msiltop 0.2.0__py3-none-any.whl

msiltop/parsers.py

@@ -0,0 +1,62 @@
+import math
+
+def parse_thermal_pressure(powermetrics_parse):
+    return powermetrics_parse["thermal_pressure"]
+
+
+def parse_cpu_metrics(powermetrics_parse):
+    e_core = []
+    p_core = []
+    cpu_metrics = powermetrics_parse["processor"]
+    cpu_metric_dict = {}
+    # cpu_clusters
+    cpu_clusters = cpu_metrics["clusters"]
+    for cluster in cpu_clusters:
+        name = cluster["name"]
+        cluster_cpus = cluster["cpus"]
+        # Calculate total active time across all cores instead of averaging
+        # This better matches Activity Monitor's methodology
+        total_active = sum(map(lambda x: 1.0 - x["idle_ratio"], cluster_cpus))
+        # Express as percentage of total cluster capacity (number of cores = 100% each)
+        cpu_metric_dict[name+"_active"] = int(
+            100 * total_active / len(cluster_cpus)
+        )
+        for cpu in cluster["cpus"]:
+            name = 'E-Cluster' if name[0] == 'E' else 'P-Cluster'
+            core = e_core if name[0] == 'E' else p_core
+            core.append(cpu["cpu"])
+            cpu_idle_ratio = cpu["idle_ratio"]
+            
+            if math.isnan(cpu_idle_ratio):
+                cpu_idle_ratio = 0
+            cpu_metric_dict[name + str(cpu["cpu"]) + "_active"] = int((1 - cpu_idle_ratio) * 100)
+            
+    cpu_metric_dict["e_core"] = e_core
+    cpu_metric_dict["p_core"] = p_core
+    if "E-Cluster_active" not in cpu_metric_dict:
+        # M1 Ultra
+        cpu_metric_dict["E-Cluster_active"] = int(
+            (cpu_metric_dict["E0-Cluster_active"] + cpu_metric_dict["E1-Cluster_active"])/2)
+    if "P-Cluster_active" not in cpu_metric_dict:
+        if "P2-Cluster_active" in cpu_metric_dict:
+            # M1 Ultra
+            cpu_metric_dict["P-Cluster_active"] = int((cpu_metric_dict["P0-Cluster_active"] + cpu_metric_dict["P1-Cluster_active"] +
+                                                      cpu_metric_dict["P2-Cluster_active"] + cpu_metric_dict["P3-Cluster_active"]) / 4)
+        else:
+            cpu_metric_dict["P-Cluster_active"] = int(
+                (cpu_metric_dict["P0-Cluster_active"] + cpu_metric_dict["P1-Cluster_active"])/2)
+    # power
+    cpu_metric_dict["ane_W"] = cpu_metrics["ane_energy"]/1000
+    #cpu_metric_dict["dram_W"] = cpu_metrics["dram_energy"]/1000
+    cpu_metric_dict["cpu_W"] = cpu_metrics["cpu_energy"]/1000
+    cpu_metric_dict["gpu_W"] = cpu_metrics["gpu_energy"]/1000
+    cpu_metric_dict["package_W"] = cpu_metrics["combined_power"]/1000
+    return cpu_metric_dict
+
+
+def parse_gpu_metrics(powermetrics_parse):
+    gpu_metrics = powermetrics_parse["gpu"]
+    gpu_metrics_dict = {
+        "active": int((1 - gpu_metrics["idle_ratio"])*100),
+    }
+    return gpu_metrics_dict

