Fix/windows llama cpp serve and test upstream (#2669)

* fix: code runner base64, Windows serve paths, endpoint cache clear, copy-log guards, model-picker remove-recent

* Revert model-picker 'remove from recent' feature and remove stray PR_DRAFT.md

routes/cookbook_helpers.py

@@ -342,6 +342,15 @@ def _cached_model_scan_script(model_dirs: list[str] | None = None) -> str:
         "                if f.is_file(): nf += 1; sz += f.stat().st_size",
         "                if f.name.endswith('.incomplete'): ic = True",
         "        snap = os.path.join(cache, d, 'snapshots')",
+        "        # Windows HF cache stores files directly in snapshots/; blobs/ may be empty.",
+        "        # Fallback: scan snapshots for real files when blobs yielded nothing.",
+        "        if sz == 0 and os.path.isdir(snap):",
+        "            for sd in os.listdir(snap):",
+        "                sf = os.path.join(snap, sd)",
+        "                if not os.path.isdir(sf): continue",
+        "                for f in os.scandir(sf):",
+        "                    if f.is_file(): nf += 1; sz += f.stat().st_size",
+        "                    if f.name.endswith('.incomplete'): ic = True",
         "        is_diffusion = False; gguf_files = []",
         "        if os.path.isdir(snap):",
         "            for sd in os.listdir(snap):",

routes/cookbook_routes.py

@@ -799,6 +799,10 @@ def setup_cookbook_routes() -> APIRouter:
                 existing.name = display_name
                 if supports_tools is not None:
                     existing.supports_tools = supports_tools
+                # Wipe stale model lists so the picker re-probes and discovers
+                # the newly-served model instead of showing the old one.
+                existing.cached_models = None
+                existing.hidden_models = None
                 db.commit()
                 logger.info(f"Updated existing local model endpoint: {base_url}")
                 return existing.id
@@ -1089,13 +1093,23 @@ def setup_cookbook_routes() -> APIRouter:
                 runner_lines.append('    ODYSSEUS_OLLAMA_PORT="$_ody_try_port"')
                 runner_lines.append('    break')
                 runner_lines.append('  fi')
-                runner_lines.append('  exec 3<&-; exec 3>&-')
+                runner_lines.append('  echo "[odysseus] Ollama API ready on port ${ODYSSEUS_OLLAMA_PORT}: ${ODYSSEUS_OLLAMA_URL}"')
-                runner_lines.append('done')
+                runner_lines.append('  echo "[odysseus] This task is monitoring an existing Ollama server; stopping it here will not stop an external Docker/system service."')
+                if local_windows:
+                    # Windows detached process has no TTY; exec bash -i crashes.
+                    # Keep the monitoring task alive with a sleep loop.
+                    runner_lines.append('  while true; do sleep 60; done')
+                else:
+                    runner_lines.append('  exec bash -i')
+                runner_lines.append('fi')
                 runner_lines.append('if ! command -v ollama &>/dev/null; then')
                 runner_lines.append('  echo "ERROR: Ollama not found on this server. Install it from https://ollama.com/download or `curl -fsSL https://ollama.com/install.sh | sh`."')
                 runner_lines.append('  echo')
                 runner_lines.append('  echo "=== Process exited with code 127 ==="')
-                runner_lines.append('  exec bash -i')
+                if local_windows:
+                    runner_lines.append('  exit 127')
+                else:
+                    runner_lines.append('  exec bash -i')
                 runner_lines.append('fi')
                 runner_lines.append('ODYSSEUS_OLLAMA_URL="http://${ODYSSEUS_OLLAMA_HOST}:${ODYSSEUS_OLLAMA_PORT}"')
                 if remote and _ollama_host in ("0.0.0.0", "::"):
@@ -1103,10 +1117,13 @@ def setup_cookbook_routes() -> APIRouter:
                     runner_lines.append('echo "[odysseus] Ollama has no built-in authentication; expose this only on a trusted LAN/VPN or provide an explicit OLLAMA_HOST with your own access controls."')
                 runner_lines.append('echo "Starting ollama server on ${ODYSSEUS_OLLAMA_HOST}:${ODYSSEUS_OLLAMA_PORT}..."')
                 runner_lines.append('OLLAMA_HOST="${ODYSSEUS_OLLAMA_HOST}:${ODYSSEUS_OLLAMA_PORT}" ollama serve')
-                runner_lines.append('_ody_exit=$?')
+                if local_windows:
-                runner_lines.append('echo')
+                    _append_serve_exit_code_lines(runner_lines, keep_shell_open=False)
-                runner_lines.append('echo "=== Process exited with code ${_ody_exit} ==="')
+                else:
-                runner_lines.append('exec bash -i')
+                    runner_lines.append('_ody_exit=$?')
+                    runner_lines.append('echo')
+                    runner_lines.append('echo "=== Process exited with code ${_ody_exit} ==="')
+                    runner_lines.append('exec bash -i')
             elif "vllm serve" in req.cmd:
                 # vLLM is CUDA/ROCm-only and does not run on macOS at all.
                 runner_lines.append('if [ "$(uname -s)" = "Darwin" ]; then')
@@ -2104,11 +2121,13 @@ def setup_cookbook_routes() -> APIRouter:
                 "inc=os.path.isdir(blobs) and any(x.endswith('.incomplete') for x in os.listdir(blobs));"
                 "sys.exit(0 if ok and not inc else 1)"
             )
-            if not remote_host:
+            if remote_host:
-                import sys
-                cmd = [sys.executable, "-c", py, repo_id]
-            else:
                 cmd = ["python3", "-c", py, repo_id]
+            else:
+                # Local Windows: python3 can hit the Microsoft Store stub. Use the
+                # real Python Odysseus is running under (guaranteed to exist).
+                import sys as _sys_local
+                cmd = [_sys_local.executable, "-c", py, repo_id]
             try:
                 if remote_host:
                     ssh_base = ["ssh"]

