bokeh.io¶

bokeh.io.doc¶

curdoc()[source]¶

Return the document for the current default state.

Returns

the current default document object.

Return type

Document

set_curdoc(doc)[source]¶

Configure the current document (returned by curdoc()).

Parameters

doc (Document) – Document we will output.

Returns

None

Warning

Calling this function will replace any existing document.

bokeh.io.export¶

create_webdriver()[source]¶

Create a new webdriver.

Note

Here for compatibility. Prefer methods on the webdriver_control object.

export_png(obj, filename=None, height=None, width=None, webdriver=None, timeout=5)[source]¶

Export the LayoutDOM object or document as a PNG.

If the filename is not given, it is derived from the script name (e.g. /foo/myplot.py will create /foo/myplot.png)

Parameters

• obj (LayoutDOM or Document) – a Layout (Row/Column), Plot or Widget object or Document to export.

• filename (str, optional) – filename to save document under (default: None) If None, infer from the filename.

• height (int) – the desired height of the exported layout obj only if it’s a Plot instance. Otherwise the height kwarg is ignored.

• width (int) – the desired width of the exported layout obj only if it’s a Plot instance. Otherwise the width kwarg is ignored.

• webdriver (selenium.webdriver) – a selenium webdriver instance to use to export the image.

• timeout (int) – the maximum amount of time (in seconds) to wait for Bokeh to initialize (default: 5) (Added in 1.1.1).

Returns

the filename where the static file is saved.

Return type

filename (str)

If you would like to access an Image object directly, rather than save a file to disk, use the lower-level get_screenshot_as_png() function.

Warning

Responsive sizing_modes may generate layouts with unexpected size and aspect ratios. It is recommended to use the default fixed sizing mode.

export_svgs(obj, filename=None, height=None, width=None, webdriver=None, timeout=5)[source]¶

Export the SVG-enabled plots within a layout. Each plot will result in a distinct SVG file.

If the filename is not given, it is derived from the script name (e.g. /foo/myplot.py will create /foo/myplot.svg)

Parameters

• obj (LayoutDOM object) – a Layout (Row/Column), Plot or Widget object to display

• filename (str, optional) – filename to save document under (default: None) If None, infer from the filename.

• height (int) – the desired height of the exported layout obj only if it’s a Plot instance. Otherwise the height kwarg is ignored.

• width (int) – the desired width of the exported layout obj only if it’s a Plot instance. Otherwise the width kwarg is ignored.

• webdriver (selenium.webdriver) – a selenium webdriver instance to use to export the image.

• timeout (int) – the maximum amount of time (in seconds) to wait for Bokeh to initialize (default: 5) (Added in 1.1.1).

Returns

the list of filenames where the SVGs files are saved.

Return type

filenames (list(str))

Warning

Responsive sizing_modes may generate layouts with unexpected size and aspect ratios. It is recommended to use the default fixed sizing mode.

get_layout_html(obj, resources=<bokeh.resources.Resources object>, **kwargs)[source]¶

get_screenshot_as_png(obj, driver=None, timeout=5, **kwargs)[source]¶

Get a screenshot of a LayoutDOM object.

Parameters

• obj (LayoutDOM or Document) – a Layout (Row/Column), Plot or Widget object or Document to export.

• driver (selenium.webdriver) – a selenium webdriver instance to use to export the image.

• timeout (int) – the maximum amount of time to wait for initialization. It will be used as a timeout for loading Bokeh, then when waiting for the layout to be rendered.

Returns

a pillow image loaded from PNG.

Return type

cropped_image (PIL.Image.Image)

Warning

Responsive sizing_modes may generate layouts with unexpected size and aspect ratios. It is recommended to use the default fixed sizing mode.

get_svgs(obj, driver=None, timeout=5, **kwargs)[source]¶

bokeh.io.notebook¶

class CommsHandle(comms, cell_doc)[source]¶

destroy_server(server_id)[source]¶

Given a UUID id of a div removed or replaced in the Jupyter notebook, destroy the corresponding server sessions and stop it.

get_comms(target_name)[source]¶

Create a Jupyter comms object for a specific target, that can be used to update Bokeh documents in the Jupyter notebook.

Parameters

target_name (str) – the target name the Comms object should connect to

Returns

Jupyter Comms

install_notebook_hook(notebook_type, load, show_doc, show_app, overwrite=False)[source]¶

Install a new notebook display hook.

Bokeh comes with support for Jupyter notebooks built-in. However, there are other kinds of notebooks in use by different communities. This function provides a mechanism for other projects to instruct Bokeh how to display content in other notebooks.

