wsgi-lineprof 0.5.0

WSGI middleware for line-by-line profiling

wsgi_lineprof is a WSGI middleware for line-by-line profiling.

wsgi_lineprof shows results of line-by-line profiling per request. You can use this project with many WSGI-compatible applications and frameworks:

• Django

• Pyramid

• Flask

• Bottle

• etc…

At a Glance

You can use wsgi_lineprof as a WSGI middleware of existing applications.

$ pip install wsgi_lineprof

Example usage with Bottle:

import time

import bottle
from wsgi_lineprof.middleware import LineProfilerMiddleware

app = bottle.default_app()


@app.route('/')
def index():
    time.sleep(1)
    return "Hello world!!"

if __name__ == "__main__":
    # Add wsgi_lineprof as a WSGI middleware!
    app = LineProfilerMiddleware(app)
    bottle.run(app=app)

Run the above script to start web server, then access http://127.0.0.1:8080.

The results are outputted to stdout by default. You can see the results like this:

... (snip) ...

File: ./app.py
Name: index
Total time: 1.00518 [sec]
  Line      Hits         Time  Code
===================================
     9                         @app.route('/')
    10                         def index():
    11         1      1005175      time.sleep(1)
    12         1            4      return "Hello world!!"

... (snip) ...

Results contain many other functions, you can remove unnecessary results by using filters.

Requirements

• Python 2.7

• Python 3.4

• Python 3.5

• Python 3.6

• Python 3.7

Filters

You can get results from specific files or sort results by using filters. For example, use FilenameFilter to filter results with filename and use TotalTimeSorter to sort results by total_time.

import time

import bottle
from wsgi_lineprof.filters import FilenameFilter, TotalTimeSorter
from wsgi_lineprof.middleware import LineProfilerMiddleware

app = bottle.default_app()


def get_name():
    # Get some data...
    time.sleep(1)
    return "Monty Python"

@app.route('/')
def index():
    name = get_name()
    return "Hello, {}!!".format(name)

if __name__ == "__main__":
    filters = [
        # Results which filename contains "app2.py"
        FilenameFilter("app2.py"),
        # Sort by total time of results
        TotalTimeSorter(),
    ]
    # Add wsgi_lineprof as a WSGI middleware
    app = LineProfilerMiddleware(app, filters=filters)

    bottle.run(app=app)

Run the above script to start web server, then access http://127.0.0.1:8080. You can see results in stdout.

$ ./app2.py
Bottle v0.12.10 server starting up (using WSGIRefServer())...
Listening on http://127.0.0.1:8080/
Hit Ctrl-C to quit.

Time unit: 1e-06 [sec]

File: ./app2.py
Name: index
Total time: 1.00526 [sec]
  Line      Hits         Time  Code
===================================
    15                         @app.route('/')
    16                         def index():
    17         1      1005250      name = get_name()
    18         1           11      return "Hello, {}!!".format(name)

File: ./app2.py
Name: get_name
Total time: 1.00523 [sec]
  Line      Hits         Time  Code
===================================
    10                         def get_name():
    11                             # Get some data...
    12         1      1005226      time.sleep(1)
    13         1            4      return "Monty Python"

127.0.0.1 - - [30/Nov/2016 17:21:12] "GET / HTTP/1.1" 200 21

There are some useful filters in wsgi_lineprof.filters.

Stream

By using stream option, you can output results to a file. For example, you can output logs to lineprof.log.

f = open("lineprof.log", "w")
app = LineProfilerMiddleware(app, stream=f)
bottle.run(app=app)

Async Stream

By using async_stream option, wsgi_lineprof starts a new thread for writing results. This option is useful when you do not want the main thread blocked for writing results.

# Start a new thread for writing results
app = LineProfilerMiddleware(app, async_stream=True)
bottle.run(app=app)

Links

• GitHub: ymyzk/wsgi_lineprof

• WSGI ミドルウェアとして使えるラインプロファイラを作った話 – ymyzk’s blog

Special Thanks

This project uses code from the following projects:

• kainosnoema/rack-lineprof

This project is inspired by the following projects:

• rkern/line_profiler

wsgi_lineprof is integrated with the following projects:

• kobinpy/wsgicli