routes/shell_routes.py

@@ -373,7 +373,11 @@ async def _create_shell(command: str, **kwargs):
     and env variable expansion errors under Git Bash.
     """
     if IS_WINDOWS:
-        if command.strip().lower().startswith("powershell"):
+        # PowerShell commands (used by the frontend for Windows log-file polling
+        # and session management) must run directly — passing them through
+        # bash -c mangles $env:VAR syntax and breaks the command.
+        cmd_trim = command.strip()
+        if cmd_trim.startswith("powershell") or cmd_trim.startswith("cmd "):
             return await asyncio.create_subprocess_shell(command, **kwargs)
         bash = find_bash()
         if bash:
@@ -758,7 +762,7 @@ def setup_shell_routes() -> APIRouter:
             return {"stdout": "", "stderr": "No command provided", "exit_code": 1}
 
         logger.info("User shell exec requested: length=%d", len(cmd))
-        result = await _exec_shell(cmd, timeout=EXEC_TIMEOUT)
+        result = await _exec_shell(cmd, timeout=req.timeout if req.timeout is not None else EXEC_TIMEOUT)
         return result
 
     @router.post("/api/shell/stream")

static/js/codeRunner.js

@@ -310,11 +310,15 @@ try {
  */
 export async function runServer(code, panel, lang) {
   showLoading(panel, 'Running on server...');
+  // Base64-encode the script so newlines survive the shell quoting intact.
+  // JSON.stringify turns \n into literal \\n which python3 -c sees as backslash-n;
+  // base64 avoids every quoting/escaping pitfall.
+  const b64 = btoa(unescape(encodeURIComponent(code)));
   var command;
   if (lang === 'python' || lang === 'py') {
-    command = 'python3 -c ' + JSON.stringify(code);
+    command = `python3 -c "import base64; exec(base64.b64decode('${b64}').decode('utf-8'))"`;
   } else {
-    command = 'bash -c ' + JSON.stringify(code);
+    command = `python3 -c "import base64, subprocess, sys; sys.exit(subprocess.run(['bash','-c',base64.b64decode('${b64}').decode('utf-8')]).returncode)"`;
   }
   try {
     var res = await fetch('/api/shell/exec', {

