OpenJarvis 1.0.2.dev700__tar.gz → 1.0.2.dev707__tar.gz

{openjarvis-1.0.2.dev700 → openjarvis-1.0.2.dev707}/.gitignore

@@ -48,7 +48,11 @@ Thumbs.db
 *.log
 results/
 logs/
-traces/
+# Anchored to repo root — DO NOT use the unanchored form `traces/`.
+# hatchling honors .gitignore when building the wheel; an unanchored
+# `traces/` pattern matches src/openjarvis/traces/ and silently drops
+# the runtime module from the wheel (issue #372).
+/traces/
 coding_task_*
 get-pip.py
 # Junk from mocked-path tests that write to their mock's __repr__ as a path

{openjarvis-1.0.2.dev700 → openjarvis-1.0.2.dev707}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: OpenJarvis
-Version: 1.0.2.dev700
+Version: 1.0.2.dev707
 Summary: OpenJarvis — modular AI assistant backend with composable intelligence primitives
 Project-URL: Homepage, https://github.com/open-jarvis/OpenJarvis
 Project-URL: Documentation, https://open-jarvis.github.io/OpenJarvis/
@@ -23,9 +23,9 @@ Requires-Dist: click>=8
 Requires-Dist: datasets>=4.5.0
 Requires-Dist: ddgs>=9.11.4
 Requires-Dist: httpx>=0.27
+Requires-Dist: nvidia-ml-py>=12.560.30
 Requires-Dist: openai>=1.30
 Requires-Dist: posthog>=3.0
-Requires-Dist: pynvml>=13.0.1
 Requires-Dist: python-telegram-bot>=22.6
 Requires-Dist: rich>=13
 Requires-Dist: tomli>=2.0; python_version < '3.11'
@@ -84,7 +84,7 @@ Requires-Dist: mkdocs>=1.6; extra == 'docs'
 Requires-Dist: mkdocstrings[python]>=0.25; extra == 'docs'
 Provides-Extra: energy-all
 Requires-Dist: amdsmi>=6.1; extra == 'energy-all'
-Requires-Dist: pynvml>=12.0; extra == 'energy-all'
+Requires-Dist: nvidia-ml-py>=12.560.30; extra == 'energy-all'
 Requires-Dist: zeus-ml[apple]; extra == 'energy-all'
 Provides-Extra: energy-amd
 Requires-Dist: amdsmi>=6.1; extra == 'energy-amd'
@@ -98,7 +98,7 @@ Requires-Dist: wandb>=0.17; extra == 'eval-wandb'
 Provides-Extra: framework-comparison
 Requires-Dist: polars>=1.0; extra == 'framework-comparison'
 Provides-Extra: gpu-metrics
-Requires-Dist: pynvml>=12.0; extra == 'gpu-metrics'
+Requires-Dist: nvidia-ml-py>=12.560.30; extra == 'gpu-metrics'
 Provides-Extra: inference-cloud
 Requires-Dist: anthropic>=0.30; extra == 'inference-cloud'
 Requires-Dist: openai>=1.30; extra == 'inference-cloud'
@@ -195,10 +195,29 @@ OpenJarvis is that stack. It is a framework for local-first personal AI, built a
 
 ## Installation
 
+**macOS / Linux:**
+
 ```bash
 curl -fsSL https://openjarvis.ai/install.sh | bash
 ```
 
