mindmeld.markup module

The markup module contains functions for interacting with the MindMeld Markup language for representing annotations of query text inline.

mindmeld.markup.bootstrap_query_file(input_file, output_file, nlp, **kwargs)

Apply predicted annotations to a file of text queries

Parameters:

• input_file (str) -- filename of queries to be processed
• output_file (str or None) -- filename for processed queries
• nlp (NaturalLanguageProcessor) -- an application's NLP with built models
• kwargs (dict) -- A dictionary of additional args

mindmeld.markup.bootstrap_query_row(proc_query, show_confidence, **kwargs)

Produce predicted annotation values and confidences for a single query

Parameters:

• proc_query (ProcessedQuery) -- a labeled query
• show_confidence (bool) -- whether to generate confidence columns
• **kwargs -- flags indicating which columns to generate

Returns:

(dict)

mindmeld.markup.cache_query_file(file_path, query_cache, query_factory=None, app_path=None, domain=None, intent=None, is_gold=False)

Loads the specified query file into the query cache

Parameters:

• file_path (str) -- The path of the file to load
• query_cache (QueryCache) -- A container containing cache query objects
• query_factory (QueryFactory, optional) -- An object which can create queries.
• app_path (str) -- The app path
• domain (str, optional) -- The name of the domain annotated for the query.
• intent (str, optional) -- The name of the intent annotated for the query.
• is_gold (bool, optional) -- True if the markup passed in is a reference, human-labeled example. Defaults to False.

Returns:

List of cached query ids

mindmeld.markup.dump_queries(queries, markup_format='mindmeld', **kwargs)

Converts a collection of processed queries to marked up query text

Parameters:

• queries (iterable) -- A collection of processed queries
• markup_format (str, optional) -- The format to use. Valid formats include 'mindmeld' and 'brat'. Defaults to 'mindmeld'
• **kwargs -- additional format specific parameters may be passed in as keyword arguments.

Yields:

str or tuple -- A marked up representation of the query

mindmeld.markup.dump_query(processed_query, markup_format='mindmeld', **kwargs)

Converts a processed query into marked up query text.

Parameters:

• processed_query (ProcessedQuery) -- The query to convert
• markup_format (str, optional) -- The format to use. Valid formats include 'mindmeld' and 'brat'. Defaults to 'mindmeld'
• **kwargs -- additional format specific parameters may be passed in as keyword arguments.

Returns:

A marked up representation of the query

Return type:

str

Raises:

ValueError

mindmeld.markup.load_query(markup, query_factory=None, app_path=None, domain=None, intent=None, is_gold=False, query_options=None)

Creates a processed query object from marked up query text.

Parameters:

• markup (str) -- The marked up query text.
• query_factory (QueryFactory, optional) -- An object which can create queries.
• app_path (str, optional) -- The dir path of the application
• domain (str, optional) -- The name of the domain annotated for the query.
• intent (str, optional) -- The name of the intent annotated for the query.
• is_gold (bool, optional) -- True if the markup passed in is a reference, human-labeled example. Defaults to False.
• query_options (dict, optional) -- A dict containing options for creating a Query, such as language, time_zone and timestamp

Returns:

a processed query

Return type:

ProcessedQuery

mindmeld.markup.mark_down(markup)

Removes all entity mark up from a string

Parameters:markup (str) -- A marked up string Returns:A clean string with no mark up Return type:str

mindmeld.markup.mark_down_file(file_path)

Read all annotated queries from the input file and remove all the annotations

Parameters:file_path (str) -- The path of the file to load Yields:(str) -- marked down query text for each line

mindmeld.markup.process_markup(markup, query_factory, query_options)

This function takes in some text and returns a constructed Query object associated with the

text, along with other objects like a list of entities.

Parameters:

• markup (str) -- The markup string to process
• query_factory (QueryFactory) -- The factory used to construct Query objects
• query_options (dict) -- A dictionary containing options for language, time_zone and time_stamp

Returns:

Returns a tuple of the raw text, the Query object associated with the text and a list of entities (ProcessedQuery) associated with the text

Return type:

(str, Query, list)

mindmeld.markup.read_query_file(file_path)

Summary

Parameters:file_path (str) -- The path of the file to load Yields:str -- query text for each line

mindmeld.markup.validate_markup(markup, query_factory)

Checks whether the markup text is well-formed.

Parameters:

• markup (str) -- The marked up query text
• query_factory (QueryFactory) -- An object which can create queries

Returns:

True if the markup is valid

Return type:

bool