sciveo 0.0.43__tar.gz → 0.0.45__tar.gz

{sciveo-0.0.43 → sciveo-0.0.45}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sciveo
-Version: 0.0.43
+Version: 0.0.45
 Description-Content-Type: text/markdown
 Provides-Extra: mon
 
@@ -10,6 +10,9 @@ Provides-Extra: mon
 There are few configuration params samplers, which allows easy parameter tuning. The "auto" sampler perhaps is the easiest to use, but also
 "random" and "grid" ones are available.
 
+There is also the sciveo.monitor() which will start monitoring machine CPU/RAM/GPU etc.
+
+
 ## Features
 
 - **Experiment Tracking:** Easily log and track your machine learning experiments.
@@ -19,8 +22,12 @@ There are few configuration params samplers, which allows easy parameter tuning.
 
 ## Installation
 
+ - main sciveo
 pip install sciveo
 
+ - optional for sciveo monitoring
+pip install sciveo[mon]
+
 ## Example usage
 
 There are few public examples in sciveo.com.
@@ -30,6 +37,10 @@ The library has local and remote mode. The local one is ready to use, but for th
 When have sciveo account:
 export SCIVEO_SECRET_ACCESS_KEY='my_sciveo_user_auth_token'
 
+When using sciveo monitoring just run, using suitable python environment
+python -c "import sciveo; sciveo.monitor(period=120)"
+
+When using sciveo experimental projects management
 
 ```python
 

{sciveo-0.0.43 → sciveo-0.0.45}/README.md

@@ -4,6 +4,9 @@
 There are few configuration params samplers, which allows easy parameter tuning. The "auto" sampler perhaps is the easiest to use, but also
 "random" and "grid" ones are available.
 
+There is also the sciveo.monitor() which will start monitoring machine CPU/RAM/GPU etc.
+
+
 ## Features
 
 - **Experiment Tracking:** Easily log and track your machine learning experiments.
@@ -13,8 +16,12 @@ There are few configuration params samplers, which allows easy parameter tuning.
 
 ## Installation
 
+ - main sciveo
 pip install sciveo
 
+ - optional for sciveo monitoring
+pip install sciveo[mon]
+
 ## Example usage
 
 There are few public examples in sciveo.com.
@@ -24,6 +31,10 @@ The library has local and remote mode. The local one is ready to use, but for th
 When have sciveo account:
 export SCIVEO_SECRET_ACCESS_KEY='my_sciveo_user_auth_token'
 
+When using sciveo monitoring just run, using suitable python environment
+python -c "import sciveo; sciveo.monitor(period=120)"
+
+When using sciveo experimental projects management
 
 ```python
 

{sciveo-0.0.43 → sciveo-0.0.45}/sciveo/__init__.py

@@ -20,12 +20,11 @@ try:
   from sciveo.common.tools.daemon import TasksDaemon, __upload_content__
   from sciveo.content.runner import ProjectRunner
   from sciveo.content.dataset import Dataset
-  from sciveo.monitoring.monitor import BaseMonitor
+  from sciveo.monitoring.start import MonitorStart
   from sciveo.version import __version__
 
 
   TasksDaemon.current = TasksDaemon(num_threads=int(os.environ.get("SCIVEO_TASKS_NUM_THREADS", 1)))
-  TasksDaemon.current.start()
 
 
   # New Experiment
@@ -36,6 +35,7 @@ try:
       error("there is no started project")
 
   def start(project, function, configuration={}, **kwargs):
+    TasksDaemon.current.start()
     ProjectRunner.current = ProjectRunner(project=project, function=function, configuration=configuration, **kwargs)
     ProjectRunner.current.run()
 
@@ -44,11 +44,8 @@ try:
     return Dataset.get(info)
 
   # Monitoring start
-  def monitor(period=120):
-    mon = BaseMonitor(period=period)
-    mon.start()
-    while(True):
-      time.sleep(60)
+  def monitor(**kwargs):
+    MonitorStart(**kwargs)()
 
 except ImportError as e:
   pass

{sciveo-0.0.43 → sciveo-0.0.45}/sciveo/monitoring/monitor.py

@@ -15,6 +15,7 @@ import time
 import datetime
 import socket
 import psutil
+import uuid
 import numpy as np
 
 from sciveo.common.tools.logger import *
@@ -45,7 +46,7 @@ class BaseMonitor(DaemonBase):
 
     machine_serial = self.getserial()
 
