Skip to main content
Version: 0.94

templates.UtteranceContextTemplateSchema

class templates.UtteranceContextTemplateSchema(*args, **kwargs)

Utterance Context template

LF Template specific to conversational use-case. Looks for keywords/regex pattern in specified number of previous or next utterances.

Parameters:

  • operator ({"BEFORE", "AFTER", "BEFORE OR AFTER"}) – Indicates whether to search before or after the current utterance.

  • window (int, default 1) – How many utterances should we restrict the search window to.

  • speakers ({"any", "current", "other", or speakername}) – List of speakers whose utterances will be considered for matching

  • value (str) – The value we should be searching for.

  • regex (bool, default False) – If this is True, value should be a RegEx Pattern.

  • case_sensitive (bool, default False) – Case sensitive or not

  • tokenize (bool, default False) – Only can be True if regex is False, we will search for an exact word match between value and span context and not match on substrings.