#----------------------------------------------------------------------------- # Copyright (c) 2012 - 2020, Anaconda, Inc., and Bokeh Contributors. # All rights reserved. # # The full license is in the file LICENSE.txt, distributed with this software. #----------------------------------------------------------------------------- ''' Provide events that represent various changes to Bokeh Documents. These events are used internally to signal changes to Documents. For information about user-facing (e.g. UI or tool) events, see the reference for :ref:`bokeh.events`. ''' #----------------------------------------------------------------------------- # Boilerplate #----------------------------------------------------------------------------- import logging # isort:skip log = logging.getLogger(__name__) #----------------------------------------------------------------------------- # Imports #----------------------------------------------------------------------------- # Standard library imports from typing import Any, Union # Bokeh imports from ..util.dependencies import import_optional from ..util.serialization import make_id #----------------------------------------------------------------------------- # Globals and constants #----------------------------------------------------------------------------- pd = import_optional('pandas') __all__ = ( 'ColumnDataChangedEvent', 'ColumnsStreamedEvent', 'ColumnsPatchedEvent', 'DocumentChangedEvent', 'DocumentPatchedEvent', 'ModelChangedEvent', 'RootAddedEvent', 'RootRemovedEvent', 'SessionCallbackAdded', 'SessionCallbackRemoved', 'TitleChangedEvent', 'MessageSentEvent', ) #----------------------------------------------------------------------------- # General API #----------------------------------------------------------------------------- #----------------------------------------------------------------------------- # Dev API #----------------------------------------------------------------------------- [docs]class DocumentChangedEvent(object): ''' Base class for all internal events representing a change to a Bokeh Document. ''' [docs] def __init__(self, document, setter=None, callback_invoker=None): ''' Args: document (Document) : A Bokeh document that is to be updated. setter (ClientSession or ServerSession or None, optional) : This is used to prevent "boomerang" updates to Bokeh apps. (default: None) In the context of a Bokeh server application, incoming updates to properties will be annotated with the session that is doing the updating. This value is propagated through any subsequent change notifications that the update triggers. The session can compare the event setter to itself, and suppress any updates that originate from itself. callback_invoker (callable, optional) : A callable that will invoke any Model callbacks that should be executed in response to the change that triggered this event. (default: None) ''' self.document = document self.setter = setter self.callback_invoker = callback_invoker [docs] def combine(self, event): ''' ''' return False [docs] def dispatch(self, receiver): ''' Dispatch handling of this event to a receiver. This method will invoke ``receiver._document_changed`` if it exists. ''' if hasattr(receiver, '_document_changed'): receiver._document_changed(self) [docs]class DocumentPatchedEvent(DocumentChangedEvent): ''' A Base class for events that represent updating Bokeh Models and their properties. ''' [docs] def dispatch(self, receiver): ''' Dispatch handling of this event to a receiver. This method will invoke ``receiver._document_patched`` if it exists. ''' super().dispatch(receiver) if hasattr(receiver, '_document_patched'): receiver._document_patched(self) [docs] def generate(self, references, buffers): ''' Create a JSON representation of this event suitable for sending to clients. *Sub-classes must implement this method.* Args: references (dict[str, Model]) : If the event requires references to certain models in order to function, they may be collected here. **This is an "out" parameter**. The values it contains will be modified in-place. buffers (set) : If the event needs to supply any additional Bokeh protocol buffers, they may be added to this set. **This is an "out" parameter**. The values it contains will be modified in-place. ''' raise NotImplementedError() [docs]class MessageSentEvent(DocumentPatchedEvent): """ """ def __init__(self, document, msg_type: str, msg_data: Union[Any, bytes], setter=None, callback_invoker=None): super(MessageSentEvent, self).