static/js/cookbook-hwfit.js

@@ -443,6 +443,9 @@ export async function _hwfitFetch(fresh = false) {
   if (_cached) {
     _hwfitCache = _cached;
     _hwfitRenderHw(hw, _cached.system);
+    if (!remoteHost && _cached.system && _cached.system.platform) {
+      _envState.platform = _cached.system.platform;
+    }
     _hwfitRenderList(list, _applyEngineFilter(_cached.models));
   } else {
     // Show spinner while scanning — stack the spinner above a text label
@@ -578,6 +581,11 @@ export async function _hwfitFetch(fresh = false) {
     }
     _hwfitCache = data;
     _hwfitRenderHw(hw, data.system);
+    // Propagate local platform from hardware probe so _isWindows(task) works
+    // for local tasks (menu items, shell commands, etc.).
+    if (!remoteHost && data.system && data.system.platform) {
+      _envState.platform = data.system.platform;
+    }
     // Sort client-side by the active column so the highest↔lowest toggle is
     // deterministic (the previous array .reverse() didn't reliably flip).
     // 1st click on a column = highest first; clicking it again = lowest first.

static/js/cookbookRunning.js

@@ -1862,7 +1862,17 @@ export function _renderRunningTab() {
       const startNow = el.querySelector('.cookbook-task-start-now');
       if (startNow) startNow.style.display = (task.type === 'download' && task.status === 'queued') ? '' : 'none';
       const terminalDiag = _terminalServeDiagnosis(task, el.querySelector('.cookbook-output-pre')?.textContent || task.output || '');
-      if (terminalDiag) _showDiagnosis(el, terminalDiag, el.querySelector('.cookbook-output-pre')?.textContent || task.output || '');
+      if (terminalDiag) {
+        _showDiagnosis(el, terminalDiag, el.querySelector('.cookbook-output-pre')?.textContent || task.output || '');
+      } else {
+        const existingDiag = el.querySelector('.cookbook-diagnosis');
+        // Keep diagnosis for failed tasks even if output was cleared and we
+        // can no longer re-derive the exact message — removing it would hide
+        // the crash reason from the user.
+        if (existingDiag && !['stopped', 'error', 'crashed', 'failed'].includes(task.status)) {
+          existingDiag.remove();
+        }
+      }
     }
     if (!task) {
       if (el._uptimeInterval) { clearInterval(el._uptimeInterval); el._uptimeInterval = null; }
@@ -2201,6 +2211,10 @@ export function _renderRunningTab() {
         items.push({ label: 'Copy last 50 lines', action: 'copy-log', custom: () => {
           const out = (el.querySelector('.cookbook-output-pre')?.textContent || task.output || '');
           const last = out.split('\n').slice(-50).join('\n');
+          if (!last.trim()) {
+            uiModule.showToast('No log content available yet');
+            return;
+          }
           _copyText(last);
           uiModule.showToast('Copied last 50 lines');
         }});
@@ -2437,6 +2451,10 @@ export function _renderRunningTab() {
     el.querySelector('.cookbook-output-copy').addEventListener('click', (e) => {
       e.stopPropagation();
       const text = el.querySelector('.cookbook-output-pre')?.textContent || '';
+      if (!text.trim()) {
+        uiModule.showToast('No log content available yet');
+        return;
+      }
       _copyText(text).then(() => {
         const btn = el.querySelector('.cookbook-output-copy');
         const origHTML = btn.innerHTML;

static/js/cookbookServe.js

@@ -242,6 +242,21 @@ function _shellPathExpr(path) {
 function _selectedGgufExpr(model, repo, relPath) {
   const rel = String(relPath || '').replace(/^\/+/, '');
   if (!rel) return '';
+  if (_isWindows()) {
+    // PowerShell: plain path — no bash $() syntax (backend validator rejects
+    // $( ) in non-prelude commands, and PowerShell doesn't have printf).
+    const relW = rel.replace(/\//g, '\\');
+    if (model.is_local_dir && model.path) {
+      const base = String(model.path || '').replace(/\/+$/, '').replace(/\//g, '\\');
+      return `${base}\\${repo.replace(/\//g, '\\')}\\${relW}`;
+    }
+    if (model.path) {
+      const base = String(model.path || '').replace(/\/+$/, '').replace(/\//g, '\\');
+      return `${base}\\models--${repo.replace(/\//g, '--')}\\snapshots\\${relW}`;
+    }
+    const cacheRepo = repo.replace(/\//g, '--');
+    return `$env:USERPROFILE\\.cache\\huggingface\\hub\\models--${cacheRepo}\\snapshots\\${relW}`;
+  }
   if (model.is_local_dir && model.path) {
     const base = String(model.path || '').replace(/\/+$/, '');
     return `$(printf %s ${_shellPathExpr(`${base}/${repo}/${rel}`)})`;
@@ -255,6 +270,15 @@ function _selectedGgufExpr(model, repo, relPath) {
 }
 
