Integration with python/pytest logging functionality

#### I'm submitting a ...
- [x] feature request

#### What is the current behavior?
logger calls from python logging module are not shown on the report.

#### What is the expected behavior?
they be shown on the report

#### What is the motivation / use case for changing the behavior?
the report should contain as much info as possible. Log calls are very relevant information.

#### Please tell us about your environment:

- Allure version: 2.6.0
- Test framework: pytest@3.4
- Allure adaptor: allure-pytest@2.3.2b1

#### Other information

I actually have a simple working workaround for this, by defining a custom logging handler for Allure (you can use it on conftest.py or as a plugin):

```
import logging
import allure

class AllureLoggingHandler(logging.Handler):
def log(self, message):
with allure.step('Log {}'.format(message)):
pass

def emit(self, record):
self.log("({}) {}".format(record.levelname, record.getMessage()))

class AllureCatchLogs:
def __init__(self):
self.rootlogger = logging.getLogger()
self.allurehandler = AllureLoggingHandler()
def __enter__(self):
if self.allurehandler not in self.rootlogger.handlers:
self.rootlogger.addHandler(self.allurehandler)
def __exit__(self, exc_type, exc_value, traceback):
self.rootlogger.removeHandler(self.allurehandler)

@pytest.hookimpl(hookwrapper=True)
def pytest_runtest_setup():
with AllureCatchLogs():
yield

@pytest.hookimpl(hookwrapper=True)
def pytest_runtest_call():
with AllureCatchLogs():
yield

@pytest.hookimpl(hookwrapper=True)
def pytest_runtest_teardown():
with AllureCatchLogs():
yield
```

Then you can test with a simple test:

```
def test_print():
logging.info("Logging an info message")
logging.debug("Logging a DEBUG message")
logging.warning("Sample time is too low!")
raise Exception
```
It can be added for the test like this example or easily added also in the config section of a custom pytest plugin. Not sure if this could be merged into allure pytest plugin though.

It is still missing more complete formatting (module name). Also the handler log level is not dynamically adjusted as done in the pytest logging plugin, but it seems to work fine like this because pytest sets both the handler and the logger to the same level.

It even works with pytest logging switches. For instance, if we run this test with --log_level=INFO, the debug log record won't appear:


首先查看 issue 中的 pytest 集成点和 conftest.py 的 hookwrapper 示例，然后跟踪测试生命周期事件如何成为 Allure 报告数据。使用提供的 logging handler 和示例测试，检查 Python logging 记录是否会出现在报告中，同时遵循 pytest 的日志级别开关；完成标准是集成无需自定义 handler 即可正常工作。

由索引模型根据 Issue 内容生成。