__init__(document, setter, callback_invoker) self.msg_type = msg_type self.msg_data = msg_data [docs] def dispatch(self, receiver): super(MessageSentEvent, self).dispatch(receiver) if hasattr(receiver, "_document_message_sent"): receiver._document_message_sent(self) [docs] def generate(self, references, buffers): msg = { "kind": "MessageSent", "msg_type": self.msg_type, } if not isinstance(self.msg_data, bytes): msg["msg_data"] = self.msg_data else: buffer_id = make_id() buf = (dict(id=buffer_id), self.msg_data) buffers.append(buf) return msg [docs]class ModelChangedEvent(DocumentPatchedEvent): ''' A concrete event representing updating an attribute and value of a specific Bokeh Model. This is the "standard" way of updating most Bokeh model attributes. For special casing situations that can optimized (e.g. streaming, etc.), a ``hint`` may be supplied that overrides normal mechanisms. ''' [docs] def __init__(self, document, model, attr, old, new, serializable_new, hint=None, setter=None, callback_invoker=None): ''' Args: document (Document) : A Bokeh document that is to be updated. model (Model) : A Model to update attr (str) : The name of the attribute to update on the model. old (object) : The old value of the attribute new (object) : The new value of the attribute serializable_new (object) : A serialized (JSON) version of the new value. It may be ``None`` if a hint is supplied. hint (DocumentPatchedEvent, optional) : When appropriate, a secondary event may be supplied that modifies the normal update process. For example, in order to stream or patch data more efficiently than the standard update mechanism. setter (ClientSession or ServerSession or None, optional) : This is used to prevent "boomerang" updates to Bokeh apps. (default: None) See :class:`~bokeh.document.events.DocumentChangedEvent` for more details. callback_invoker (callable, optional) : A callable that will invoke any Model callbacks that should be executed in response to the change that triggered this event. (default: None) ''' if setter is None and isinstance(hint, (ColumnsStreamedEvent, ColumnsPatchedEvent)): setter = hint.setter super().__init__(document, setter, callback_invoker) self.model = model self.attr = attr self.old = old self.new = new self.serializable_new = serializable_new self.hint = hint [docs] def combine(self, event): ''' ''' if not isinstance(event, ModelChangedEvent): return False # If these are not true something weird is going on, maybe updates from # Python bokeh.client, don't try to combine if self.setter != event.setter: return False if self.document != event.document: return False if self.hint: return self.hint.combine(event.hint) if (self.model == event.model) and (self.attr == event.attr): self.new = event.new self.serializable_new = event.serializable_new self.callback_invoker = event.callback_invoker return True return False [docs] def dispatch(self, receiver): ''' Dispatch handling of this event to a receiver. This method will invoke ``receiver._document_model_changed`` if it exists. ''' super().dispatch(receiver) if hasattr(receiver, '_document_model_changed'): receiver._document_model_changed(self) [docs] def generate(self, references, buffers): ''' Create a JSON representation of this event suitable for sending to clients. Args: references (dict[str, Model]) : If the event requires references to certain models in order to function, they may be collected here. **This is an "out" parameter**. The values it contains will be modified in-place. buffers (set) : If the event needs to supply any additional Bokeh protocol buffers, they may be added to this set. **This is an "out" parameter**. The values it contains will be modified in-place. ''' from ..model import collect_models if self.hint is not None: return self.hint.generate(references, buffers) value = self.serializable_new # the new value is an object that may have # not-yet-in-the-remote-doc references, and may also # itself not be in the remote doc yet. the remote may # already have some of the references, but # unfortunately we don't have an easy way to know # unless we were to check BEFORE the attr gets changed # (we need the old _all_models before setting the # property). So we have to send all the references the # remote could need, even though it could be inefficient. # If it turns out we need to fix this we could probably # do it by adding some complexity. value_refs = set(collect_models(value)) # we know we don't want a whole new copy of the obj we're patching # unless it's also the new value if self.model != value: value_refs.discard(self.model) references.update(value_refs) return { 'kind' : 'ModelChanged', 'model' : self.model.ref, 'attr' : self.attr, 'new' : value } [docs]class ColumnDataChangedEvent(DocumentPatchedEvent): ''' A concrete event representing efficiently replacing *all* existing data for a :class:`~bokeh.models.sources.ColumnDataSource` ''' [docs] def __init__(self, document, column_source, cols=None, setter=None, callback_invoker=None): ''' Args: document (Document) : A Bokeh document that is to be updated. column_source (ColumnDataSource) : cols (list[str]) : optional explicit list of column names to update. If None, all columns will be updated (default: None) setter (ClientSession or ServerSession or None, optional) : This is used to prevent "boomerang" updates to Bokeh apps. (default: None) See :class:`~bokeh.document.events.DocumentChangedEvent` for more details. callback_invoker (callable, optional) : A callable that will invoke any Model callbacks that should be executed in response to the change that triggered this event. (default: None) ''' super().__init__(document, setter, callback_invoker) self.column_source = column_source self.cols = cols [docs] def dispatch(self, receiver): ''' Dispatch handling of this event to a receiver. This method will invoke ``receiver._column_data_changed`` if it exists. ''' super().dispatch(receiver) if hasattr(receiver, '_column_data_changed'): receiver._column_data_changed(self) [docs] def generate(self, references, buffers): ''' Create a JSON representation of this event suitable for sending to clients. .. code-block:: python { 'kind' : 'ColumnDataChanged' 'column_source' : <reference to a CDS> 'new' : <new data to steam to column_source> 'cols' : <specific columns to update> } Args: references (dict[str, Model]) : If the event requires references to certain models in order to function, they may be collected here. **This is an "out" parameter**. The values it contains will be modified in-place. buffers (set) : If the event needs to supply any additional Bokeh protocol buffers, they may be added to this set. **This is an "out" parameter**. The values it contains will be modified in-place. ''' from ..util.serialization import transform_column_source_data data_dict = transform_column_source_data(self.column_source.data, buffers=buffers, cols=self.cols) return { 'kind' : 'ColumnDataChanged', 'column_source' : self.column_source.ref, 'new' : data_dict, 'cols' : self.cols} [docs]class ColumnsStreamedEvent(DocumentPatchedEvent): ''' A concrete event representing efficiently streaming new data to a :class:`~bokeh.models.sources.ColumnDataSource` ''' [docs] def __init__(self, document, column_source, data, rollover, setter=None, callback_invoker=None): ''' Args: document (Document) : A Bokeh document that is to be updated. column_source (ColumnDataSource) : The data source to stream new data to. data (dict or DataFrame) : New data to stream. If a DataFrame, will be stored as ``{c: df[c] for c in df.columns}`` rollover (int) : A rollover limit. If the data source columns exceed this limit, earlier values will be discarded to maintain the column length under the limit. setter (ClientSession or ServerSession or None, optional) : This is used to prevent "boomerang" updates to Bokeh apps. (default: None) See :class:`~bokeh.document.events.DocumentChangedEvent` for more details. callback_invoker (callable, optional) : A callable that will invoke any Model callbacks that should be executed in response to the change that triggered this event. (default: None) ''' super().__init__(document, setter, callback_invoker) self.column_source = column_source if pd and isinstance(data, pd.DataFrame): data = {c: data[c] for c in data.columns} self.data = data self.rollover = rollover [docs] def dispatch(self, receiver): ''' Dispatch handling of this event to a receiver. This method will invoke ``receiver._columns_streamed`` if it exists. ''' super().dispatch(receiver) if hasattr(receiver, '_columns_streamed'): receiver._columns_streamed(self) [docs] def generate(self, references, buffers): ''' Create a JSON representation of this event suitable for sending to clients. .. code-block:: python { 'kind' : 'ColumnsStreamed' 'column_source' : <reference to a CDS> 'data' : <new data to steam to column_source> 'rollover' : <rollover limit> } Args: references (dict[str, Model]) : If the event requires references to certain models in order to function, they may be collected here. **This is an "out" parameter**. The values it contains will be