2021/09/22 - Amazon Lex Model Building V2 - 2 new api methods
Changes This release adds support for utterances statistics for bots built using Lex V2 console and APIs. For details, see: https://docs.aws.amazon.com/lexv2/latest/dg/monitoring-utterances.html
Deletes stored utterances.
Amazon Lex stores the utterances that users send to your bot. Utterances are stored for 15 days for use with the operation, and then stored indefinitely for use in improving the ability of your bot to respond to user input..
Use the DeleteUtterances operation to manually delete utterances for a specific session. When you use the DeleteUtterances operation, utterances stored for improving your bot's ability to respond to user input are deleted immediately. Utterances stored for use with the ListAggregatedUtterances operation are deleted after 15 days.
See also: AWS API Documentation
Request Syntax
client.delete_utterances(
botId='string',
localeId='string',
sessionId='string'
)
string
[REQUIRED]
The unique identifier of the bot that contains the utterances.
string
The identifier of the language and locale where the utterances were collected. The string must match one of the supported locales. For more information, see Supported languages.
string
The unique identifier of the session with the user. The ID is returned in the response from the and operations.
dict
Response Syntax
{}
Response Structure
(dict) --
Provides a list of utterances that users have sent to the bot.
Utterances are aggregated by the text of the utterance. For example, all instances where customers used the phrase "I want to order pizza" are aggregated into the same line in the response.
You can see both detected utterances and missed utterances. A detected utterance is where the bot properly recognized the utterance and activated the associated intent. A missed utterance was not recognized by the bot and didn't activate an intent.
Utterances can be aggregated for a bot alias or for a bot version, but not both at the same time.
Utterances statistics are not generated under the following conditions:
The childDirected field was set to true when the bot was created.
You are using slot obfuscation with one or more slots.
You opted out of participating in improving Amazon Lex.
See also: AWS API Documentation
Request Syntax
client.list_aggregated_utterances(
botId='string',
botAliasId='string',
botVersion='string',
localeId='string',
aggregationDuration={
'relativeAggregationDuration': {
'timeDimension': 'Hours'|'Days'|'Weeks',
'timeValue': 123
}
},
sortBy={
'attribute': 'HitCount'|'MissedCount',
'order': 'Ascending'|'Descending'
},
filters=[
{
'name': 'Utterance',
'values': [
'string',
],
'operator': 'CO'|'EQ'
},
],
maxResults=123,
nextToken='string'
)
string
[REQUIRED]
The unique identifier of the bot associated with this request.
string
The identifier of the bot alias associated with this request. If you specify the bot alias, you can't specify the bot version.
string
The identifier of the bot version associated with this request. If you specify the bot version, you can't specify the bot alias.
string
[REQUIRED]
The identifier of the language and locale where the utterances were collected. For more information, see Supported languages.
dict
[REQUIRED]
The time window for aggregating the utterance information. You can specify a time between one hour and two weeks.
relativeAggregationDuration (dict) -- [REQUIRED]
The desired time window for aggregating utterances.
timeDimension (string) -- [REQUIRED]
The type of time period that the timeValue field represents.
timeValue (integer) -- [REQUIRED]
The period of the time window to gather statistics for. The valid value depends on the setting of the timeDimension field.
Hours - 1/3/6/12/24
Days - 3
Weeks - 1/2
dict
Specifies sorting parameters for the list of utterances. You can sort by the hit count, the missed count, or the number of distinct sessions the utterance appeared in.
attribute (string) -- [REQUIRED]
The utterance attribute to sort by.
order (string) -- [REQUIRED]
Specifies whether to sort the aggregated utterances in ascending or descending order.
list
Provides the specification of a filter used to limit the utterances in the response to only those that match the filter specification. You can only specify one filter and one string to filter on.
(dict) --
Filters responses returned by the ListAggregatedUtterances operation.
name (string) -- [REQUIRED]
The name of the field to filter the utterance list.
values (list) -- [REQUIRED]
The value to use for filtering the list of bots.