msiltop/soc_info.json

@@ -0,0 +1,18 @@
+{
+  "Apple M1 Max": {"cpu_max_power": 30, "gpu_max_power": 60},
+  "Apple M1 Pro": {"cpu_max_power": 30, "gpu_max_power": 30},
+  "Apple M1": {"cpu_max_power": 20, "gpu_max_power": 20},
+  "Apple M1 Ultra": {"cpu_max_power": 60, "gpu_max_power": 120},
+  "Apple M2": {"cpu_max_power": 25, "gpu_max_power": 15},
+  "Apple M2 Pro": {"cpu_max_power": 30, "gpu_max_power": 19},
+  "Apple M2 Max": {"cpu_max_power": 38, "gpu_max_power": 38},
+  "Apple M2 Ultra": {"cpu_max_power": 76, "gpu_max_power": 76},
+  "Apple M3": {"cpu_max_power": 22, "gpu_max_power": 13},
+  "Apple M3 Pro": {"cpu_max_power": 37, "gpu_max_power": 19},
+  "Apple M3 Max": {"cpu_max_power": 54, "gpu_max_power": 47},
+  "Apple M3 Ultra": {"cpu_max_power": 108, "gpu_max_power": 94},
+  "Apple M4": {"cpu_max_power": 22, "gpu_max_power": 13},
+  "Apple M4 Pro": {"cpu_max_power": 42, "gpu_max_power": 23},
+  "Apple M4 Max": {"cpu_max_power": 68, "gpu_max_power": 57},
+  "Apple M4 Ultra": {"cpu_max_power": 136, "gpu_max_power": 114}
+}

