feat: async analysis with polling + nested archive extraction

- Backend /api/v1/analyze now returns 202 with analysis_id immediately
- New GET /api/v1/analyze/{id} for polling analysis status
- Background thread handles extraction, disk discovery, OS/EFI detection
- Nested archive extraction: handles chained zips and split zips (.z01-.zNN)
- Frontend polls /session/analyze/status/{id} every 2s until complete
- SCP page now has complete confirm form flow with job polling

backend/app.py

@@ -22,11 +22,12 @@ from models import (
     CleanupResponse,
     AnalyzeRequest,
     AnalyzeResponse,
+    AnalyzeStatusResponse,
     CloneRequest,
     HealthResponse,
     ErrorResponse,
 )
-from converter import extract_if_needed, discover_disk, detect_os, detect_efi
+from converter import extract_if_needed, discover_disk, detect_os, detect_efi, start_analysis, get_analysis_status
 from provisioner import submit_job, get_job_status, cleanup_staging, clone_vm
 
 # ---------------------------------------------------------------------------
@@ -93,39 +94,33 @@ def health() -> HealthResponse:
     return HealthResponse(status="ok")
 
 
-@app.post(f"{API_PREFIX}/analyze", response_model=AnalyzeResponse)
+@app.post(f"{API_PREFIX}/analyze", response_model=AnalyzeStatusResponse, status_code=202)
-def analyze(req: AnalyzeRequest) -> AnalyzeResponse:
+def analyze(req: AnalyzeRequest) -> AnalyzeStatusResponse:
-    """Analyze a source file: extract if archive, find disk, detect OS + EFI."""
+    """Analyze a source file asynchronously: extraction + disk probing + OS detection."""
     _validate_filename(req.source_filename)
     logger.info("Analyze request: vmid=%d file=%s", req.vmid, req.source_filename)
-
     try:
-        source = extract_if_needed(req.source_filename)
+        result = start_analysis(req.vmid, req.source_filename)
-        logger.info("Source ready: %s", source)
+        return AnalyzeStatusResponse(
+            analysis_id=result["analysis_id"],
+            vmid=req.vmid,
+            status=result["status"],
+            message=result["message"],
+        )
     except Exception as exc:
-        logger.error("Extraction failed: %s", exc)
+        logger.error("Analysis start failed: %s", exc)
         raise HTTPException(status_code=400, detail=str(exc))
 
+
+@app.get(f"{API_PREFIX}/analyze/{{analysis_id}}", response_model=AnalyzeStatusResponse)
+def analyze_status(analysis_id: str) -> AnalyzeStatusResponse:
+    """Poll async analysis status."""
     try:
-        disk = discover_disk(source)
+        result = get_analysis_status(analysis_id)
-        logger.info("Disk found: %s (format=%s, size=%.1f GiB)", disk.path, disk.format, disk.size_gb)
+        return AnalyzeStatusResponse(**result)
-    except Exception as exc:
+    except KeyError:
-        logger.error("Disk discovery failed: %s", exc)
+        logger.warning("Analysis not found: %s", analysis_id)
-        raise HTTPException(status_code=400, detail=f"Disk discovery failed: {exc}")
+        raise HTTPException(status_code=404, detail=f"Analysis not found: {analysis_id}")
-
-    extract_dir = source if source.is_dir() else source.parent
-    os_type = detect_os(disk.path, extract_dir)
-    efi = detect_efi(disk.path)
-    logger.info("Analysis result: os=%s efi=%s", os_type, efi)
-
-    return AnalyzeResponse(
-        vmid=req.vmid,
-        filename=disk.path.name,
-        disk_format=disk.format,
-        disk_size_gb=disk.size_gb,
-        os_type=os_type,
-        efi_detectable=efi,
-    )
 
 
 @app.post(f"{API_PREFIX}/jobs", response_model=JobStatusResponse, status_code=202)