modified in-place. buffers (set) : If the event needs to supply any additional Bokeh protocol buffers, they may be added to this set. **This is an "out" parameter**. The values it contains will be modified in-place. ''' return { 'kind' : 'ColumnsStreamed', 'column_source' : self.column_source.ref, 'data' : self.data, 'rollover' : self.rollover } [docs]class ColumnsPatchedEvent(DocumentPatchedEvent): ''' A concrete event representing efficiently applying data patches to a :class:`~bokeh.models.sources.ColumnDataSource` ''' [docs] def __init__(self, document, column_source, patches, setter=None, callback_invoker=None): ''' Args: document (Document) : A Bokeh document that is to be updated. column_source (ColumnDataSource) : The data source to apply patches to. patches (list) : setter (ClientSession or ServerSession or None, optional) : This is used to prevent "boomerang" updates to Bokeh apps. (default: None) See :class:`~bokeh.document.events.DocumentChangedEvent` for more details. callback_invoker (callable, optional) : A callable that will invoke any Model callbacks that should be executed in response to the change that triggered this event. (default: None) ''' super().__init__(document, setter, callback_invoker) self.column_source = column_source self.patches = patches [docs] def dispatch(self, receiver): ''' Dispatch handling of this event to a receiver. This method will invoke ``receiver._columns_patched`` if it exists. ''' super().dispatch(receiver) if hasattr(receiver, '_columns_patched'): receiver._columns_patched(self) [docs] def generate(self, references, buffers): ''' Create a JSON representation of this event suitable for sending to clients. .. code-block:: python { 'kind' : 'ColumnsPatched' 'column_source' : <reference to a CDS> 'patches' : <patches to apply to column_source> } Args: references (dict[str, Model]) : If the event requires references to certain models in order to function, they may be collected here. **This is an "out" parameter**. The values it contains will be modified in-place. buffers (set) : If the event needs to supply any additional Bokeh protocol buffers, they may be added to this set. **This is an "out" parameter**. The values it contains will be modified in-place. ''' return { 'kind' : 'ColumnsPatched', 'column_source' : self.column_source.ref, 'patches' : self.patches } [docs]class TitleChangedEvent(DocumentPatchedEvent): ''' A concrete event representing a change to the title of a Bokeh Document. ''' [docs] def __init__(self, document, title, setter=None, callback_invoker=None): ''' Args: document (Document) : A Bokeh document that is to be updated. title (str) : The new title to set on the Document setter (ClientSession or ServerSession or None, optional) : This is used to prevent "boomerang" updates to Bokeh apps. (default: None) See :class:`~bokeh.document.events.DocumentChangedEvent` for more details. callback_invoker (callable, optional) : A callable that will invoke any Model callbacks that should be executed in response to the change that triggered this event. (default: None) ''' super().__init__(document, setter, callback_invoker) self.title = title [docs] def combine(self, event): ''' ''' if not isinstance(event, TitleChangedEvent): return False # If these are not true something weird is going on, maybe updates from # Python bokeh.client, don't try to combine if self.setter != event.setter: return False if self.document != event.document: return False self.title = event.title self.callback_invoker = event.callback_invoker return True [docs] def generate(self, references, buffers): ''' Create a JSON representation of this event suitable for sending to clients. .. code-block:: python { 'kind' : 'TitleChanged' 'title' : <new title to set> } Args: references (dict[str, Model]) : If the event requires references to certain models in order to function, they may be collected here. **This is an "out" parameter**. The values it contains will be modified in-place. buffers (set) : If the event needs to supply any additional Bokeh protocol buffers, they may be added to this set. **This is an "out" parameter**. The values it contains will be modified in-place. ''' return { 'kind' : 'TitleChanged', 'title' : self.title } [docs]class RootAddedEvent(DocumentPatchedEvent): ''' A concrete event representing a change to add a new Model to a Document's collection of "root" models. ''' [docs] def __init__(self, document, model, setter=None, callback_invoker=None): ''' Args: document (Document) : A Bokeh document that is to be updated. model (Model) : The Bokeh Model to add as a Document root. setter (ClientSession or ServerSession or None, optional) : This is used to prevent "boomerang" updates to Bokeh apps. (default: None) See :class:`~bokeh.document.events.DocumentChangedEvent` for more details. callback_invoker (callable, optional) : A callable that will invoke any Model callbacks that should be executed in response to the change that triggered this event. (default: None) ''' super().