pytype.metrics

module documentation

(source)

Classes for instrumenting code to collect various metrics. Instrumentation consists of creating the metric and then updating it. Creation can be performed at the module level or as a class attribute. Since the metric namespace is global, metrics should not be created by instances unless that instance is certain to be a singleton. Sample code: _my_counter = metrics.Counter("my-counter") def foo(): _my_counter.inc() # calls to foo() count as 1 unit. def bar(n): _my_counter.inc(n) # calls to bar() count as n units.

Class	`Counter`	A monotonically increasing metric.
Class	`Distribution`	A metric to track simple statistics from a distribution of values.
Class	`MapCounter`	A set of related counters keyed by an arbitrary string.
Class	`Metric`	Abstract base class for metrics.
Class	`MetricsContext`	A context manager that configures metrics and writes their output.
Class	`ReentrantStopWatch`	A watch that supports being called multiple times and recursively.
Class	`Snapshot`	A metric to track memory usage via tracemalloc snapshots.
Class	`StopWatch`	A counter that measures the time spent in a "with" statement.
Function	`dump_all`	Write a list of metrics to a json file.
Function	`get_cpu_clock`	Returns CPU clock to keep compatibility with various Python versions.
Function	`get_metric`	Return an existing metric or create a new one for the given name.
Function	`get_report`	Return a string listing all metrics, one per line.
Function	`load_all`	Read a list of metrics from a json file.
Function	`merge_from_file`	Merge metrics recorded in another file into the current metrics.
Class	`_RegistryMeta`	Metaclass that registers subclasses in _METRIC_TYPES.
Function	`_deserialize`	Construct a Metric from a typename and payload loaded from json.
Function	`_prepare_for_test`	Setup metrics collection for a test.
Function	`_serialize`	Return a json-serializable form of object.
Function	`_validate_metric_name`	Undocumented
Constant	`_METRIC_NAME_RE`	Undocumented
Constant	`_METRIC_TYPES`	Undocumented
Variable	`_enabled`	Undocumented
Variable	`_platform_timer`	Undocumented
Variable	`_registered_metrics`	Undocumented

def dump_all(objs, fp): (source) ¶

Write a list of metrics to a json file.

def get_cpu_clock(): (source) ¶

Returns CPU clock to keep compatibility with various Python versions.

def get_metric(name, constructor, *args, **kwargs): (source) ¶

Return an existing metric or create a new one for the given name. Args: name: The name of the metric. constructor: A class to instantiate if a new metric is required. *args: Additional positional args to pass to the constructor. **kwargs: Keyword args for the constructor. Returns: The current metric registered to name, or a new one created by invoking constructor(name, *args, **kwargs).

def get_report(): (source) ¶

Return a string listing all metrics, one per line.

def load_all(fp): (source) ¶

Read a list of metrics from a json file.

def merge_from_file(metrics_file): (source) ¶

Merge metrics recorded in another file into the current metrics.

def _deserialize(typ, payload): (source) ¶

Construct a Metric from a typename and payload loaded from json.

def _prepare_for_test(enabled=True): (source) ¶

Setup metrics collection for a test.

def _serialize(obj): (source) ¶

Return a json-serializable form of object.