backend/converter.py

@@ -11,6 +11,8 @@ import os
 import re
 import shutil
 import subprocess
+import threading
+import time
 from pathlib import Path
 from dataclasses import dataclass
 from typing import Optional
@@ -320,9 +322,166 @@ def detect_efi(disk_path: Path) -> Optional[bool]:
         )
         if result.returncode != 0:
             return None
-        # EFI System Partition is typically VFAT; look for it
         if "vfat" in result.stdout.lower():
             return True
         return False
     except Exception:
         return None
+
+
+# ---------------------------------------------------------------------------
+# Async analysis — background extraction + probing
+# ---------------------------------------------------------------------------
+
+_analyses: dict[str, dict] = {}
+_analyses_lock = threading.Lock()
+
+
+def start_analysis(vmid: int, source_filename: str) -> dict:
+    analysis_id = f"analysis_{vmid}_{int(time.time())}"
+    with _analyses_lock:
+        _analyses[analysis_id] = {
+            "vmid": vmid,
+            "status": "queued",
+            "message": "Queued",
+            "result": None,
+            "error": None,
+        }
+    thread = threading.Thread(target=_process_analysis, args=(analysis_id, vmid, source_filename), daemon=True)
+    thread.start()
+    return {"analysis_id": analysis_id, "status": "queued", "message": "Analysis queued"}
+
+
+def get_analysis_status(analysis_id: str) -> dict:
+    with _analyses_lock:
+        a = _analyses.get(analysis_id)
+    if not a:
+        raise KeyError(analysis_id)
+    return {
+        "analysis_id": analysis_id,
+        "vmid": a["vmid"],
+        "status": a["status"],
+        "message": a.get("message", ""),
+        "result": a.get("result"),
+        "error_details": a.get("error"),
+    }
+
+
+def _update_analysis(analysis_id: str, status: str, message: str = "", result=None, error: str = None):
+    with _analyses_lock:
+        a = _analyses.get(analysis_id)
+        if not a:
+            return
+        a["status"] = status
+        if message:
+            a["message"] = message
+        if result is not None:
+            a["result"] = result
+        if error is not None:
+            a["error"] = error
+
+
+def _process_analysis(analysis_id: str, vmid: int, source_filename: str):
+    try:
+        _update_analysis(analysis_id, "processing", "Extracting archive...")
+        source = extract_if_needed(source_filename)
+        logger.info("Extraction done: %s", source)
+
+        _update_analysis(analysis_id, "processing", "Extracting nested archives...")
+        source = _extract_nested(source)
+
+        _update_analysis(analysis_id, "processing", "Discovering disk image...")
+        disk = discover_disk(source)
+        logger.info("Disk found: %s (format=%s, size=%.1f GiB)", disk.path, disk.format, disk.size_gb)
+
+        _update_analysis(analysis_id, "processing", "Detecting OS and boot type...")
+        extract_dir = source if source.is_dir() else source.parent
+        os_type = detect_os(disk.path, extract_dir)
+        efi = detect_efi(disk.path)
+        logger.info("Analysis done: os=%s efi=%s", os_type, efi)
+
+        from models import AnalyzeResponse
+        result = AnalyzeResponse(
+            vmid=vmid,
+            filename=disk.path.name,
+            disk_format=disk.format,
+            disk_size_gb=disk.size_gb,
+            os_type=os_type,
+            efi_detectable=efi,
+        )
+        _update_analysis(analysis_id, "completed", "Analysis complete", result=result.model_dump())
+
+    except Exception as exc:
+        logger.exception("Analysis %s failed", analysis_id)
+        _update_analysis(analysis_id, "failed", "Analysis failed", error=str(exc))
+
+
+# ---------------------------------------------------------------------------
+# Nested archive extraction — handles split zips, chained archives
+# ---------------------------------------------------------------------------
+
+def _extract_nested(base: Path) -> Path:
+    """Recursively extract nested archives until disk images are found
+    or no more archives remain. Handles split zips (.z01-.zNN + .zip)."""
+    for _ in range(5):
+        if not base.is_dir():
+            return base
+
+        cand = _find_nested_archive(base)
+        if cand is None:
+            break
+
+        logger.info("Extracting nested archive: %s", cand)
+        _extract_one_archive(cand, base)
+        try:
+            cand.unlink()
+        except Exception:
+            pass
+
+        base = _flatten_nested(base)
+    return base
+
+
+def _find_nested_archive(base: Path) -> Optional[Path]:
+    """Find the next archive to extract inside a directory. Prioritises
+    split zip masters if .z01 segments are present."""
+    files = sorted([f for f in base.iterdir() if f.is_file()], key=lambda f: f.name)
+
+    split_master = None
+    has_segments = False
+    for f in files:
+        name = f.name.lower()
+        if re.search(r'\.z\d{2,3}$', name):
+            has_segments = True
+        if name.endswith(".zip") and not re.search(r'\.z\d{2,3}$', name):
+            split_master = f
+
+    if has_segments and split_master:
+        return split_master
+
+    for f in files:
+        if _detect_archive_ext(f.name.lower()) is not None:
+            return f
+    return None
+
+
+def _extract_one_archive(archive_path: Path, target_dir: Path):
+    name = archive_path.name.lower()
+    ext = _detect_archive_ext(name)
+    tool_map = {
+        ".7z":     ("7z",   ["7z", "x", "-y", str(archive_path)]),
+        ".zip":    ("7z",   ["7z", "x", "-y", str(archive_path)]),
+        ".rar":    ("unrar", ["unrar", "x", "-y", str(archive_path)]),
+        ".tar.gz": ("tar",   ["tar", "-xzf", str(archive_path)]),
+        ".tgz":    ("tar",   ["tar", "-xzf", str(archive_path)]),
+        ".tar":    ("tar",   ["tar", "-xf", str(archive_path)]),
+        ".gz":     ("gunzip", ["gunzip", "-f", str(archive_path)]),
+        ".bz2":    ("bunzip2", ["bunzip2", "-f", str(archive_path)]),
+        ".xz":     ("xz",    ["xz", "-d", str(archive_path)]),
+    }
+    tool, cmd = tool_map[ext]
+    if not shutil.which(tool):
+        raise RuntimeError(f"Extraction tool '{tool}' not found on host")
+    result = subprocess.run(cmd, cwd=str(target_dir), capture_output=True, text=True, timeout=900)
+    if result.returncode != 0:
+        logger.warning("Nested extraction %s failed: %s", archive_path.name, result.stderr.strip()[:200])