msiltop/utils.py

@@ -0,0 +1,161 @@
+import os
+import glob
+import subprocess
+from subprocess import PIPE
+import psutil
+from .parsers import *
+import plistlib
+import json
+
+
+def parse_powermetrics(path='/tmp/msiltop_powermetrics', timecode="0", tail_bytes=1024 * 512):
+    try:
+        with open(path + timecode, 'rb') as fp:
+            fp.seek(0, os.SEEK_END)
+            size = fp.tell()
+            fp.seek(max(size - tail_bytes, 0))
+            data = fp.read()
+        parts = data.rsplit(b'\x00', maxsplit=2)
+        for part in reversed(parts):
+            if part.strip():
+                powermetrics_parse = plistlib.loads(part)
+                thermal_pressure = parse_thermal_pressure(powermetrics_parse)
+                cpu_metrics_dict = parse_cpu_metrics(powermetrics_parse)
+                gpu_metrics_dict = parse_gpu_metrics(powermetrics_parse)
+                bandwidth_metrics = None
+                timestamp = powermetrics_parse["timestamp"]
+                return cpu_metrics_dict, gpu_metrics_dict, thermal_pressure, bandwidth_metrics, timestamp
+    except Exception:
+        return False
+
+
+def clear_console():
+    command = 'clear'
+    os.system(command)
+
+
+def convert_to_GB(value):
+    return round(value/1024/1024/1024, 1)
+
+
+def run_powermetrics_process(timecode, nice=10, interval=1000):
+    #ver, *_ = platform.mac_ver()
+    #major_ver = int(ver.split(".")[0])
+    for tmpf in glob.glob("/tmp/msiltop_powermetrics*"):
+        os.remove(tmpf)
+    output_file_flag = "-o"
+    command = " ".join([
+        "sudo nice -n",
+        str(nice),
+        "powermetrics",
+        "--samplers cpu_power,gpu_power,thermal",
+        output_file_flag,
+        "/tmp/msiltop_powermetrics"+timecode,
+        "-f plist",
+        "-i",
+        str(interval)
+    ])
+    process = subprocess.Popen(
+        command.split(" "),
+        stdin=PIPE,
+        stdout=subprocess.DEVNULL,
+        stderr=subprocess.DEVNULL,
+    )
+    return process
+
+
+def get_ram_metrics_dict():
+    ram_metrics = psutil.virtual_memory()
+    swap_metrics = psutil.swap_memory()
+    total_GB = convert_to_GB(ram_metrics.total)
+    free_GB = convert_to_GB(ram_metrics.available)
+    used_GB = convert_to_GB(ram_metrics.total-ram_metrics.available)
+    swap_total_GB = convert_to_GB(swap_metrics.total)
+    swap_used_GB = convert_to_GB(swap_metrics.used)
+    swap_free_GB = convert_to_GB(swap_metrics.total-swap_metrics.used)
+    if swap_total_GB > 0:
+        swap_free_percent = int(100-(swap_free_GB/swap_total_GB*100))
+    else:
+        swap_free_percent = None
+    ram_metrics_dict = {
+        "total_GB": round(total_GB, 1),
+        "free_GB": round(free_GB, 1),
+        "used_GB": round(used_GB, 1),
+        "free_percent": int(100-(ram_metrics.available/ram_metrics.total*100)),
+        "swap_total_GB": swap_total_GB,
+        "swap_used_GB": swap_used_GB,
+        "swap_free_GB": swap_free_GB,
+        "swap_free_percent": swap_free_percent,
+    }
+    return ram_metrics_dict
+
+
+def get_cpu_info():
+    cpu_info = os.popen('sysctl -a | grep machdep.cpu').read()
+    cpu_info_lines = cpu_info.split("\n")
+    data_fields = ["machdep.cpu.brand_string", "machdep.cpu.core_count"]
+    cpu_info_dict = {}
+    for l in cpu_info_lines:
+        for h in data_fields:
+            if h in l:
+                value = l.split(":")[1].strip()
+                cpu_info_dict[h] = value
+    return cpu_info_dict
+
+
+def get_core_counts():
+    cores_info = os.popen('sysctl -a | grep hw.perflevel').read()
+    cores_info_lines = cores_info.split("\n")
+    data_fields = ["hw.perflevel0.logicalcpu", "hw.perflevel1.logicalcpu"]
+    cores_info_dict = {}
+    for l in cores_info_lines:
+        for h in data_fields:
+            if h in l:
+                value = int(l.split(":")[1].strip())
+                cores_info_dict[h] = value
+    return cores_info_dict
+
+
+def get_gpu_cores():
+    try:
+        cores = os.popen(
+            "system_profiler -detailLevel basic SPDisplaysDataType | grep 'Total Number of Cores'").read()
+        cores = int(cores.split(": ")[-1])
+    except:
+        cores = "?"
+    return cores
+
+
+def get_soc_info():
+    cpu_info_dict = get_cpu_info()
+    core_counts_dict = get_core_counts()
+    try:
+        e_core_count = core_counts_dict["hw.perflevel1.logicalcpu"]
+        p_core_count = core_counts_dict["hw.perflevel0.logicalcpu"]
+    except:
+        e_core_count = "?"
+        p_core_count = "?"
+    soc_info = {
+        "name": cpu_info_dict["machdep.cpu.brand_string"],
+        "core_count": int(cpu_info_dict["machdep.cpu.core_count"]),
+        "cpu_max_power": None,
+        "gpu_max_power": None,
+        "e_core_count": e_core_count,
+        "p_core_count": p_core_count,
+        "gpu_core_count": get_gpu_cores()
+    }
+    # TDP (power)
+    soc_info_defaults = {
+        "cpu_max_power": 20,
+        "gpu_max_power": 20,
+    }
+    try:
+        config_path = os.path.join(os.path.dirname(__file__), "soc_info.json")
+        with open(config_path, "r", encoding="utf-8") as config_file:
+            soc_info_map = json.load(config_file)
+        soc_info_defaults.update(soc_info_map.get(soc_info["name"], {}))
+    except Exception:
+        pass
+    soc_info["cpu_max_power"] = soc_info_defaults["cpu_max_power"]
+    soc_info["gpu_max_power"] = soc_info_defaults["gpu_max_power"]
+    return soc_info