+**Windows:** the installer is a `bash` script and won't run in PowerShell or `cmd`. Pick one of:
+
+- **WSL2 (recommended for the CLI / Python SDK)** — one-time setup in an admin PowerShell, then run the same `curl … | bash` inside Ubuntu:
+  ```powershell
+  wsl --install -d Ubuntu-24.04
+  ```
+  Open the Ubuntu shell that gets installed, then follow [WSL2 install instructions](https://open-jarvis.github.io/OpenJarvis/getting-started/wsl2/).
+- **Desktop app** — download the `.exe` from the [Releases page](https://github.com/open-jarvis/OpenJarvis/releases) for the GUI experience, no terminal required.
+
+> **If `curl` fails on `openjarvis.ai` with `sslv3 alert handshake failure`** ([issue #337](https://github.com/open-jarvis/OpenJarvis/issues/337)), use the GitHub mirror until the domain is restored:
+>
+> ```bash
+> curl -fsSL https://raw.githubusercontent.com/open-jarvis/OpenJarvis/main/scripts/install/install.sh | bash
+> ```
+>
+> Same script, served straight from this repo. The installer itself fetches everything else (uv, the project source, Ollama) from independent CDNs, so the rest of install proceeds normally.
+
 That's it. The installer handles everything: uv, the Python venv, Ollama, and pulling a small starter model. About 3 minutes on a typical broadband connection. Then:
 
 ```bash
@@ -207,7 +226,7 @@ jarvis
 
 The Rust extension and bigger models continue downloading in the background while you chat. Run `jarvis doctor` to see status.
 
-**Platforms:** macOS (Intel + Apple Silicon), Linux, WSL2 on Windows.
+**Platforms:** macOS (Intel + Apple Silicon), Linux, WSL2 on Windows. Native Windows is not supported — use WSL2 or the desktop binary.
 
 **Manual install / contributors:** see [docs/getting-started/install.md](docs/getting-started/install.md).
 

{openjarvis-1.0.2.dev700 → openjarvis-1.0.2.dev707}/README.md

@@ -31,10 +31,29 @@ OpenJarvis is that stack. It is a framework for local-first personal AI, built a
 
 ## Installation
 
+**macOS / Linux:**
+
 ```bash
 curl -fsSL https://openjarvis.ai/install.sh | bash
 ```
 
+**Windows:** the installer is a `bash` script and won't run in PowerShell or `cmd`. Pick one of:
+
+- **WSL2 (recommended for the CLI / Python SDK)** — one-time setup in an admin PowerShell, then run the same `curl … | bash` inside Ubuntu:
+  ```powershell
+  wsl --install -d Ubuntu-24.04
+  ```
+  Open the Ubuntu shell that gets installed, then follow [WSL2 install instructions](https://open-jarvis.github.io/OpenJarvis/getting-started/wsl2/).
+- **Desktop app** — download the `.exe` from the [Releases page](https://github.com/open-jarvis/OpenJarvis/releases) for the GUI experience, no terminal required.
+
+> **If `curl` fails on `openjarvis.ai` with `sslv3 alert handshake failure`** ([issue #337](https://github.com/open-jarvis/OpenJarvis/issues/337)), use the GitHub mirror until the domain is restored:
+>
+> ```bash
+> curl -fsSL https://raw.githubusercontent.com/open-jarvis/OpenJarvis/main/scripts/install/install.sh | bash
+> ```
+>
+> Same script, served straight from this repo. The installer itself fetches everything else (uv, the project source, Ollama) from independent CDNs, so the rest of install proceeds normally.
+
 That's it. The installer handles everything: uv, the Python venv, Ollama, and pulling a small starter model. About 3 minutes on a typical broadband connection. Then:
 
 ```bash
@@ -43,7 +62,7 @@ jarvis
 
 The Rust extension and bigger models continue downloading in the background while you chat. Run `jarvis doctor` to see status.
 
-**Platforms:** macOS (Intel + Apple Silicon), Linux, WSL2 on Windows.
+**Platforms:** macOS (Intel + Apple Silicon), Linux, WSL2 on Windows. Native Windows is not supported — use WSL2 or the desktop binary.
 
 **Manual install / contributors:** see [docs/getting-started/install.md](docs/getting-started/install.md).
 

{openjarvis-1.0.2.dev700 → openjarvis-1.0.2.dev707}/docs/getting-started/install.md