backend/models.py

@@ -35,6 +35,13 @@ class JobStatus(str, Enum):
     FAILED = "failed"
 
 
+class AnalyzeStatus(str, Enum):
+    QUEUED = "queued"
+    PROCESSING = "processing"
+    COMPLETED = "completed"
+    FAILED = "failed"
+
+
 # ---------------------------------------------------------------------------
 # Request schemas
 # ---------------------------------------------------------------------------
@@ -117,3 +124,12 @@ class CloneRequest(BaseModel):
     source_vmid: int
     target_vmid: int
     target_name: str
+
+
+class AnalyzeStatusResponse(BaseModel):
+    analysis_id: str
+    vmid: int
+    status: AnalyzeStatus
+    message: str = ""
+    result: Optional[AnalyzeResponse] = None
+    error_details: Optional[str] = None

frontend/api_client.py

@@ -15,7 +15,6 @@ from typing import Optional
 
 BACKEND_URL = os.getenv("BACKEND_URL", "http://10.2.0.2:9000")
 TIMEOUT = 30
-ANALYZE_TIMEOUT = 300  # guestfish needs time to probe large disk images
 
 
 class ApiError(Exception):
@@ -44,12 +43,16 @@ class ApiClient:
         return self._get("/api/v1/health")
 
     def analyze(self, vmid: int, filename: str, source_type: str = "upload") -> dict:
-        """POST /api/v1/analyze — probe source file for OS, size, format."""
+        """POST /api/v1/analyze — start async analysis, returns analysis_id."""
         return self._post("/api/v1/analyze", {
             "vmid": vmid,
             "source_filename": filename,
             "source_type": source_type,
-        }, timeout=ANALYZE_TIMEOUT)
+        })
+
+    def get_analysis(self, analysis_id: str) -> dict:
+        """GET /api/v1/analyze/{analysis_id} — poll analysis status."""
+        return self._get(f"/api/v1/analyze/{analysis_id}")
 
     def create_job(self, payload: dict) -> dict:
         """POST /api/v1/jobs — submit conversion job."""

frontend/app.py

@@ -299,7 +299,7 @@ async def session_progress(filename: str):
     })
 
 
-@app.post("/session/analyze", response_class=HTMLResponse)
+@app.post("/session/analyze")
 async def session_analyze(
     request: Request,
     vmid: int = Form(...),
@@ -307,28 +307,63 @@ async def session_analyze(
     vm_name: str = Form(""),
     session_id: str = Form(""),
 ):
-    """Phase 2 — call the backend /analyze endpoint and render the result."""
+    """Phase 2 — start async backend analysis, return analysis_id for polling."""
     err = _validate_vmid(vmid)
     if err:
-        return render("_analysis.html", request=request, error=err)
+        return JSONResponse({"phase": "error", "error": err}, status_code=400)
 
-    # Prepend session subdirectory to the filename for the backend
-    # Files are in tmp/{guid}/in/, so the backend needs {guid}/in/filename
     backend_filename = f"{session_id}/in/{filename}" if session_id else filename
 
     try:
         analysis = api.analyze(vmid=vmid, filename=backend_filename)
     except ApiError as exc:
-        return render("_analysis.html", request=request,
+        return JSONResponse({"phase": "error", "error": f"Backend analysis failed: {exc.detail}"}, status_code=502)
-                      error=f"Backend analysis failed: {exc.detail}")
 