msiltop-0.2.0.dist-info/METADATA

@@ -0,0 +1,276 @@
+Metadata-Version: 2.4
+Name: msiltop
+Version: 0.2.0
+Summary: Real-time macOS hardware performance monitoring for Apple Silicon (M1/M2/M3/M4) with AI workload focus - enhanced asitop alternative
+Project-URL: Homepage, https://github.com/pratikdevnani/msiltop
+Project-URL: Repository, https://github.com/pratikdevnani/msiltop
+Project-URL: Issues, https://github.com/pratikdevnani/msiltop/issues
+Project-URL: Documentation, https://github.com/pratikdevnani/msiltop#readme
+Project-URL: Changelog, https://github.com/pratikdevnani/msiltop/releases
+Author-email: Fluid Inference <dev@fluidinference.com>
+License-Expression: MIT
+License-File: LICENSE
+Keywords: ai,ane,apple-silicon,artificial-intelligence,asitop,cli,command-line,cpu,gpu,hardware,htop,m1,m2,m3,m4,machine-learning,macos,memory,ml,monitoring,msiltop,neural-engine,nvtop,performance,powermetrics,realtime,system-monitor,terminal,workload
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Science/Research
+Classifier: Intended Audience :: System Administrators
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: MacOS
+Classifier: Operating System :: MacOS :: MacOS X
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Software Development :: Debuggers
+Classifier: Topic :: System :: Hardware
+Classifier: Topic :: System :: Monitoring
+Classifier: Topic :: System :: Systems Administration
+Classifier: Topic :: Utilities
+Requires-Python: >=3.10
+Requires-Dist: click>=8.2.0
+Requires-Dist: psutil>=7.0.0
+Requires-Dist: textual-plotext>=1.0.0
+Requires-Dist: textual>=3.0.0
+Description-Content-Type: text/markdown
+
+# MSILTOP - MacOS Performance Monitor for Local AI
+
+[![PyPI version](https://badge.fury.io/py/msiltop.svg)](https://badge.fury.io/py/msiltop)
+[![Python 3.8+](https://img.shields.io/badge/python-3.8+-blue.svg)](https://www.python.org/downloads/)
+[![macOS](https://img.shields.io/badge/macOS-Monterey%2B-green.svg)](https://www.apple.com/macos/)
+[![Apple Silicon](https://img.shields.io/badge/Apple%20Silicon-M1%2FM2%2FM3%2FM4-orange.svg)](https://www.apple.com/mac/)
+[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
+
+**Real-time macOS hardware performance monitoring CLI tool for Apple Silicon Macs (M1/M2/M3/M4+) with AI workload focus**
+
+MSILTOP is a Python-based `nvtop`-inspired command line tool specifically designed for Apple Silicon Macs. This enhanced and actively maintained fork of the original [asitop](https://github.com/tlkh/asitop) project provides comprehensive hardware monitoring with additional features, support for newer Apple Silicon chips, and optimizations for modern terminal emulators including Ghostty.
+
+
+## 📦 Installation & Usage
+
+### Quick Start with UV (Recommended)
+
+```bash
+# Install UV package manager (if not already installed)
+curl -LsSf https://astral.sh/uv/install.sh | sh
+```
+
+```bash
+# Run MSILTOP directly without installation (one-time usage)
+sudo uvx msiltop
+
+# Run with custom options
+sudo uvx msiltop --interval 2 --color 5 --avg 60
+
+# Or install as a tool for regular usages
+uv tool install msiltop@latest -U
+sudo uv tool run msiltop
+```
+
+```bash
+# to update msiltop
+uv tool install msiltop@latest -U
+```
+
+```bash
+# Run with custom options
+sudo uvx msiltop --interval 2 --color 5 --avg 60
+```
+
+### Install from PyPI
+
+```bash
+# Install with pip
+pip install msiltop
+
+# Run after installation
+sudo msiltop
+```
+
+## 🚀 Key Features & Capabilities
+
+### Hardware Monitoring
+* **Real-time CPU monitoring** - Individual core utilization and frequency tracking
+* **GPU performance tracking** - Apple GPU usage, memory, and power consumption  
+* **Neural Engine (ANE) monitoring** - AI/ML workload detection and utilization
+* **Memory usage monitoring** - RAM and swap memory tracking
+* **Power consumption analysis** - CPU/GPU power draw with thermal throttling detection
+* **Temperature monitoring** - System thermal state and throttling alerts
+
+### Apple Silicon Support
+* **Complete Apple Silicon coverage** - M1, M2, M3, M4, and future chip support
+* **Optimized for modern terminals** - Enhanced Ghostty compatibility and performance
+* **Hardware-specific metrics** - TDP specifications for all variants
+* **Individual core monitoring** - Detailed per-core performance and efficiency tracking
+
+### AI & Machine Learning Focus
+* **AI workload detection** - Specialized monitoring for machine learning tasks
+* **Neural Engine utilization** - Track AI inference and training workloads
+* **Memory usage optimization** - Critical for large model performance
+* **Future ML framework integration** - Planned support for popular AI libraries
+
+### Development Installation
+
+```bash
+# Clone repository
+git clone https://github.com/pratikdevnani/msiltop.git
+cd msiltop
+
+# Install in development mode
+uv sync
+sudo uv run msiltop
+```
+
+## 🎛️ Command Line Options & Configuration
+
+```bash
+msiltop [OPTIONS]
+
+Options:
+  --interval INTERVAL   Display refresh rate in seconds (default: 1.0)