@@ -6,6 +6,18 @@ OpenJarvis ships a one-line installer for macOS, Linux, and WSL2.
 curl -fsSL https://openjarvis.ai/install.sh | bash
 ```
 
+!!! warning "`openjarvis.ai` SSL fallback (issue #337)"
+    If `curl` fails on `openjarvis.ai` with `sslv3 alert handshake failure`,
+    fetch the same script from the GitHub mirror until the domain is restored:
+
+    ```bash
+    curl -fsSL https://raw.githubusercontent.com/open-jarvis/OpenJarvis/main/scripts/install/install.sh | bash
+    ```
+
+    The installer itself pulls everything else (uv, the project source,
+    Ollama) from independent CDNs (`astral.sh`, `github.com`,
+    `ollama.com`), so the rest of install proceeds normally.
+
 About 3 minutes on a typical broadband connection. Type `jarvis` to start chatting.
 
 ## What the installer does

{openjarvis-1.0.2.dev700 → openjarvis-1.0.2.dev707}/frontend/src-tauri/src/lib.rs

@@ -495,14 +495,33 @@ async fn boot_backend(backend: SharedBackend, status: SharedStatus) {
 
     let uv_bin = resolve_bin("uv");
 
-    // Verify uv is actually installed
+    // Verify uv is actually installed. Concrete per-OS instructions —
+    // the generic "install it from astral.sh" was the #1 source of
+    // confusion on the Discord support thread; users couldn't tell whether
+    // to use winget, scoop, pip, or the official installer.
     if !std::path::Path::new(&uv_bin).exists() && uv_bin == "uv" {
         let mut s = status.lock().await;
-        s.error = Some(
-            "Could not find 'uv' (Python package manager). \
-             Install it from https://astral.sh/uv then relaunch."
-                .into(),
-        );
+        #[cfg(target_os = "windows")]
+        let msg = "Could not find 'uv' (Python package manager). \
+                   To install on Windows, open PowerShell and run:\n\n\
+                   powershell -ExecutionPolicy Bypass -c \"irm https://astral.sh/uv/install.ps1 | iex\"\n\n\
+                   Then close and relaunch this app. \
+                   (If the install completes but the app still can't find uv, \
+                   you may need to log out and back in so PATH refreshes.)";
+        #[cfg(target_os = "macos")]
+        let msg = "Could not find 'uv' (Python package manager). \
+                   To install on macOS, open Terminal and run:\n\n\
+                   curl -LsSf https://astral.sh/uv/install.sh | sh\n\n\
+                   Then relaunch this app.";
+        #[cfg(target_os = "linux")]
+        let msg = "Could not find 'uv' (Python package manager). \
+                   To install on Linux, open a terminal and run:\n\n\
+                   curl -LsSf https://astral.sh/uv/install.sh | sh\n\n\
+                   Then relaunch this app.";
+        #[cfg(not(any(target_os = "windows", target_os = "macos", target_os = "linux")))]
+        let msg = "Could not find 'uv' (Python package manager). \
+                   Install it from https://astral.sh/uv then relaunch.";
+        s.error = Some(msg.into());
         return;
     }
 
@@ -643,12 +662,25 @@ async fn boot_backend(backend: SharedBackend, status: SharedStatus) {
 
     let root = project_root.as_ref().unwrap();
 
-    // Install dependencies automatically (handles fresh clones)
+    // Install dependencies automatically (handles fresh clones).
+    //
+    // Previously we ran `uv sync` with both stdout AND stderr piped to
+    // /dev/null and discarded the exit code (`let _ = …`). When `uv sync`
+    // failed — Windows path issues, network problems, lockfile conflicts —
+    // the user saw no error, the boot continued, `uv run jarvis serve`
+    // then ran in an under-provisioned venv, and the user waited the full
+    // 600s health-check window before getting "Jarvis server did not
+    // become healthy in time" with no actionable detail (issue #331).
+    //
+    // Now: capture stderr, check the exit status, surface a useful error
+    // to the user BEFORE the long server-start wait. The status detail
+    // message also indicates this can take a couple of minutes on first
+    // boot so users don't restart the app thinking it's stuck.
     {
         let mut s = status.lock().await;
-        s.detail = "Installing dependencies...".into();