-    vm_name = vm_name.strip()
+    return JSONResponse({
-    if not vm_name:
+        "phase": "analyzing",
-        vm_name = (analysis.get("os_type") or "vm") + f"-{vmid}"
+        "analysis_id": analysis.get("analysis_id"),
+        "vmid": vmid,
+        "filename": filename,
+        "vm_name": vm_name.strip(),
+    })
+
+
+@app.get("/session/analyze/status/{analysis_id}")
+async def session_analyze_status(analysis_id: str):
+    """Poll backend for async analysis status."""
+    try:
+        return api.get_analysis(analysis_id)
+    except ApiError as exc:
+        return JSONResponse({"error": exc.detail}, status_code=502)
+
+
+@app.get("/session/analyze/result/{analysis_id}", response_class=HTMLResponse)
+async def session_analyze_result(
+    request: Request,
+    analysis_id: str,
+    vmid: str = "",
+    source_filename: str = "",
+    vm_name: str = "",
+):
+    """Render the analysis result after polling completes."""
+    try:
+        status = api.get_analysis(analysis_id)
+        result = status.get("result")
+        if not result:
+            return render("_analysis.html", request=request,
+                          error="Analysis result not ready.")
+    except ApiError as exc:
+        return render("_analysis.html", request=request,
+                      error=f"Failed to fetch analysis: {exc.detail}")
+
+    analysis = result
+    name = vm_name.strip()
+    if not name:
+        name = (analysis.get("os_type") or "vm") + f"-{vmid}"
 
     return render("_analysis.html", request=request,
-                  vmid=vmid, source_filename=filename,
+                  vmid=vmid, source_filename=source_filename,
-                  vm_name=vm_name, analysis=analysis)
+                  vm_name=name, analysis=analysis)
 
 
 @app.post("/session/confirm", response_class=HTMLResponse)

frontend/templates/index.html

@@ -147,10 +147,41 @@ async function runAnalysis(vmid, filename, vmName) {
   fd.append('vmid', vmid); fd.append('filename', filename); fd.append('vm_name', vmName);
   fd.append('session_id', window.VM_BENCH_SID || '');
   const resp = await fetch('/session/analyze', { method: 'POST', body: fd });
-  const html = await resp.text();
+  const data = await resp.json();
-  document.getElementById('analysis-section').innerHTML = html;
+
-  document.getElementById('session-status').classList.add('pve-hidden');
+  if (data.phase === 'error') {
-  initConfirmForm();
+    setPhase('Analysis failed', 0, false);
+    showError(data.error);
+    return;
+  }
+
+  var analysisId = data.analysis_id;
+  var startTime = Date.now();
+
+  for (;;) {
+    await sleep(2000);
+    try {
+      var pr = await fetch('/session/analyze/status/' + analysisId);
+      var sdata = await pr.json();
+      var elapsed = Math.round((Date.now() - startTime) / 1000);
+      setPhase('Analysing... ' + (sdata.message || '') + ' (' + elapsed + 's)', sdata.status === 'completed' ? 100 : 50, sdata.status !== 'completed');
+
+      if (sdata.status === 'completed') {
+        await sleep(500);
+        var rr = await fetch('/session/analyze/result/' + analysisId + '?source_filename=' + encodeURIComponent(filename) + '&vm_name=' + encodeURIComponent(vmName) + '&vmid=' + vmid);
+        var html = await rr.text();
+        document.getElementById('analysis-section').innerHTML = html;
+        document.getElementById('session-status').classList.add('pve-hidden');
+        initConfirmForm();
+        return;
+      }
+      if (sdata.status === 'failed') {
+        setPhase('Analysis failed', 0, false);
+        showError(sdata.error_details || 'Analysis failed');
+        return;
+      }
+    } catch (_) {}
+  }
 }
 
