deploying the bokeh server

29 January 2016

overview

  1. provide

    1. publishing bokeh plots for wider audiences

    2. streaming data to automatically updating plots

    3. interactively visualizing very large datasets

    4. building dashboards and apps with sophisticated interactions

  2. built on top of flask

concepts

  1. core architecture of bokeh-server

    1. Document

    2. User

running

  1. if bokeh was installed by setup.py or conda

         $ bokeh-server

  2. development mode

         $ python ./bokeh-server

storing plots on the server

  1. demo

         from bokeh.plotting import figure, output_server, show

     output_server('line')

     p = figure(plot_width=400, plot_height=400)

     # add a line renderer
     p.line([1, 2, 3, 4, 5], [6, 7, 2, 4, 5], line_width=2)

     show(p)

  2. publishing to the server

    1. running in multi user mode

       output_server('myplot')

 # make some plots
 cursession().publish()

streaming data with the server

  1. steps

    1. write code demo.py

       import time
 from random import shuffle
 from bokeh.plotting import figure, output_server, cursession, show

 # prepare output to server
 output_server('animated_line')

 p = figure(plot_width=400, plot_height=400)
 p.line([1, 2, 3, 4, 5], [6, 7, 2, 4, 5], name='ex_line')
 show(p)

 # create some simple animation..
 # first get our figure example data source
 renderer = p.select(dict(name='ex_line'))
 ds = renderer[0].data_source

 while True:
     # update y data of the source object
     shuffle(ds.data['y'])

     # store the updated source on the server
     cursession().store_objects(ds)
     time.sleep(0.5)

    2. start bokeh-server

       $ bokeh-server

 Bokeh Server Configuration
 ==========================
 python version : 2.7.11
 bokeh version  : 0.10.0
 listening      : 127.0.0.1:5006
 backend        : memory
 python options : debug:OFF, verbose:OFF, filter-logs:OFF, multi-user:OFF
 js options     : debugjs:OFF

    3. run demo.py

       $ python demo.py
 Using saved session configuration for http://localhost:5006/
 To override, pass 'load_from_config=False' to Session

    4. from bokeh-server side you will see

       2016-01-30 18:13:27,521:INFO:tornado.access:200 GET /bokeh/userinfo/ (127.0.0.1) 1.52ms
 2016-01-30 18:13:27,547:INFO:tornado.access:200 POST /bokeh/doc/ (127.0.0.1) 2.18ms
 2016-01-30 18:13:27,550:INFO:tornado.access:200 GET /bokeh/getdocapikey/797be60d-03f7-4558-a60f-6b6b904d5960 (127.0.0.1) 1.12ms
 2016-01-30 18:13:27,570:INFO:tornado.access:200 POST /bokeh/bb/797be60d-03f7-4558-a60f-6b6b904d5960/gc (127.0.0.1) 17.87ms
 2016-01-30 18:13:27,574:INFO:tornado.access:200 GET /bokeh/bb/797be60d-03f7-4558-a60f-6b6b904d5960/ (127.0.0.1) 1.37ms
 2016-01-30 18:13:27,630:INFO:tornado.access:200 POST /bokeh/bb/797be60d-03f7-4558-a60f-6b6b904d5960/bulkupsert (127.0.0.1) 26.75ms
 2016-01-30 18:13:27,739:INFO:tornado.access:200 POST /bokeh/bb/797be60d-03f7-4558-a60f-6b6b904d5960/bulkupsert (127.0.0.1) 18.40ms
 2016-01-30 18:13:27,991:INFO:tornado.access:200 GET /bokeh/doc/797be60d-03f7-4558-a60f-6b6b904d5960/c4c6deb9-2c17-4370-93bf-9e1a8a81e9f9 (127.0.0.1) 17.51ms
 2016-01-30 18:13:28,263:INFO:tornado.access:200 POST /bokeh/bb/797be60d-03f7-4558-a60f-6b6b904d5960/bulkupsert (127.0.0.1) 15.45ms
 2016-01-30 18:13:28,441:INFO:tornado.access:200 GET /bokeh/wsurl/ (127.0.0.1) 0.97ms
 2016-01-30 18:13:28,489:INFO:tornado.access:200 GET /bokeh/objinfo/797be60d-03f7-4558-a60f-6b6b904d5960/c4c6deb9-2c17-4370-93bf-9e1a8a81e9f9 (127.0.0.1) 18.74ms
 2016-01-30 18:13:28,784:INFO:tornado.access:200 POST /bokeh/bb/797be60d-03f7-4558-a60f-6b6b904d5960/bulkupsert (127.0.0.1) 14.50ms
 2016-01-30 18:13:29,304:INFO:tornado.access:200 POST /bokeh/bb/797be60d-03f7-4558-a60f-6b6b904d5960/bulkupsert (127.0.0.1) 14.42ms
 2016-01-30 18:13:29,826:INFO:tornado.access:200 POST /bokeh/bb/797be60d-03f7-4558-a60f-6b6b904d5960/bulkupsert (127.0.0.1) 14.21ms
 ...

fixed

  1. but if your bokeh’s version is higher than 0.10.0

    1. you can’t run bokeh-server

       $ bokeh-server

 The 'bokeh-server' command from versions 0.10 and earlier no longer exists.
 Please see the documentation for the new 'bokeh' command here:

 http://bokeh.pydata.org/en/latest/docs/user_guide/cli.html

    2. you can’t import cursession from bokeh.plotting

       # so you must downgrade your bokeh
 $ conda install bokeh=0.10.0

  2. you can run bokeh with bokeh’s version higher than 0.10.0

    1. bokeh -h

       $ bokeh -h
 usage: /Users/hqlgree2/anaconda/bin/bokeh [-h] [-v]
                                           {html,json,secret,serve} ...

 positional arguments:
   {html,json,secret,serve}
                         Sub-commands
     html                Create standalone HTML files for one or more
                         applications
     json                Create JSON files for one or more applications
     secret              Create a Bokeh secret key for use with Bokeh server
     serve               Run a Bokeh server hosting one or more applications

 optional arguments:
   -h, --help            show this help message and exit
   -v, --version         show program's version number and exit

    2. bokeh serve

       $ bokeh serve
 DEBUG:bokeh.server.tornado:Allowed Host headers: ['localhost:5006']
 DEBUG:bokeh.server.tornado:These host origins can connect to the websocket: ['localhost:5006']
 DEBUG:bokeh.server.tornado:Patterns are: [('/?', <class 'bokeh.server.views.doc_handler.DocHandler'>, {'application_context': <bokeh.server.application_context.ApplicationContext object at 0x106a2f250>, 'bokeh_websocket_path': '/ws'}), ('/ws', <class 'bokeh.server.views.ws.WSHandler'>, {'application_context': <bokeh.server.application_context.ApplicationContext object at 0x106a2f250>, 'bokeh_websocket_path': '/ws'}), ('/autoload.js', <class 'bokeh.server.views.autoload_js_handler.AutoloadJsHandler'>, {'application_context': <bokeh.server.application_context.ApplicationContext object at 0x106a2f250>, 'bokeh_websocket_path': '/ws'}), ('/static/(.*)', <class 'bokeh.server.views.static_handler.StaticHandler'>)]
 INFO:bokeh.command.subcommands.serve:Starting Bokeh server on port 5006 with applications at paths ['/']
 DEBUG:bokeh.server.tornado:[pid 17900] 0 clients connected
blog comments powered by Disqus