+        s.detail = "Installing dependencies (uv sync — may take 1-2 min on first boot)...".into();
     }
-    let _ = tokio::process::Command::new(&uv_bin)
+    let sync_output = tokio::process::Command::new(&uv_bin)
         .args([
             "sync",
             "--extra", "server",
@@ -656,10 +688,47 @@ async fn boot_backend(backend: SharedBackend, status: SharedStatus) {
             "--extra", "inference-google",
         ])
         .stdout(std::process::Stdio::null())
-        .stderr(std::process::Stdio::null())
+        .stderr(std::process::Stdio::piped())
         .current_dir(root)
-        .status()
+        .output()
         .await;
+    match sync_output {
+        Ok(out) if !out.status.success() => {
+            let stderr = String::from_utf8_lossy(&out.stderr);
+            // Take the last ~800 chars — uv's most useful diagnostic line
+            // is usually at the tail of the stream.
+            let tail: String = stderr
+                .chars()
+                .rev()
+                .take(800)
+                .collect::<String>()
+                .chars()
+                .rev()
+                .collect();
+            let mut s = status.lock().await;
+            s.error = Some(format!(
+                "`uv sync` failed in {} (exit {}). Last output:\n\n{}\n\n\
+                 Try opening a terminal in that directory and running \
+                 `uv sync --extra server` manually for the full output.",
+                root.display(),
+                out.status.code().unwrap_or(-1),
+                tail.trim(),
+            ));
+            return;
+        }
+        Err(e) => {
+            let mut s = status.lock().await;
+            s.error = Some(format!(
+                "Could not run `uv sync`: {}. Verify uv is installed at \
+                 `{}` and the OpenJarvis repo is at `{}`.",
+                e,
+                uv_bin,
+                root.display(),
+            ));
+            return;
+        }
+        Ok(_) => {} // success — fall through
+    }
 
     {
         let mut s = status.lock().await;

{openjarvis-1.0.2.dev700 → openjarvis-1.0.2.dev707}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "OpenJarvis"
-version = "1.0.2.dev700"
+version = "1.0.2.dev707"
 description = "OpenJarvis — modular AI assistant backend with composable intelligence primitives"
 readme = "README.md"
 requires-python = ">=3.10"
@@ -30,7 +30,7 @@ dependencies = [
     "httpx>=0.27",
     "openai>=1.30",
     "posthog>=3.0",
-    "pynvml>=13.0.1",
+    "nvidia-ml-py>=12.560.30",
     "python-telegram-bot>=22.6",
     "rich>=13",
     "tomli>=2.0; python_version < '3.11'",
@@ -80,10 +80,10 @@ server = [
     "python-multipart>=0.0.9",
 ]
 openhands = ["openhands-sdk>=1.0; python_version >= '3.12'"]
-gpu-metrics = ["pynvml>=12.0"]
+gpu-metrics = ["nvidia-ml-py>=12.560.30"]
 energy-amd = ["amdsmi>=6.1"]
 energy-apple = ["zeus-ml[apple]"]
-energy-all = ["pynvml>=12.0", "amdsmi>=6.1", "zeus-ml[apple]"]
+energy-all = ["nvidia-ml-py>=12.560.30", "amdsmi>=6.1", "zeus-ml[apple]"]
 orchestrator-training = ["torch>=2.0", "transformers>=4.40"]
 learning-dspy = ["dspy>=2.6"]
 learning-gepa = ["gepa>=0.1"]

{openjarvis-1.0.2.dev700 → openjarvis-1.0.2.dev707}/scripts/install/install.sh

@@ -29,6 +29,40 @@ for arg in "$@"; do
     esac
 done
 
+# ---- non-WSL Windows refusal ----
+# Running the installer in Git Bash / MSYS2 / Cygwin on native Windows
+# (i.e. NOT inside WSL2) gets the user into a confusing failure state:
+# uv/git tooling installs to Windows paths the rest of OpenJarvis can't
+# reach, and Ollama integration silently breaks. The supported Windows
+# path is WSL2. Bail early with a clear next step rather than letting
+# users discover this 3 minutes into a doomed install.
+case "$(uname -s 2>/dev/null)" in
+    MINGW*|MSYS*|CYGWIN*)
+        cat >&2 <<'EOF'
+install.sh: native Windows (Git Bash / MSYS2 / Cygwin) is not supported.
+
+OpenJarvis runs on Windows via WSL2. Two paths:
+
+  1. WSL2 (recommended for the CLI). One-time setup in an admin PowerShell:
+
+       wsl --install -d Ubuntu-24.04
+
+     Open the Ubuntu shell that gets installed, then re-run:
+
+       curl -fsSL https://openjarvis.ai/install.sh | bash
+