 function initConfirmForm() {

frontend/templates/scp.html

@@ -106,15 +106,42 @@ async function startScpPull(e) {
           setScpPhase('Pull complete (' + (pdata.file_size_gb || 0) + ' GiB)', 100, false);
           await new Promise(r => setTimeout(r, 500));
           setScpPhase('Step 2/2: Analysing source image...', 0, true);
-          const fd = new FormData();
+          var fd = new FormData();
           fd.append('vmid', document.getElementById('scp-vmid').value);
           fd.append('filename', filename);
           fd.append('vm_name', document.getElementById('scp-vm-name').value.trim());
           fd.append('session_id', SESSION_ID);
-          const ar = await fetch('/session/analyze', { method: 'POST', body: fd });
+          var ar = await fetch('/session/analyze', { method: 'POST', body: fd });
-          document.getElementById('scp-analysis-section').innerHTML = await ar.text();
+          var adata = await ar.json();
-          document.getElementById('scp-status').classList.add('pve-hidden');
+          if (adata.phase === 'error') {
-          btn.disabled = false; return;
+            setScpPhase('Analysis failed', 0, false);
+            showScpError(adata.error);
+            btn.disabled = false; return;
+          }
+          var analysisId = adata.analysis_id;
+          var vmid = document.getElementById('scp-vmid').value;
+          var vmName = document.getElementById('scp-vm-name').value.trim();
+          var astart = Date.now();
+          for (;;) {
+            await new Promise(r => setTimeout(r, 2000));
+            var sr = await fetch('/session/analyze/status/' + analysisId);
+            var sdata = await sr.json();
+            var elapsed = Math.round((Date.now() - astart) / 1000);
+            setScpPhase('Analysing... ' + (sdata.message || '') + ' (' + elapsed + 's)', sdata.status === 'completed' ? 100 : 50, sdata.status !== 'completed');
+            if (sdata.status === 'completed') {
+              await new Promise(r => setTimeout(r, 500));
+              var rr = await fetch('/session/analyze/result/' + analysisId + '?source_filename=' + encodeURIComponent(filename) + '&vm_name=' + encodeURIComponent(vmName) + '&vmid=' + vmid);
+              document.getElementById('scp-analysis-section').innerHTML = await rr.text();
+              document.getElementById('scp-status').classList.add('pve-hidden');
+              scpInitConfirmForm();
+              btn.disabled = false; return;
+            }
+            if (sdata.status === 'failed') {
+              setScpPhase('Analysis failed', 0, false);
+              showScpError(sdata.error_details || 'Analysis failed');
+              btn.disabled = false; return;
+            }
+          }
         }
         if (pdata.phase === 'error') {
           setScpPhase('SCP pull failed', 0, false);
@@ -147,6 +174,68 @@ function showScpError(msg) {
   const el = document.getElementById('scp-error');
   if (el) el.innerHTML += '<div class="pve-alert pve-alert-error" style="margin-top:0.5rem;">' + msg + '</div>';
 }
+
+function scpInitConfirmForm() {
+  const form = document.getElementById('confirm-form');
+  if (!form) return;
+  form.addEventListener('submit', scpSubmitJob);
+  const bootSel = form.querySelector('select[name="auto_detect_boot"]');
+  if (bootSel) {
+    bootSel.addEventListener('change', function() {
+      document.getElementById('boot-type-group').style.display =
+        this.value === 'false' ? 'block' : 'none';
+    });
+  }
+}
+
+async function scpSubmitJob(e) {
+  e.preventDefault();
+  const form = e.target;
+  const btn = form.querySelector('#submit-btn');
+  btn.disabled = true;
+  const formData = new FormData(form);
+  formData.append('session_id', SESSION_ID);
+  try {
+    const resp = await fetch('/session/confirm', { method: 'POST', body: formData });
+    const html = await resp.text();
+    document.getElementById('scp-form-panel').style.display = 'none';
+    document.getElementById('scp-analysis-section').innerHTML = '';
+    document.getElementById('scp-analysis-section').innerHTML = html;
+    scpInitPolling();
+  } catch (err) {
+    btn.disabled = false;
+  }
+}
+
+var _scpPollTimer = null;
+function scpInitPolling() {
+  const container = document.getElementById('polling-container');
+  if (!container) return;
+  const jobId = container.dataset.jobId;
+  const startTime = Date.now();
+  var completed = false;
+
+  async function poll() {
+    try {
+      const resp = await fetch('/session/status/' + jobId);
+      const data = await resp.json();
+      const pct = data.progress_percentage || 0;
+      var fill = document.getElementById('pve-progress-bar');
+      if (fill) { fill.style.width = pct + '%'; fill.textContent = pct + '%'; }
+      var elMsg = document.getElementById('status-message');
+      if (elMsg) elMsg.textContent = data.message || '';
+      var badge = document.getElementById('status-badge');
+      if (badge) {
+        if (data.status === 'completed') { badge.className = 'pve-badge pve-badge-completed'; badge.textContent = 'Completed'; }
+        else if (data.status === 'failed') { badge.className = 'pve-badge pve-badge-failed'; badge.textContent = 'Failed'; }
+        else { badge.className = 'pve-badge pve-badge-running'; badge.innerHTML = '<span class="pve-spinner"></span> Processing'; }
+      }
+      if (data.status === 'completed' || data.status === 'failed') { completed = true; return; }
+    } catch (err) {}
+    if (!completed) _scpPollTimer = setTimeout(poll, 2000);
+  }
+  _scpPollTimer = setTimeout(poll, 3000);
+}
 </script>
 
 {% endblock %}