This function is primarily of use to developers wishing to integrate Bokeh with new notebook types.

Parameters

• notebook_type (str) –

  A name for the notebook type, e.e. 'Jupyter' or 'Zeppelin'

  If the name has previously been installed, a RuntimeError will be raised, unless overwrite=True

• load (callable) –

  A function for loading BokehJS in a notebook type. The function will be called with the following arguments:

  load(
      resources,   # A Resources object for how to load BokehJS
      verbose,     # Whether to display verbose loading banner
      hide_banner, # Whether to hide the output banner entirely
      load_timeout # Time after which to report a load fail error
  )
  

• show_doc (callable) –

  A function for displaying Bokeh standalone documents in the notebook type. This function will be called with the following arguments:

  show_doc(
      obj,            # the Bokeh object to display
      state,          # current bokeh.io "state"
      notebook_handle # whether a notebook handle was requested
  )
  

  If the notebook platform is capable of supporting in-place updates to plots then this function may return an opaque notebook handle that can be used for that purpose. The handle will be returned by show(), and can be used by as appropriate to update plots, etc. by additional functions in the library that installed the hooks.

• show_app (callable) –

  A function for displaying Bokeh applications in the notebook type. This function will be called with the following arguments:

  show_app(
      app,          # the Bokeh Application to display
      state,        # current bokeh.io "state"
      notebook_url, # URL to the current active notebook page
      **kw          # any backend-specific keywords passed as-is
  )
  

• overwrite (bool, optional) – Whether to allow an existing hook to be overwritten by a new definition (default: False)

Returns

None

Raises

RuntimeError – If notebook_type is already installed and overwrite=False

install_jupyter_hooks()[source]¶

load_notebook(resources=None, verbose=False, hide_banner=False, load_timeout=5000)[source]¶

Prepare the IPython notebook for displaying Bokeh plots.

Parameters

• resources (Resource, optional) – how and where to load BokehJS from (default: CDN)

• verbose (bool, optional) – whether to report detailed settings (default: False)

• hide_banner (bool, optional) – whether to hide the Bokeh banner (default: False)

• load_timeout (int, optional) – Timeout in milliseconds when plots assume load timed out (default: 5000)

Warning

Clearing the output cell containing the published BokehJS resources HTML code may cause Bokeh CSS styling to be removed.

Returns

None

publish_display_data(*args, **kw)[source]¶

push_notebook(document=None, state=None, handle=None)[source]¶

Update Bokeh plots in a Jupyter notebook output cells with new data or property values.

When working the the notebook, the show function can be passed the argument notebook_handle=True, which will cause it to return a handle object that can be used to update the Bokeh output later. When push_notebook is called, any property updates (e.g. plot titles or data source values, etc.) since the last call to push_notebook or the original show call are applied to the Bokeh output in the previously rendered Jupyter output cell.

Several example notebooks can be found in the GitHub repository in the examples/howto/notebook_comms directory.

Parameters

• document (Document, optional) – A Document to push from. If None, uses curdoc(). (default: None)

• state (State, optional) – A State object. If None, then the current default state (set by output_file, etc.) is used. (default: None)

Returns

None

Examples

Typical usage is typically similar to this:

from bokeh.plotting import figure
from bokeh.io import output_notebook, push_notebook, show

output_notebook()

plot = figure()
plot.circle([1,2,3], [4,6,5])

handle = show(plot, notebook_handle=True)

# Update the plot title in the earlier cell
plot.title.text = "New Title"
push_notebook(handle=handle)

run_notebook_hook(notebook_type, action, *args, **kw)[source]¶

Run an installed notebook hook with supplied arguments.

Parameters

• noteboook_type (str) – Name of an existing installed notebook hook

• actions (str) – Name of the hook action to execute, 'doc' or 'app'

All other arguments and keyword arguments are passed to the hook action exactly as supplied.

Returns

Result of the hook action, as-is

Raises

RuntimeError – If the hook or specific action is not installed

show_app(app, state, notebook_url, port=0, **kw)[source]¶

Embed a Bokeh server application in a Jupyter Notebook output cell.

Parameters

• app (Application or callable) – A Bokeh Application to embed inline in a Jupyter notebook.

• state (State) – ** Unused **

• notebook_url (str or callable) –

  The URL of the notebook server that is running the embedded app.

  If notebook_url is a string, the value string is parsed to construct the origin and full server URLs.

