.get("mtime") == mtime
+            ):
+                return cache.get("value")
+            with open(path, 'r', encoding='utf-8') as fh:
+                value = fh.read().strip()
+            cache.update(
+                {
+                    "source": "file",
+                    "path": path,
+                    "mtime": mtime,
+                    "extra": None,
+                    "value": value or None,
+                }
+            )
             return cache.get("value")
-        with open(path, 'r', encoding='utf-8') as fh:
-            value = fh.read().strip()
-        cache.update({"path": path, "mtime": mtime, "value": value or None})
-        return cache.get("value")
+
+        for root in _iter_hash_roots():
+            git_dir = os.path.join(root, '.git')
+            head_path = os.path.join(git_dir, 'HEAD')
+            if not os.path.isfile(head_path):
+                continue
+            head_mtime = os.path.getmtime(head_path)
+            packed_path = os.path.join(git_dir, 'packed-refs')
+            packed_mtime = os.path.getmtime(packed_path) if os.path.isfile(packed_path) else None
+            if (
+                cache.get("source") == "git"
+                and cache.get("path") == head_path
+                and cache.get("mtime") == head_mtime
+                and cache.get("extra") == packed_mtime
+            ):
+                return cache.get("value")
+            commit = _resolve_git_head_hash(root)
+            cache.update(
+                {
+                    "source": "git",
+                    "path": head_path,
+                    "mtime": head_mtime,
+                    "extra": packed_mtime,
+                    "value": commit or None,
+                }
+            )
+            if commit:
+                return commit
+        cache.update({"source": None, "path": None, "mtime": None, "extra": None, "value": None})
+        return None
     except Exception:
         try:
             _AGENT_HASH_CACHE.update({"value": None})
@@ -803,6 +912,12 @@ def _build_details_fallback() -> dict:
         'storage': collect_storage(),
         'network': network,
     }
+    try:
+        agent_hash_value = _read_agent_hash()
+        if agent_hash_value:
+            details.setdefault('summary', {})['agent_hash'] = agent_hash_value
+    except Exception:
+        pass
     return details
 
 

Data/Server/WebUI/src/Devices/Device_List.jsx

@@ -126,17 +126,15 @@ export default function DeviceList({ onSelectDevice }) {
 
   const fetchLatestRepoHash = useCallback(async () => {
     try {
-      const resp = await fetch(
-        "https://api.github.com/repos/bunny-lab-io/Borealis/branches/main",
-        {
-          headers: {
-            Accept: "application/vnd.github+json",
-          },
-        }
-      );
-      if (!resp.ok) throw new Error(`GitHub status ${resp.status}`);
+      const params = new URLSearchParams({ repo: "bunny-lab-io/Borealis", branch: "main" });
+      const resp = await fetch(`/api/agent/repo_hash?${params.toString()}`);
       const json = await resp.json();
-      const sha = (json?.commit?.sha || "").trim();
+      const sha = (json?.sha || "").trim();
+      if (!resp.ok || !sha) {
+        const err = new Error(`Latest hash status ${resp.status}${json?.error ? ` - ${json.error}` : ""}`);
+        err.response = json;
+        throw err;
+      }
       setRepoHash((prev) => sha || prev || null);
       return sha || null;
     } catch (err) {
@@ -167,7 +165,7 @@ export default function DeviceList({ onSelectDevice }) {
       const arr = Object.entries(data || {}).map(([id, a]) => {
         const hostname = a.hostname || id || "unknown";
         const details = detailsByHost[hostname] || {};
-        const agentHash = (a.agent_hash || "").trim();
+        const agentHash = (a.agent_hash || details.agentHash || "").trim();
         return {
           id,
           hostname,
@@ -230,43 +228,47 @@ export default function DeviceList({ onSelectDevice }) {
                 const externalIp = summary.external_ip || "";
                 const lastReboot = summary.last_reboot || "";
                 const description = summary.description || "";
+                const agentHashValue = (summary.agent_hash || "").trim();
+                const enriched = {
+                  lastUser,
+                  created: createdRaw,
+                  createdTs,
+                  type: deviceType,
+                  internalIp,
+                  externalIp,
+                  lastReboot,
+                  description,
+                  agentHash: agentHashValue,
+                };
                 setDetailsByHost((prev) => ({
                   ...prev,
-                  [h]: {
-                    lastUser,
-                    created: createdRaw,
-                    createdTs,
-                    type: deviceType,
-                    internalIp,
-                    externalIp,
-                    lastReboot,
-                    description,
-                  },
+                  [h]: enriched,
                 }));
+                setRows((prev) =>
+                  prev.map((r) => {
+                    if (r.hostname !== h) return r;
+                    const nextHash = agentHashValue || r.agentHash;
+                    return {
+                      ...r,
+                      lastUser: enriched.lastUser || r.lastUser,
+                      type: enriched.type || r.type,
+                      created: enriched.created || r.created,
+                      createdTs: enriched.createdTs || r.createdTs,
+                      internalIp: enriched.internalIp || r.internalIp,
+                      externalIp: enriched.externalIp || r.externalIp,
+                      lastReboot: enriched.lastReboot || r.lastReboot,
+                      description: enriched.description || r.description,
+                      agentHash: nextHash,
+                      agentVersion: computeAgentVersion(nextHash, repoSha),
+                    };
+                  })
+                );
               } catch {
                 // ignore per-host failure
               }
             })
           );
         }
-        // After caching, refresh rows to apply newly available details
-        setRows((prev) =>
-          prev.map((r) => {
-            const det = detailsByHost[r.hostname];
-            if (!det) return r;
-            return {
-              ...r,
-              lastUser: det.lastUser || r.lastUser,
-              type: det.type || r.type,
-              created: det.created || r.created,
-              createdTs: det.createdTs || r.createdTs,
-              internalIp: det.internalIp || r.internalIp,
-              externalIp: det.externalIp || r.externalIp,
-              lastReboot: det.lastReboot || r.lastReboot,
-              description: det.description || r.description,
-            };
-          })
-        );
       }
     } catch (e) {
       console.warn("Failed to load agents:", e);