+  --color COLOR        Color theme selection 0-8 (default: 2)
+  --avg AVG           Power averaging window in seconds (default: 30)
+  --show_cores        Enable individual CPU core monitoring
+  --max_count COUNT   Restart powermetrics after N samples (stability)
+  -h, --help          Show help message and exit
+```
+
+### Usage Examples
+
+```bash
+# Basic monitoring with 1-second refresh
+sudo msiltop
+
+# High-frequency monitoring for AI workloads
+sudo msiltop --interval 0.5 --show_cores
+
+# Long-term monitoring with 60-second power averaging
+sudo msiltop --avg 60 --max_count 1000
+
+# Custom color theme
+sudo msiltop --color 5
+```
+
+## 🔧 How MSILTOP Works
+
+MSILTOP leverages macOS's built-in [`powermetrics`](https://www.unix.com/man-page/osx/1/powermetrics/) utility to access hardware performance counters with minimal system impact. Root privileges are required due to `powermetrics` security requirements.
+
+### Technical Architecture
+
+* **CPU/GPU Utilization:** Active residency measurements via `powermetrics`
+* **Power Consumption:** Hardware energy counters and thermal state monitoring
+* **Memory Statistics:** [`psutil`](https://github.com/giampaolo/psutil) virtual memory and swap tracking
+* **System Information:** `sysctl` CPU details and `system_profiler` GPU specifications
+* **Hardware Database:** Built-in TDP specifications for all Apple Silicon variants
+
+**Note**: Bandwidth monitoring was removed as macOS 13 deprecated bandwidth support in powermetrics.
+
+### System Requirements
+
+* **Hardware:** Apple Silicon Mac (M1, M2, M3, M4, or newer)
+* **Operating System:** macOS Monterey (12.0) or later
+* **Python:** Python 3.8+ (automatically managed with UV)
+* **Privileges:** Root access required for `powermetrics`
+
+### Why Fork asitop?
+
+The original `asitop` provided excellent Apple Silicon monitoring but lacked:
+
+- **Modern hardware support** - M3, M4+ compatibility
+- **Terminal compatibility** - Ghostty and modern terminal optimization  
+- **AI workload focus** - Machine learning specific monitoring
+- **Active maintenance** - Regular updates and bug fixes
+
+### Migration from asitop
+
+MSILTOP is a drop-in replacement for asitop with identical command-line interface:
+
+```bash
+# Replace this:
+sudo asitop
+
+# With this:
+sudo msiltop
+```
+
+## 📚 Documentation
+
+For detailed information, troubleshooting, and frequently asked questions, see our comprehensive [FAQ](docs/FAQ.md) which covers:
+
+- **Apple Silicon Architecture**: Understanding E-CPU vs P-CPU cores
+- **System Requirements**: Supported chips and compatibility  
+- **Monitoring Details**: Power accuracy, Neural Engine tracking, and more
+- **Troubleshooting**: Common issues and solutions
+- **Migration Guide**: Differences from asitop and compatibility
+
+## 🗺️ Roadmap & Development
+
+### Completed Features
+- ✅ Enhanced hardware support (M1-M4+)
+- ✅ Ghostty terminal optimization  
+- ✅ Improved documentation and user experience
+- ✅ PyPI publishing and UV integration
+
+### In Development
+- 🔄 monitor a specific process/application
+- 🔄 GUI-less monitoring and data collection
+- 🔄 Performance data export capabilities (CSV, JSON)
+- API interface for other Python services
+
+## 🛠️ Development & Publishing
+
+### Automated Release Process
+
+MSILTOP uses GitHub Actions to automate the entire release process. No manual steps required!
+
+#### 🧪 Development Workflow
+
+```bash
+# Setup development environment
+git clone https://github.com/pratikdevnani/msiltop.git
+cd msiltop
+uv sync
+
+# Make changes to code
+# Test changes locally
+sudo uv run msiltop
+
+# Run with development options
+sudo uv run msiltop --interval 0.5 --show_cores
+
+# Test build (optional)
+uv build
+sudo uv run msiltop --help  # Verify build works
+```
+
+
+## 🐛 Known Issues & Contributing
+
+### Current Issues
+- Chart height doesn't adapt to terminal height (width works correctly)
+- Plot colors don't always respect theme selection
+- Long-running sessions may require periodic restart
+
+### Contributing
+- Report bugs and request features via [GitHub Issues](https://github.com/pratikdevnani/msiltop/issues)
+- Submit pull requests for bug fixes and improvements
+- Improve documentation and examples
+- Test on different Apple Silicon variants
+
+## 📄 License
+
+MIT License - maintaining compatibility with the original asitop project.
+
+## 🙏 Acknowledgments
+
+This project builds upon the excellent foundation created by [Timothy Liu](https://github.com/tlkh) with the original [asitop](https://github.com/tlkh/asitop) project. We extend our gratitude for creating the groundwork for Apple Silicon performance monitoring.
+
+This project adds to the MSILTOP project created by [pratikdevnani](https://github.com/pratikdevnani/msiltop).
+
+---
+
+**Keywords:** Apple Silicon monitoring, M1 M2 M3 M4 performance, macOS system monitor, AI workload tracking, Neural Engine monitoring, GPU utilization, real-time hardware stats, terminal performance tool, powermetrics CLI, asitop alternative

