1.0.0-preview.3 (2020-03-10)
- [Breaking] Renamed
id
todataSourceEntityId
in theLinkedEntity
type. - Added special handling for the string
"none"
as thecountryHint
parameter of theTextAnalyticsClient.detectLanguage
."none"
is now treated the same as the empty string, and indicates that the default language detection model should be used. - [Breaking] Renamed
offset
tographemeOffset
andlength
tographemeLength
in fields of response objects as appropriate in order to make it clear that the offsets and lengths are in units of Unicode graphemes. - [Breaking] Renamed
sentimentScores
on bothDocumentSentiment
andSentenceSentiment
toconfidenceScores
, and renamed the typeSentimentScorePerLabel
toSentimentConfidenceScores
. - [Breaking] Renamed
characterCount
tographemeCount
in theDocumentStatistics
interface, to align with the change to usinggrapheme
in the lengths/offsets of response objects.