  If notebook_url is a callable, it must accept one parameter, which will be the server port, or None. If passed a port, the callable must generate the server URL, otherwise if passed None, it must generate the origin URL for the server.

• port (int) –

  A port for the embedded server will listen on.

  By default the port is 0, which results in the server listening on a random dynamic port.

Any additional keyword arguments are passed to Server (added in version 1.1)

Returns

None

show_doc(obj, state, notebook_handle)[source]¶

bokeh.io.output¶

output_file(filename, title='Bokeh Plot', mode=None, root_dir=None)[source]¶

Configure the default output state to generate output saved to a file when show() is called.

Does not change the current Document from curdoc(). File and notebook output may be active at the same time, so e.g., this does not clear the effects of output_notebook().

Parameters

• filename (str) – a filename for saving the HTML document

• title (str, optional) – a title for the HTML document (default: “Bokeh Plot”)

• mode (str, optional) – how to include BokehJS (default: 'cdn') One of: 'inline', 'cdn', 'relative(-dev)' or 'absolute(-dev)'. See bokeh.resources.Resources for more details.

• root_dir (str, optional) – root directory to use for ‘absolute’ resources. (default: None) This value is ignored for other resource types, e.g. INLINE or CDN.

Returns

None

Note

Generally, this should be called at the beginning of an interactive session or the top of a script.

Warning

This output file will be overwritten on every save, e.g., each time show() or save() is invoked.

output_notebook(resources=None, verbose=False, hide_banner=False, load_timeout=5000, notebook_type='jupyter')[source]¶

Configure the default output state to generate output in notebook cells when show() is called. Note that, show() may be called multiple times in a single cell to display multiple objects in the output cell. The objects will be displayed in order.

Parameters

• resources (Resource, optional) – How and where to load BokehJS from (default: CDN)

• verbose (bool, optional) – whether to display detailed BokehJS banner (default: False)

• hide_banner (bool, optional) – whether to hide the Bokeh banner (default: False)

• load_timeout (int, optional) – Timeout in milliseconds when plots assume load timed out (default: 5000)

• notebook_type (string, optional) – Notebook type (default: jupyter)

Returns

None

Note

Generally, this should be called at the beginning of an interactive session or the top of a script.

reset_output(state=None)[source]¶

Clear the default state of all output modes.

Returns

None

bokeh.io.saving¶

save(obj, filename=None, resources=None, title=None, template=None, state=None, **kwargs)[source]¶

Save an HTML file with the data for the current document.

Will fall back to the default output state (or an explicitly provided State object) for filename, resources, or title if they are not provided. If the filename is not given and not provided via output state, it is derived from the script name (e.g. /foo/myplot.py will create /foo/myplot.html)

Parameters

• obj (LayoutDOM object) – a Layout (Row/Column), Plot or Widget object to display

• filename (str, optional) – filename to save document under (default: None) If None, use the default state configuration.

• resources (Resources, optional) – A Resources config to use (default: None) If None, use the default state configuration, if there is one. otherwise use resources.INLINE.

• title (str, optional) –

  a title for the HTML document (default: None)

  If None, use the default state title value, if there is one. Otherwise, use “Bokeh Plot”

  state (State, optional) :

  A State object. If None, then the current default implicit state is used. (default: None).

Returns

the filename where the HTML file is saved.

Return type

str

bokeh.io.showing¶

show(obj, browser=None, new='tab', notebook_handle=False, notebook_url='localhost:8888', **kw)[source]¶

Immediately display a Bokeh object or application.

show() may be called multiple times in a single Jupyter notebook cell to display multiple objects. The objects are displayed in order.

Parameters

• obj (LayoutDOM or Application or callable) –

  A Bokeh object to display.

  Bokeh plots, widgets, layouts (i.e. rows and columns) may be passed to show in order to display them. When output_file has been called, the output will be to an HTML file, which is also opened in a new browser window or tab. When output_notebook has been called in a Jupyter notebook, the output will be inline in the associated notebook output cell.

  In a Jupyter notebook, a Bokeh application or callable may also be passed. A callable will be turned into an Application using a FunctionHandler. The application will be run and displayed inline in the associated notebook output cell.

• browser (str, optional) –

  Specify the browser to use to open output files(default: None)

  For file output, the browser argument allows for specifying which browser to display in, e.g. “safari”, “firefox”, “opera”, “windows-default”. Not all platforms may support this option, see the documentation for the standard library webbrowser module for more information

• new (str, optional) –

  Specify the browser mode to use for output files (default: “tab”)

  For file output, opens or raises the browser window showing the current output file. If new is ‘tab’, then opens a new tab. If new is ‘window’, then opens a new window.