msiltop-0.2.0.dist-info/RECORD

@@ -0,0 +1,10 @@
+msiltop/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+msiltop/msiltop.py,sha256=ESm4BlF1QwQ3BmOzvuPNsxPhNTcp9rOYp8joBte35wg,38298
+msiltop/parsers.py,sha256=WTrI09rjKVKP7wxv3dWckn-EgXG3Qr053BTb61iYqV4,2701
+msiltop/soc_info.json,sha256=EPSMp7pGZyhE3Z8Pqc1VOqxTLM3k4tZq3ek_ly2hcwg,990
+msiltop/utils.py,sha256=szm69SZNiu_yH_RaTB1B6CSpdlpo-3UFulqOiHD_K64,5266
+msiltop-0.2.0.dist-info/METADATA,sha256=kGKkpMkxJSLpMaGK6_Vl-sPUBaLQd_xuGEmL_g2R1Go,10327
+msiltop-0.2.0.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+msiltop-0.2.0.dist-info/entry_points.txt,sha256=7q0YR2ZtNUtndkgfCHP8rvO3Elg6nEqGxUXIk-A7zCk,49
+msiltop-0.2.0.dist-info/licenses/LICENSE,sha256=Ks65kWYKLsrlfXMglBPLfcXvRDpNo6miAedt3h-lICs,1103
+msiltop-0.2.0.dist-info/RECORD,,

msiltop-0.2.0.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.28.0
+Root-Is-Purelib: true
+Tag: py3-none-any

msiltop-0.2.0.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+msiltop = msiltop.msiltop:main

msiltop-0.2.0.dist-info/licenses/LICENSE

@@ -0,0 +1,22 @@
+MIT License
+
+Copyright (c) 2021 Timothy Liu
+Copyright (c) 2025 Fluid Inference
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.