Logging system

doc/source/how_to.rst

@@ -51,11 +51,47 @@ Global configuration
 Verbosity
 ---------
 
-Tensorflow is very verbose by default.
-When ``vegasflow`` is imported the environment variable ``TF_CPP_MIN_LOG_LEVEL``
-is set to 1, hiding most warnings.
-If you want to recover the usual Tensorflow logging level you can
-set your enviroment to ``export TF_CPP_MIN_LOG_LEVEL=0``.
+``VegasFlow`` uses the internal logging capabilities of python by
+creating a new logger handled named ``vegasflow``.
+You can modify the behavior of the logger as with any sane python library with the following lines:
+
+.. code-block:: python
+
+  import logging
+  log_dict = {
+        "0" : logging.ERROR,
+        "1" : logging.WARNING,
+        "2" : logging.INFO,
+        "3" : logging.DEBUG
+        }
+  logger_vegasflow = logging.getLogger('vegasflow')
+  logger_vegasflow.setLevel(log_dict["0"])
+
+Where the where the log level can be any level defined in the ``log_dict`` dictionary.
+
+Since ``VegasFlow`` is to be interfaced with non-python code it is also
+possible to control the behaviour through the environment variable ``VEGASFLOW_LOG_LEVEL``, in that case any of the keys in ``log_dict`` can be used. For instance:
+
+.. code-block:: bash
+
+  export VEGASFLOW_LOG_LEVEL=1
+
+will suppress all logger information other than ``WARNING`` and ``ERROR``.
+
+
+Environment
+-----------
+
+``VegasFlow`` is based on ``TensorFlow`` and as such all environment variable that
+have an effect on ``TensorFlow``s behavior will also have an effect on ``VegasFlow``.
+
+Here we describe only some of what we found to be the most useful variables.
+For a complete description on the variables controlling the GPU-behavior of ``TensorFlow`` please refer to
+the `nvidia official documentation <https://docs.nvidia.com/deeplearning/frameworks/tensorflow-user-guide/index.html#variablestf>`_.
+
+- ``TF_CPP_MIN_LOG_LEVEL``: controls the ``TensorFlow`` logging level. It is set to 1 by default so that only errors are printed.
+- ``VEGASFLOW_LOG_LEVEL``: controls the ``VegasFlow`` logging level. Set to 3 by default so that everything is printed.
+
 
 Choosing integration device
 ---------------------------
@@ -166,7 +202,3 @@ Note that here we are only filling one histograms and so the histogram tuple con
 
 
 We ship an example of an integrand which generates histograms in the github repository: `here <https://github.com/N3PDF/vegasflow/blob/master/examples/histogram_ex.py>`_.
-
-
-
-

src/vegasflow/configflow.py

@@ -1,56 +1,85 @@
 """
 Define some constants, header style
 """
+# Most of this can be moved to a yaml file without loss of generality
 import os
 
+import numpy as np
+
+# Set TF to only log errors
 os.environ.setdefault("TF_CPP_MIN_LOG_LEVEL", "1")
-# Most of this can be moved to a yaml file without loss of generality
 import tensorflow as tf
 
 # uncomment this line for debugging to avoid compiling any tf.function
-# tf.config.experimental_run_functions_eagerly(True)
+# tf.config.run_functions_eagerly(True)
+
 
-# Configure logging
+def run_eager(flag=True):
+    """ Wraper around `run_functions_eagerly` """
+    tf.config.run_functions_eagerly(flag)
+
+
+# Configure vegasflow logging
 import logging
 
 module_name = __name__.split(".")[0]
 logger = logging.getLogger(module_name)
+
+# Read the log level from environment, 3 (default) == debug, 2 == info, 1 == warning, 0 == error
+DEFAULT_LOG_LEVEL = "3"
+log_level_idx = os.environ.get("VEGASFLOW_LOG_LEVEL", DEFAULT_LOG_LEVEL)
+log_dict = {"0": logging.ERROR, "1": logging.WARNING, "2": logging.INFO, "3": logging.DEBUG}
+bad_log_warning = None
+if log_level_idx not in log_dict:
+    bad_log_warning = log_level_idx
+    log_level_idx = DEFAULT_LOG_LEVEL
+log_level = log_dict[log_level_idx]
+
 # Set level debug for development
-logger.setLevel(logging.DEBUG)
+logger.setLevel(log_level)
 # Create a handler and format it
 console_handler = logging.StreamHandler()
-console_handler.setLevel(logging.DEBUG)
+console_handler.setLevel(log_level)
 console_format = logging.Formatter("[%(levelname)s] %(message)s")
 console_handler.setFormatter(console_format)
 logger.addHandler(console_handler)
 
-# Define the tf.numberic types
+# Now that the logging has been created, warn about the bad logging level
+if bad_log_warning is not None:
+    logger.warning(
+        "Accepted log levels are: %s, received: %s", list(log_dict.keys()), bad_log_warning
+    )
+    logger.warning(f"Setting log level to its default value: {DEFAULT_LOG_LEVEL}")
+
+# Define the tensorflow number types
 DTYPE = tf.float64
 DTYPEINT = tf.int32
+FMAX = tf.constant(np.finfo(np.float64).max, dtype=DTYPE)
 
-# Define some default parameters for Vegas
-BINS_MAX = 50
-ALPHA = 1.5
-
-# Set up the logistics of the integration
-# Events Limit limits how many events are done in one single run of the event_loop
-# set it lower if hitting memory problems
-MAX_EVENTS_LIMIT = int(1e6)
-# Select the list of devices to look for
-DEFAULT_ACTIVE_DEVICES = ["GPU"]  # , 'CPU']
-
-# Create wrappers in order to have numbers of the correct type
+# The wrappers below transform tensors and array to the correct type
 def int_me(i):
-    """ Casts any interger to DTYPEINT """
+    """ Cast the input to the `DTYPEINT` type """
     return tf.cast(i, dtype=DTYPEINT)
 
 
 def float_me(i):
-    """ Cast any float to DTYPE """
+    """ Cast the input to the `DTYPE` type """
     return tf.cast(i, dtype=DTYPE)
 
 
 ione = int_me(1)
 izero = int_me(0)
 fone = float_me(1)
 fzero = float_me(0)
+
+
+# Define some default parameters for Vegas
+BINS_MAX = 50
+ALPHA = 1.5
+
+# Set up the logistics of the integration
+# Events Limit limits how many events are done in one single run of the event_loop
+# set it lower if hitting memory problems
+MAX_EVENTS_LIMIT = int(1e6)
+# Select the list of devices to look for
+DEFAULT_ACTIVE_DEVICES = ["GPU"]  # , 'CPU']