 function _ggufSearchDirExpr(model, repo) {
+  if (_isWindows()) {
+    if (model.is_local_dir && model.path) {
+      return `${String(model.path || '').replace(/\/+$/, '').replace(/\//g, '\\')}\\${repo.replace(/\//g, '\\')}`;
+    }
+    if (model.path) {
+      return `${String(model.path || '').replace(/\/+$/, '').replace(/\//g, '\\')}\\models--${repo.replace(/\//g, '--')}\\snapshots`;
+    }
+    return `$env:USERPROFILE\\.cache\\huggingface\\hub\\models--${repo.replace(/\//g, '--')}\\snapshots`;
+  }
   if (model.is_local_dir && model.path) return _shellQuote(`${String(model.path || '').replace(/\/+$/, '')}/${repo}`);
   if (model.path) return _shellQuote(`${String(model.path || '').replace(/\/+$/, '')}/models--${repo.replace(/\//g, '--')}/snapshots`);
   return `"$HOME/.cache/huggingface/hub/models--${repo.replace(/\//g, '--')}/snapshots"`;
@@ -800,17 +824,27 @@ function _rerenderCachedModels() {
           // model the file lives under "<path>/<repo>" — search there just like we
           // search the HF snapshots dir, so serving a GGUF from a custom dir works
           // instead of handing llama.cpp a directory (which fails).
-          const _ldir = m.path ? _shellQuote(`${m.path}/${repo}`) : '""';
+          const _ldir = m.path
-          f._gguf_path = selectedGguf
+            ? (_isWindows() ? `${m.path.replace(/\//g, '\\')}\\${repo.replace(/\//g, '\\')}` : _shellQuote(`${m.path}/${repo}`))
-            ? _selectedGgufExpr(m, repo, selectedGguf.rel_path)
+            : (_isWindows() ? '' : '""');
-            : m.is_local_dir && m.path
+          if (selectedGguf) {
-            ? `$({ find ${_ldir} -name '*-00001-of-*.gguf' 2>/dev/null | sort; find ${_ldir} -name '*.gguf' 2>/dev/null | sort; } | head -1)`
+            f._gguf_path = _selectedGgufExpr(m, repo, selectedGguf.rel_path);
-            : `$({ find ${dir} -name '*-00001-of-*.gguf' 2>/dev/null | sort; find ${dir} -name '*.gguf' 2>/dev/null | sort; } | head -1)`;
+          } else if (_isWindows()) {
+            // Windows fallback: no bash $() available; validator rejects it.
+            // Return empty so the serve fails with a clear message.
+            f._gguf_path = '';
+          } else if (m.is_local_dir && m.path) {
+            f._gguf_path = `$({ find ${_ldir} -name '*-00001-of-*.gguf' 2>/dev/null | sort; find ${_ldir} -name '*.gguf' 2>/dev/null | sort; } | head -1)`;
+          } else {
+            f._gguf_path = `$({ find ${dir} -name '*-00001-of-*.gguf' 2>/dev/null | sort; find ${dir} -name '*.gguf' 2>/dev/null | sort; } | head -1)`;
+          }
           // Vision: auto-find the mmproj (CLIP/projector) file in the same dir.
           // Resolved at runtime so the toggle just works if an mmproj-*.gguf is
           // present (downloaded alongside the model). Empty if none → cmd omits it.
           const _vsearchdir = (m.is_local_dir && m.path) ? _ldir : dir;
-          f._mmproj_path = `$(find ${_vsearchdir} -iname 'mmproj*.gguf' 2>/dev/null | sort | head -1)`;
+          f._mmproj_path = _isWindows()
+            ? (_vsearchdir ? `${_vsearchdir}\\mmproj*.gguf` : '')
+            : `$(find ${_vsearchdir} -iname 'mmproj*.gguf' 2>/dev/null | sort | head -1)`;
         }
         if (f.reasoning_parser) {
           const _rpEl2 = panel.querySelector('[data-field="reasoning_parser"]');