Data/Server/server.py

@@ -21,6 +21,7 @@ from typing import List, Dict, Tuple, Optional, Any, Set
 import sqlite3
 import io
 import uuid
+from threading import Lock
 from datetime import datetime, timezone
 
 try:
@@ -68,6 +69,142 @@ def _write_service_log(service: str, msg: str):
         pass
 
 
+_REPO_HEAD_CACHE: Dict[str, Tuple[str, float]] = {}
+_REPO_HEAD_LOCK = Lock()
+
+_DEFAULT_REPO = os.environ.get('BOREALIS_REPO', 'bunny-lab-io/Borealis')
+_DEFAULT_BRANCH = os.environ.get('BOREALIS_REPO_BRANCH', 'main')
+try:
+    _REPO_HASH_INTERVAL = int(os.environ.get('BOREALIS_REPO_HASH_REFRESH', '60'))
+except ValueError:
+    _REPO_HASH_INTERVAL = 60
+_REPO_HASH_INTERVAL = max(30, min(_REPO_HASH_INTERVAL, 3600))
+_REPO_HASH_WORKER_STARTED = False
+_REPO_HASH_WORKER_LOCK = Lock()
+
+
+def _fetch_repo_head(owner_repo: str, branch: str = 'main', *, ttl_seconds: int = 60, force_refresh: bool = False) -> Dict[str, Any]:
+    """Resolve the latest commit hash for ``owner_repo``/``branch`` via GitHub's REST API.
+
+    The server caches the response so that a fleet of agents can reuse the
+    result without exhausting rate limits. ``ttl_seconds`` bounds how long a
+    cached value is considered fresh. When ``force_refresh`` is True the cache
+    is bypassed and a new request is attempted immediately.
+    """
+
+    key = f"{owner_repo}:{branch}"
+    now = time.time()
+
+    with _REPO_HEAD_LOCK:
+        cached = _REPO_HEAD_CACHE.get(key)
+
+    cached_sha: Optional[str] = None
+    cached_ts: Optional[float] = None
+    cached_age: Optional[float] = None
+    if cached:
+        cached_sha, cached_ts = cached
+        cached_age = max(0.0, now - cached_ts)
+
+    if cached_sha and not force_refresh and cached_age is not None and cached_age < max(30, ttl_seconds):
+        return {
+            'sha': cached_sha,
+            'cached': True,
+            'age_seconds': cached_age,
+            'error': None,
+            'source': 'cache',
+        }
+
+    headers = {
+        'Accept': 'application/vnd.github+json',
+        'User-Agent': 'Borealis-Server'
+    }
+    token = os.environ.get('BOREALIS_GITHUB_TOKEN') or os.environ.get('GITHUB_TOKEN')
+    if token:
+        headers['Authorization'] = f'Bearer {token}'
+
+    error_msg: Optional[str] = None
+    sha: Optional[str] = None
+    try:
+        resp = requests.get(
+            f'https://api.github.com/repos/{owner_repo}/branches/{branch}',
+            headers=headers,
+            timeout=20,
+        )
+        if resp.status_code == 200:
+            data = resp.json()
+            sha = (data.get('commit') or {}).get('sha')
+        else:
+            error_msg = f'GitHub REST API repo head lookup failed: HTTP {resp.status_code} {resp.text[:200]}'
+    except Exception as exc:  # pragma: no cover - defensive logging
+        error_msg = f'GitHub REST API repo head lookup raised: {exc}'
+
+    if sha:
+        sha = sha.strip()
+        with _REPO_HEAD_LOCK:
+            _REPO_HEAD_CACHE[key] = (sha, now)
+        return {
+            'sha': sha,
+            'cached': False,
+            'age_seconds': 0.0,
+            'error': None,
+            'source': 'github',
+        }
+
+    if error_msg:
+        _write_service_log('server', error_msg)
+
+    if cached_sha is not None:
+        return {
+            'sha': cached_sha,
+            'cached': True,
+            'age_seconds': cached_age,
+            'error': error_msg or 'using cached value',
+            'source': 'cache-stale',
+        }
+
+    return {
+        'sha': None,
+        'cached': False,
+        'age_seconds': None,
+        'error': error_msg or 'unable to resolve repository head',
+        'source': 'github',
+    }
+
+
+def _refresh_default_repo_hash(force: bool = False) -> Dict[str, Any]:
+    ttl = max(30, _REPO_HASH_INTERVAL)
+    try:
+        return _fetch_repo_head(_DEFAULT_REPO, _DEFAULT_BRANCH, ttl_seconds=ttl, force_refresh=force)
+    except Exception as exc:  # pragma: no cover - defensive logging
+        _write_service_log('server', f'default repo hash refresh failed: {exc}')
+        raise
+
+
+def _repo_hash_background_worker():
+    interval = max(30, _REPO_HASH_INTERVAL)
+    # Fetch immediately, then sleep between refreshes
+    while True:
+        try:
+            _refresh_default_repo_hash(force=True)
+        except Exception:
+            # _refresh_default_repo_hash already logs details
+            pass
+        eventlet.sleep(interval)
+
+
+def _ensure_repo_hash_worker():
+    global _REPO_HASH_WORKER_STARTED
+    with _REPO_HASH_WORKER_LOCK:
+        if _REPO_HASH_WORKER_STARTED:
+            return
+        _REPO_HASH_WORKER_STARTED = True
+        try:
+            eventlet.spawn_n(_repo_hash_background_worker)
+        except Exception as exc:
+            _REPO_HASH_WORKER_STARTED = False
+            _write_service_log('server', f'failed to start repo hash worker: {exc}')
+
+
 def _ansible_log_server(msg: str):
     _write_service_log('ansible', msg)
 
