1
0
mirror of https://github.com/ARM-software/devlib.git synced 2024-10-05 18:30:50 +01:00

docs: Update to include Collector information

This commit is contained in:
Marc Bonnici 2019-12-05 16:19:19 +00:00
parent a4fd57f023
commit b72fb470e7
5 changed files with 179 additions and 17 deletions

153
doc/collectors.rst Normal file
View File

@ -0,0 +1,153 @@
.. _collector:
Collectors
==========
The ``Collector`` API provide a consistent way of collecting arbitrary data from
a target. Data is collected via an instance of a class derived from
:class:`CollectorBase`.
Example
-------
The following example shows how to use a collector to read the logcat output
from an Android target.
.. code-block:: python
# import and instantiate the Target and the collector
# (note: this assumes exactly one android target connected
# to the host machine).
In [1]: from devlib import AndroidTarget, LogcatCollector
In [2]: t = AndroidTarget()
# Set up the collector on the Target.
In [3]: collector = LogcatCollector(t)
# Configure the output file path for the collector to use.
In [4]: collector.set_output('adb_log.txt')
# Reset the Collector to preform any required configuration or preparation.
In [5]: collector.reset()
# Start Collecting
In [6]: collector.start()
# Wait for some output to be generated
In [7]: sleep(10)
# Stop Collecting
In [8]: collector.stop()
# Retrieved the collected data
In [9]: output = collector.get_data()
# Display the returned ``CollectorOutput`` Object.
In [10]: output
Out[10]: [<adb_log.txt (file)>]
In [11] log_file = output[0]
# Get the path kind of the the returned CollectorOutputEntry.
In [12]: log_file.path_kind
Out[12]: 'file'
# Get the path of the returned CollectorOutputEntry.
In [13]: log_file.path
Out[13]: 'adb_log.txt'
# Find the full path to the log file.
In [14]: os.path.join(os.getcwd(), logfile)
Out[14]: '/tmp/adb_log.txt'
API
---
.. collector:
.. module:: devlib.collector
CollectorBase
~~~~~~~~~~~~~
.. class:: CollectorBase(target, \*\*kwargs)
A ``CollectorBase`` is the the base class and API that should be
implemented to allowing collecting various data from a traget e.g. traces,
logs etc.
.. method:: Collector.setup(\*args, \*\*kwargs)
This will set up the collector on the target. Parameters this method takes
are particular to subclasses (see documentation for specific collectors
below). What actions are performed by this method are also
collector-specific. Usually these will be things like installing
executables, starting services, deploying assets, etc. Typically, this method
needs to be invoked at most once per reboot of the target (unless
``teardown()`` has been called), but see documentation for the collector
you're interested in.
.. method:: CollectorBase.reset()
This can be used to configure a collector for collection. This must be invoked
before ``start()`` is called to begin collection.
.. method:: CollectorBase.start()
Starts collecting from the target.
.. method:: CollectorBase.stop()
Stops collecting from target. Must be called after
:func:`start()`.
.. method:: CollectorBase.set_output(output_path)
Configure the output path for the particular collector. This will be either
a directory or file path which will be used when storing the data. Please see
the individual Collector documentation for more information.
.. method:: CollectorBase.get_data()
The collected data will be return via the previously specified output_path.
This method will return a ``CollectorOutput`` object which is a subclassed
list object containing individual ``CollectorOutputEntry`` objects with details
about the individual output entry.
CollectorOutputEntry
~~~~~~~~~~~~~~~~~~~~
This object is designed to allow for the output of a collector to be processed
generically. The object will behave as a regular string containing the path to
underlying output path and can be used directly in ``os.path`` operations.
.. attribute:: CollectorOutputEntry.path
The file path for the corresponding output item.
.. attribute:: CollectorOutputEntry.path_kind
The type of output the is specified in the ``path`` attribute. Current valid
kinds are: ``file`` and ``directory``.
.. method:: CollectorOutputEntry.__init__(path, path_kind)
Initialises a ``CollectorOutputEntry`` object with the desired file path and
kind of file path specified.
.. collectors:
Available Collectors
---------------------
This section lists collectors that are currently part of devlib.
.. todo:: Add collectors

View File

@ -19,6 +19,7 @@ Contents:
target
modules
instrumentation
collectors
derived_measurements
platform
connection

View File

@ -1,3 +1,5 @@
.. _instrumentation:
Instrumentation
===============

View File

@ -6,8 +6,7 @@ There are currently four target interfaces:
- :class:`LinuxTarget` for interacting with Linux devices over SSH.
- :class:`AndroidTarget` for interacting with Android devices over adb.
- :class:`ChromeOsTarget`: for interacting with ChromeOS devices over SSH, and
their Android containers over adb.
- :class:`ChromeOsTarget`: for interacting with ChromeOS devices over SSH, and their Android containers over adb.
- :class:`LocalLinuxTarget`: for interacting with the local Linux host.
They all work in more-or-less the same way, with the major difference being in
@ -240,17 +239,17 @@ complete. Retrying it or bailing out is therefore a responsability of the caller
The hierarchy is as follows:
- :class:`DevlibError`
- :class:`WorkerThreadError`
- :class:`HostError`
- :class:`TargetError`
- :class:`TargetStableError`
- :class:`TargetTransientError`
- :class:`TargetNotRespondingError`
- :class:`DevlibStableError`
- :class:`TargetStableError`
- :class:`DevlibTransientError`
@ -307,12 +306,22 @@ has been successfully installed on a target, you can use ``has()`` method, e.g.
Please see the modules documentation for more detail.
Instruments and Collectors
--------------------------
Measurement and Trace
---------------------
You can retrieve multiple types of data from a target. There are two categories
of classes that allow for this:
You can collected traces (currently, just ftrace) using
:class:`TraceCollector`\ s. For example
- An :class:`Instrument` which may be used to collect measurements (such as power) from
targets that support it. Please see the
:ref:`instruments documentation <Instrumentation>` for more details.
- A :class:`Collector` may be used to collect arbitary data from a ``Target`` varying
from screenshots to trace data. Please see the
:ref:`collectors documentation <collector>` for more details.
An example workflow using :class:`FTraceCollector` is as follows:
.. code:: python
@ -326,23 +335,19 @@ You can collected traces (currently, just ftrace) using
# As a context manager, clear ftrace buffer using trace.reset(),
# start trace collection using trace.start(), then stop it Using
# trace.stop(). Using a context manager brings the guarantee that
# tracing will stop even if an exception occurs, including
# tracing will stop even if an exception occurs, including
# KeyboardInterrupt (ctr-C) and SystemExit (sys.exit)
with trace:
# Perform the operations you want to trace here...
import time; time.sleep(5)
# extract the trace file from the target into a local file
trace.get_trace('/tmp/trace.bin')
trace.get_data('/tmp/trace.bin')
# View trace file using Kernelshark (must be installed on the host).
trace.view('/tmp/trace.bin')
# Convert binary trace into text format. This would normally be done
# automatically during get_trace(), unless autoreport is set to False during
# automatically during get_data(), unless autoreport is set to False during
# instantiation of the trace collector.
trace.report('/tmp/trace.bin', '/tmp/trace.txt')
In a similar way, :class:`Instrument` instances may be used to collect
measurements (such as power) from targets that support it. Please see
instruments documentation for more details.

View File

@ -546,6 +546,7 @@ Target
:returns: ``True`` if internet seems available, ``False`` otherwise.
.. method:: Target.install_module(mod, **params)
:param mod: The module name or object to be installed to the target.
:param params: Keyword arguments used to instantiate the module.