-    debug(type(self).__name__, "init", machine_serial, "initial_cpu_usage", initial_cpu_usage)
+    debug(type(self).__name__, f"init monitor with period={period}", machine_serial, "initial_cpu_usage", initial_cpu_usage)
 
   def __call__(self):
     return self.data
@@ -81,15 +82,12 @@ class BaseMonitor(DaemonBase):
   def getserial(self):
     machine_serial = None
     try:
-      with open('/proc/cpuinfo','r') as fp:
-        for line in fp:
-          if line.startswith('Serial'):
-            machine_serial = line[10:26]
+      machine_serial = f"{socket.gethostname()}-{uuid.getnode()}"
     except Exception:
       pass
     if machine_serial is None:
       try:
-        machine_serial = socket.gethostname()
+        machine_serial = f"UUID-{uuid.getnode()}"
       except Exception:
         pass
     if machine_serial is None:
@@ -130,7 +128,7 @@ class BaseMonitor(DaemonBase):
 
 
 if __name__ == "__main__":
-  mon = BaseMonitor(period=30)
+  mon = BaseMonitor(period=10)
   mon.start()
 
   while(True):

sciveo-0.0.45/sciveo/monitoring/start.py

@@ -0,0 +1,57 @@
+import os
+import time
+
+from sciveo.common.tools.logger import *
+from sciveo.monitoring.monitor import BaseMonitor
+
+
+class MonitorStart:
+  def __init__(self, **kwargs):
+    self.default_arguments = {
+      "period": 120,
+      "block": True,
+      "fork": False,
+      "fork_type": 0
+    }
+    self.arguments = {}
+    for k, v in self.default_arguments.items():
+      self.arguments[k] = kwargs.get(k, v)
+
+    if len(kwargs) == 0:
+      info("sciveo monitoring default options", self.default_arguments)
+
+  def __call__(self):
+    if self.arguments["fork"] and self.arguments["block"]:
+      self.fork()
+    self.start()
+
+  def start_multiprocessing(self):
+    self.start()
+  def start_fork(self):
+    os.setsid()
+    self.start()
+  def fork(self):
+    if self.arguments["fork_type"] == 0:
+      import multiprocessing
+      background_process = multiprocessing.Process(target=self.start_multiprocessing)
+      background_process.daemon = True
+      background_process.start()
+      info("monitoring service started with", self.arguments)
+      exit()
+    else:
+      pid = os.fork()
+      if pid == 0:
+        self.start_fork()
+      else:
+        info("sciveo monitor service started with pid", pid, self.arguments)
+        exit()
+
+  def start(self):
+    period = max(self.arguments["period"], 5)
+    mon = BaseMonitor(period=period)
+    mon.start()
+    mon.join()
+
+    if self.arguments["block"]:
+      while(True):
+        time.sleep(60)

sciveo-0.0.45/sciveo/version.py

@@ -0,0 +1,2 @@
+
+__version__ = '0.0.45'

{sciveo-0.0.43 → sciveo-0.0.45}/sciveo.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sciveo
-Version: 0.0.43
+Version: 0.0.45
 Description-Content-Type: text/markdown
 Provides-Extra: mon
 
@@ -10,6 +10,9 @@ Provides-Extra: mon
 There are few configuration params samplers, which allows easy parameter tuning. The "auto" sampler perhaps is the easiest to use, but also
 "random" and "grid" ones are available.
 
+There is also the sciveo.monitor() which will start monitoring machine CPU/RAM/GPU etc.
+
+
 ## Features
 
 - **Experiment Tracking:** Easily log and track your machine learning experiments.
@@ -19,8 +22,12 @@ There are few configuration params samplers, which allows easy parameter tuning.
 
 ## Installation
 
+ - main sciveo
 pip install sciveo
 
+ - optional for sciveo monitoring
+pip install sciveo[mon]
+
 ## Example usage
 
 There are few public examples in sciveo.com.
@@ -30,6 +37,10 @@ The library has local and remote mode. The local one is ready to use, but for th
 When have sciveo account:
 export SCIVEO_SECRET_ACCESS_KEY='my_sciveo_user_auth_token'
 