+     (or the github mirror — see README if openjarvis.ai is down.)
+
+  2. Desktop app — download the .exe from the Releases page:
+     https://github.com/open-jarvis/OpenJarvis/releases
+
+See the WSL2 install guide for the full walkthrough:
+  https://open-jarvis.github.io/OpenJarvis/getting-started/wsl2/
+EOF
+        exit 1
+        ;;
+esac
+
 # ---- root refusal ----
 if [[ "$(id -u)" -eq 0 ]]; then
     cat >&2 <<'EOF'

{openjarvis-1.0.2.dev700 → openjarvis-1.0.2.dev707}/src/openjarvis/evals/backends/external/_subprocess_runner.py

@@ -201,7 +201,15 @@ class _NullSampler(_Sampler):
 
 def _try_start_nvml() -> Optional[_Sampler]:
     try:
-        import pynvml  # type: ignore
+        # Suppress legacy pynvml deprecation FutureWarning (#389).
+        import warnings as _warnings
+        with _warnings.catch_warnings():
+            _warnings.filterwarnings(
+                "ignore",
+                message=r"The pynvml package is deprecated.*",
+                category=FutureWarning,
+            )
+            import pynvml  # type: ignore
 
         pynvml.nvmlInit()
         handle = pynvml.nvmlDeviceGetHandleByIndex(0)

{openjarvis-1.0.2.dev700 → openjarvis-1.0.2.dev707}/src/openjarvis/server/research_router.py

@@ -149,7 +149,15 @@ class _LiveGPUSampler:
         self._t_start = 0.0
         self._t_last = 0.0
         try:
-            import pynvml  # type: ignore
+            # Suppress legacy pynvml deprecation FutureWarning (#389).
+            import warnings as _warnings
+            with _warnings.catch_warnings():
+                _warnings.filterwarnings(
+                    "ignore",
+                    message=r"The pynvml package is deprecated.*",
+                    category=FutureWarning,
+                )
+                import pynvml  # type: ignore
 
             pynvml.nvmlInit()
             count = pynvml.nvmlDeviceGetCount()

{openjarvis-1.0.2.dev700 → openjarvis-1.0.2.dev707}/src/openjarvis/telemetry/energy_nvidia.py

@@ -17,7 +17,16 @@ from openjarvis.telemetry.energy_monitor import (
 logger = logging.getLogger(__name__)
 
 try:
-    import pynvml
+    # See gpu_monitor.py for the rationale — suppress the legacy
+    # `pynvml` package's deprecation FutureWarning narrowly (#389).
+    import warnings
+    with warnings.catch_warnings():
+        warnings.filterwarnings(
+            "ignore",
+            message=r"The pynvml package is deprecated.*",
+            category=FutureWarning,
+        )
+        import pynvml
 
     _PYNVML_AVAILABLE = True
 except ImportError:

{openjarvis-1.0.2.dev700 → openjarvis-1.0.2.dev707}/src/openjarvis/telemetry/gpu_monitor.py

@@ -12,7 +12,20 @@ from typing import Dict, Generator, List, Optional
 logger = logging.getLogger(__name__)
 
 try:
-    import pynvml
+    # The legacy `pynvml` PyPI package installs a meta-path-finder shim
+    # that prints a FutureWarning on every `import pynvml`, even though
+    # our pyproject.toml depends on `nvidia-ml-py` (the official NVIDIA
+    # package, same module name, no shim). The warning still fires if
+    # `pynvml` gets pulled in transitively by torch/vllm/etc. Suppress
+    # it narrowly here so user output stays clean (issue #389).
+    import warnings
+    with warnings.catch_warnings():
+        warnings.filterwarnings(
+            "ignore",
+            message=r"The pynvml package is deprecated.*",
+            category=FutureWarning,
+        )
+        import pynvml
     _PYNVML_AVAILABLE = True
 except ImportError:
     _PYNVML_AVAILABLE = False

openjarvis-1.0.2.dev707/src/openjarvis/traces/__init__.py

@@ -0,0 +1,15 @@
+"""Trace system — full interaction-level recording and analysis.
+
+The trace system captures the complete sequence of steps an agent takes to
+handle a query.  Unlike telemetry (which records per-inference metrics), traces
+record the *decision-making process*: which model was selected, what memory was
+retrieved, which tools were called, and the final response.
+
+Traces are the primary input to the learning system.
+"""
+
+from openjarvis.traces.analyzer import TraceAnalyzer
+from openjarvis.traces.collector import TraceCollector
+from openjarvis.traces.store import TraceStore
+
+__all__ = ["TraceAnalyzer", "TraceCollector", "TraceStore"]