__init__(document, setter, callback_invoker) self.model = model [docs] def generate(self, references, buffers): ''' Create a JSON representation of this event suitable for sending to clients. .. code-block:: python { 'kind' : 'RootAdded' 'title' : <reference to a Model> } Args: references (dict[str, Model]) : If the event requires references to certain models in order to function, they may be collected here. **This is an "out" parameter**. The values it contains will be modified in-place. buffers (set) : If the event needs to supply any additional Bokeh protocol buffers, they may be added to this set. **This is an "out" parameter**. The values it contains will be modified in-place. ''' references.update(self.model.references()) return { 'kind' : 'RootAdded', 'model' : self.model.ref } [docs]class RootRemovedEvent(DocumentPatchedEvent): ''' A concrete event representing a change to remove an existing Model from a Document's collection of "root" models. ''' [docs] def __init__(self, document, model, setter=None, callback_invoker=None): ''' Args: document (Document) : A Bokeh document that is to be updated. model (Model) : The Bokeh Model to remove as a Document root. setter (ClientSession or ServerSession or None, optional) : This is used to prevent "boomerang" updates to Bokeh apps. (default: None) See :class:`~bokeh.document.events.DocumentChangedEvent` for more details. callback_invoker (callable, optional) : A callable that will invoke any Model callbacks that should be executed in response to the change that triggered this event. (default: None) ''' super().__init__(document, setter, callback_invoker) self.model = model [docs] def generate(self, references, buffers): ''' Create a JSON representation of this event suitable for sending to clients. .. code-block:: python { 'kind' : 'RootRemoved' 'title' : <reference to a Model> } Args: references (dict[str, Model]) : If the event requires references to certain models in order to function, they may be collected here. **This is an "out" parameter**. The values it contains will be modified in-place. buffers (set) : If the event needs to supply any additional Bokeh protocol buffers, they may be added to this set. **This is an "out" parameter**. The values it contains will be modified in-place. ''' return { 'kind' : 'RootRemoved', 'model' : self.model.ref } [docs]class SessionCallbackAdded(DocumentChangedEvent): ''' A concrete event representing a change to add a new callback (e.g. periodic, timeout, or "next tick") to a Document. ''' [docs] def __init__(self, document, callback): ''' Args: document (Document) : A Bokeh document that is to be updated. callback (SessionCallback) : The callback to add ''' super().__init__(document) self.callback = callback [docs] def dispatch(self, receiver): ''' Dispatch handling of this event to a receiver. This method will invoke ``receiver._session_callback_added`` if it exists. ''' super().dispatch(receiver) if hasattr(receiver, '_session_callback_added'): receiver._session_callback_added(self) [docs]class SessionCallbackRemoved(DocumentChangedEvent): ''' A concrete event representing a change to remove an existing callback (e.g. periodic, timeout, or "next tick") from a Document. ''' [docs] def __init__(self, document, callback): ''' Args: document (Document) : A Bokeh document that is to be updated. callback (SessionCallback) : The callback to remove ''' super().__init__(document) self.callback = callback [docs] def dispatch(self, receiver): ''' Dispatch handling of this event to a receiver. This method will invoke ``receiver._session_callback_removed`` if it exists. ''' super().dispatch(receiver) if hasattr(receiver, '_session_callback_removed'): receiver._session_callback_removed(self) #----------------------------------------------------------------------------- # Private API #----------------------------------------------------------------------------- #----------------------------------------------------------------------------- # Code #-----------------------------------------------------------------------------