@@ -126,6 +263,8 @@ socketio = SocketIO(
     }
 )
 
+_ensure_repo_hash_worker()
+
 # ---------------------------------------------
 # Serve ReactJS Production Vite Build from dist/
 # ---------------------------------------------
@@ -147,6 +286,44 @@ def serve_dist(path):
 def health():
     return jsonify({"status": "ok"})
 
+
+@app.route("/api/agent/repo_hash", methods=["GET"])
+def api_agent_repo_hash():
+    try:
+        repo = (request.args.get('repo') or _DEFAULT_REPO).strip()
+        branch = (request.args.get('branch') or _DEFAULT_BRANCH).strip()
+        refresh_flag = (request.args.get('refresh') or '').strip().lower()
+        ttl_raw = request.args.get('ttl')
+        if '/' not in repo:
+            return jsonify({"error": "repo must be in the form owner/name"}), 400
+        try:
+            ttl = int(ttl_raw) if ttl_raw else _REPO_HASH_INTERVAL
+        except ValueError:
+            ttl = _REPO_HASH_INTERVAL
+        ttl = max(30, min(ttl, 3600))
+        force_refresh = refresh_flag in {'1', 'true', 'yes', 'force', 'refresh'}
+        if repo == _DEFAULT_REPO and branch == _DEFAULT_BRANCH:
+            result = _refresh_default_repo_hash(force=force_refresh)
+        else:
+            result = _fetch_repo_head(repo, branch, ttl_seconds=ttl, force_refresh=force_refresh)
+        sha = (result.get('sha') or '').strip()
+        payload = {
+            'repo': repo,
+            'branch': branch,
+            'sha': sha if sha else None,
+            'cached': bool(result.get('cached')),
+            'age_seconds': result.get('age_seconds'),
+            'source': result.get('source'),
+        }
+        if result.get('error'):
+            payload['error'] = result['error']
+        if sha:
+            return jsonify(payload)
+        return jsonify(payload), 503
+    except Exception as exc:
+        _write_service_log('server', f'/api/agent/repo_hash error: {exc}')
+        return jsonify({"error": "internal error"}), 500
+
 # ---------------------------------------------
 # Server Time Endpoint
 # ---------------------------------------------