(string) --
operator (string) -- [REQUIRED]
The operator to use for the filter. Specify EQ when the ListAggregatedUtterances operation should return only utterances that equal the specified value. Specify CO when the ListAggregatedUtterances operation should return utterances that contain the specified value.
integer
The maximum number of utterances to return in each page of results. If there are fewer results than the maximum page size, only the actual number of results are returned. If you don't specify the maxResults parameter, 1,000 results are returned.
string
If the response from the ListAggregatedUtterances operation contains more results that specified in the maxResults parameter, a token is returned in the response. Use that token in the nextToken parameter to return the next page of results.
dict
Response Syntax
{
'botId': 'string',
'botAliasId': 'string',
'botVersion': 'string',
'localeId': 'string',
'aggregationDuration': {
'relativeAggregationDuration': {
'timeDimension': 'Hours'|'Days'|'Weeks',
'timeValue': 123
}
},
'aggregationWindowStartTime': datetime(2015, 1, 1),
'aggregationWindowEndTime': datetime(2015, 1, 1),
'aggregationLastRefreshedDateTime': datetime(2015, 1, 1),
'aggregatedUtterancesSummaries': [
{
'utterance': 'string',
'hitCount': 123,
'missedCount': 123,
'utteranceFirstRecordedInAggregationDuration': datetime(2015, 1, 1),
'utteranceLastRecordedInAggregationDuration': datetime(2015, 1, 1),
'containsDataFromDeletedResources': True|False
},
],
'nextToken': 'string'
}
Response Structure
(dict) --
botId (string) --
The identifier of the bot that contains the utterances.
botAliasId (string) --
The identifier of the bot alias that contains the utterances. If you specified the bot version, the bot alias ID isn't returned.
botVersion (string) --
The identifier of the bot version that contains the utterances. If you specified the bot alias, the bot version isn't returned.
localeId (string) --
The identifier of the language and locale that the utterances are in.
aggregationDuration (dict) --
The time period used to aggregate the utterance data.
relativeAggregationDuration (dict) --
The desired time window for aggregating utterances.
timeDimension (string) --
The type of time period that the timeValue field represents.
timeValue (integer) --
The period of the time window to gather statistics for. The valid value depends on the setting of the timeDimension field.
Hours - 1/3/6/12/24
Days - 3
Weeks - 1/2
aggregationWindowStartTime (datetime) --
The date and time that the aggregation window begins. Only data collected after this time is returned in the results.
aggregationWindowEndTime (datetime) --
The date and time that the aggregation window ends. Only data collected between the start time and the end time are returned in the results.
aggregationLastRefreshedDateTime (datetime) --
The last date and time that the aggregated data was collected. The time period depends on the length of the aggregation window.
Hours - for 1 hour time window, every half hour; otherwise every hour.
Days - every 6 hours
Weeks - for a one week time window, every 12 hours; otherwise, every day
aggregatedUtterancesSummaries (list) --
Summaries of the aggregated utterance data. Each response contains information about the number of times that the utterance was seen during the time period, whether it was detected or missed, and when it was seen during the time period.
(dict) --
Provides summary information for aggregated utterances. The ListAggregatedUtterances operations combines all instances of the same utterance into a single aggregated summary.
utterance (string) --
The text of the utterance. If the utterance was used with the RecognizeUtterance operation, the text is the transcription of the audio utterance.
hitCount (integer) --
The number of times that the utterance was detected by Amazon Lex during the time period. When an utterance is detected, it activates an intent or a slot.
missedCount (integer) --
The number of times that the utterance was missed by Amazon Lex An utterance is missed when it doesn't activate an intent or slot.
utteranceFirstRecordedInAggregationDuration (datetime) --
The date and time that the utterance was first recorded in the time window for aggregation. An utterance may have been sent to Amazon Lex before that time, but only utterances within the time window are counted.
utteranceLastRecordedInAggregationDuration (datetime) --
The last date and time that an utterance was recorded in the time window for aggregation. An utterance may be sent to Amazon Lex after that time, but only utterances within the time window are counted.
containsDataFromDeletedResources (boolean) --
Aggregated utterance data may contain utterances from versions of your bot that have since been deleted. When the aggregated contains this kind of data, this field is set to true.
nextToken (string) --
A token that indicates whether there are more results to return in a response to the ListAggregatedUtterances operation. If the nextToken field is present, you send the contents as the nextToken parameter of a ListAggregatedUtterances operation request to get the next page of results.