• notebook_handle (bool, optional) –

  Whether to create a notebook interaction handle (default: False)

  For notebook output, toggles whether a handle which can be used with push_notebook is returned. Note that notebook handles only apply to standalone plots, layouts, etc. They do not apply when showing Applications in the notebook.

• notebook_url (URL, optional) –

  Location of the Jupyter notebook page (default: “localhost:8888”)

  When showing Bokeh applications, the Bokeh server must be explicitly configured to allow connections originating from different URLs. This parameter defaults to the standard notebook host and port. If you are running on a different location, you will need to supply this value for the application to display properly. If no protocol is supplied in the URL, e.g. if it is of the form “localhost:8888”, then “http” will be used.

  notebook_url can also be a function that takes one int for the bound server port. If the port is provided, the function needs to generate the full public URL to the bokeh server. If None is passed, the function is to generate the origin URL.

Some parameters are only useful when certain output modes are active:

• The browser and new parameters only apply when output_file is active.

• The notebook_handle parameter only applies when output_notebook is active, and non-Application objects are being shown. It is only supported to Jupyter notebook, raise exception for other notebook types when it is True.

• The notebook_url parameter only applies when showing Bokeh Applications in a Jupyter notebook.

• Any additional keyword arguments are passed to Server when showing a Bokeh app (added in version 1.1)

Returns

When in a Jupyter notebook (with output_notebook enabled) and notebook_handle=True, returns a handle that can be used by push_notebook, None otherwise.

bokeh.io.state¶

Encapsulate implicit state that is useful for Bokeh plotting APIs.

Generating output for Bokeh plots requires coordinating several things:

Document

Groups together Bokeh models that may be shared between plots (e.g., range or data source objects) into one common strucure.

Resources

Control how JavaScript and CSS for the client library BokehJS are included and used in the generated output.

It is possible to handle the configuration of these things manually, and some examples of doing this can be found in examples/models directory. When developing sophisticated applications, it may be necessary or desirable to work at this level. However, for general use this would quickly become burdensome. This module provides a State class that encapsulates these objects and ensures their proper configuration in many common usage scenarios.

class State[source]¶

Manage state related to controlling Bokeh output.

output_file(filename, title='Bokeh Plot', mode=None, root_dir=None)[source]¶

Configure output to a standalone HTML file.

Calling output_file not clear the effects of any other calls to output_notebook, etc. It adds an additional output destination (publishing to HTML files). Any other active output modes continue to be active.

Parameters

• filename (str) – a filename for saving the HTML document

• title (str, optional) – a title for the HTML document

• mode (str, optional) –

  how to include BokehJS (default: 'cdn')

  One of: 'inline', 'cdn', 'relative(-dev)' or 'absolute(-dev)'. See Resources for more details.

• root_dir (str, optional) –

  root dir to use for absolute resources (default: None)

  This value is ignored for other resource types, e.g. INLINE or CDN.

Warning

The specified output file will be overwritten on every save, e.g., every time show() or save() is called.

output_notebook(notebook_type='jupyter')[source]¶

Generate output in notebook cells.

Calling output_notebook not clear the effects of any other calls to output_file, etc. It adds an additional output destination (publishing to notebook output cells). Any other active output modes continue to be active.

Returns

None

reset()[source]¶

Deactivate all currently active output modes and set curdoc() to a fresh empty Document.

Subsequent calls to show() will not render until a new output mode is activated.

Returns

None

property document¶

A default Document to use for all output operations.

property file¶

A dict with the default configuration for file output (READ ONLY)

The dictionary value has the following form:

{
    'filename'  : # filename to use when saving
    'resources' : # resources configuration
    'title'     : # a title for the HTML document
}

property notebook¶

Whether to generate notebook output on show operations. (READ ONLY)

property notebook_type¶

Notebook type

curstate()[source]¶

Return the current State object

Returns

the current default State object

Return type

State

bokeh.io.util¶

default_filename(ext)[source]¶

Generate a default filename with a given extension, attempting to use the filename of the currently running process, if possible.

If the filename of the current process is not available (or would not be writable), then a temporary file with the given extension is returned.

Parameters

ext (str) – the desired extension for the filename

Returns

str

Raises

RuntimeError – If the extensions requested is “.py”

detect_current_filename()[source]¶

Attempt to return the filename of the currently running Python process

Returns None if the filename cannot be detected.

temp_filename(ext)[source]¶

Generate a temporary, writable filename with the given extension