+When using sciveo monitoring just run, using suitable python environment
+python -c "import sciveo; sciveo.monitor(period=120)"
+
+When using sciveo experimental projects management
 
 ```python
 

{sciveo-0.0.43 → sciveo-0.0.45}/sciveo.egg-info/SOURCES.txt

@@ -28,6 +28,7 @@ sciveo/content/project.py
 sciveo/content/runner.py
 sciveo/monitoring/__init__.py
 sciveo/monitoring/monitor.py
+sciveo/monitoring/start.py
 test/test_configuration.py
 test/test_monitoring.py
 test/test_runner.py

sciveo-0.0.45/test/test_monitoring.py

@@ -0,0 +1,97 @@
+#
+# Pavlin Georgiev, Softel Labs
+#
+# This is a proprietary file and may not be copied,
+# distributed, or modified without express permission
+# from the owner. For licensing inquiries, please
+# contact pavlin@softel.bg.
+#
+# 2024
+#
+
+import unittest
+import numpy as np
+
+from sciveo.common.tools.logger import *
+from sciveo.monitoring.monitor import *
+
+
+class TestMonitoring(unittest.TestCase):
+  def test_cpu(self):
+    m = BaseMonitor()
+    m.get_cpu_usage()
+    print(m.data)
+
+    self.assertTrue("usage per core" in m.data["CPU"])
+    self.assertTrue("usage" in m.data["CPU"])
+
+  def test_gpu(self):
+    data = {
+      "GPU": {
+        "raw_lines": [
+          "uuid, name, index, power.draw [W], fan.speed [%], memory.total [MiB], memory.used [MiB], memory.free [MiB], utilization.gpu [%], utilization.memory [%], temperature.gpu",
+          "GPU-41e0e872-f8d5-95fb-6fd7-8c143c0db7c9, NVIDIA GeForce RTX 3060, 0, 11.57, 0, 12053, 13, 12040, 0, 0, 38",
+          "GPU-f7a1d225-58db-e209-aa84-574996d070cd, NVIDIA GeForce RTX 3060, 1, 12.08, 0, 12045, 477, 11568, 0, 39, 36"
+        ]
+      }
+    }
+    sample = {}
+    config = {
+      "CPU": {
+        "usage":              {"ratio": 1.0, "metric": "%", "ylim": [0.0, 105.0]},
+        "usage per core":     {"ratio": 1.0, "metric": "%", "ylim": [0.0, 105.0]},
+      },
+      "RAM": {
+        "used":               {"ratio": 1.0 / (1024 * 1024 * 1024), "metric": "GB"},
+      },
+      "GPU": {
+        "fan.speed":          {"ratio": 1.0, "metric": "%", "ylim": [0.0, 105.0]},
+        "power.draw":         {"ratio": 1.0, "metric": "W", "ylim": [0.0, 250.0]},
+        "memory.free":        {"ratio": 1.0 / 1024, "metric": "GB"},
+        "memory.used":        {"ratio": 1.0 / 1024, "metric": "GB"},
+        "temperature.gpu":    {"ratio": 1.0, "metric": "°C", "ylim": [10, 110]},
+        "utilization.gpu":    {"ratio": 1.0, "metric": "%", "ylim": [0.0, 105.0]},
+        "utilization.memory": {"ratio": 1.0, "metric": "%", "ylim": [0.0, 105.0]},
+      }
+    }
+
+    try:
+      lines = data["GPU"]["raw_lines"]
+      header = lines[0].split(", ")
+
+      keys = []
+      gpu_keys = []
+      for k in header:
+        k_split = k.split(' ')
+        key = k_split[0]
+        keys.append(key)
+        gpu_keys.append(f"GPU {key}")
+
+      for i, k in enumerate(keys):
+        if k in config["GPU"]:
+          sample[gpu_keys[i]] = []
+        else:
+          data["GPU"][keys[i]] = []
+
+      for i in range(1, len(lines)):
+        line_values = lines[i].split(", ")
+        for j, value in enumerate(line_values):
+          try:
+            value = float(value)
+          except ValueError:
+            pass
+          if keys[j] in config["GPU"]:
+            sample[gpu_keys[j]].append(value)
+          else:
+            data["GPU"][keys[j]].append(value)
+
+      if "memory.total" in data["GPU"]:
+        print("memory.total", data["GPU"]["memory.total"])
+        for k in ["memory.used", "memory.free"]:
+          config["GPU"][k]["ylim"] = [0, max(data["GPU"]["memory.total"]) * config["GPU"][k]["ratio"]]
+          print("ylim", config["GPU"][k]["ylim"])
+    except Exception as e:
+      print("Exception", e)
+
+    print("memory.used", config["GPU"]["memory.used"])
+    print("free", config["GPU"]["memory.free"])

sciveo-0.0.43/sciveo/version.py

@@ -1,2 +0,0 @@
-
-__version__ = '0.0.43'

sciveo-0.0.43/test/test_monitoring.py

@@ -1,26 +0,0 @@
-#
-# Pavlin Georgiev, Softel Labs
-#
-# This is a proprietary file and may not be copied,
-# distributed, or modified without express permission
-# from the owner. For licensing inquiries, please
-# contact pavlin@softel.bg.
-#
-# 2024
-#
-
-import unittest
-import numpy as np
-
-from sciveo.common.tools.logger import *
-from sciveo.monitoring.monitor import *
-
-
-class TestMonitoring(unittest.TestCase):
-  def test_cpu(self):
-    m = BaseMonitor()
-    m.get_cpu_usage()
-    print(m.data)
-
-    self.assertTrue("usage_per_core" in m.data["CPU"])
-    self.assertTrue("usage" in m.data["CPU"])