PerspectiveAnalyzeCommentResponse

Response returned by Respectify's public Perspective-compatible analyse endpoint.

Python

from respectify.schemas import PerspectiveAnalyzeCommentAttributeScore, PerspectiveAnalyzeCommentResponse

PHP

use Respectify\Schemas\PerspectiveAnalyzeCommentAttributeScore;
use Respectify\Schemas\PerspectiveAnalyzeCommentResponse;

JSON

Types are returned as JSON objects in API responses.

Fields

attributeScores

Map of requested Perspective attribute names, such as TOXICITY or INSULT, to their score objects. Each value contains a summaryScore and, when requested, optional spanScores.

Python

result.attributeScores: Dict[str, PerspectiveAnalyzeCommentAttributeScore]

PHP

$result->attributeScores: array<string, PerspectiveAnalyzeCommentAttributeScore>

JSON

"attributeScores": {
  "ATTRIBUTE_NAME": {
  "summaryScore": PerspectiveScore,
  "spanScores": [PerspectiveAnalyzeCommentSpanScore, ...]
}
}

Each key is a requested attribute name such as TOXICITY. Each value is a PerspectiveAnalyzeCommentAttributeScore object.

languages

Requested or effective language context for this response, as language codes such as en, es, or fr.

Python

result.languages: List[str]

PHP

$result->languages: str[]

JSON

"languages": ["...", "..."]

detectedLanguages

Present only if Respectify actually detected languages, as language codes such as en, es, or fr.

Python

result.detectedLanguages: Optional[List[str]]

PHP

$result->detectedLanguages: ?array<int, str>

JSON

"detectedLanguages": ["...", "..."]

clientToken

Caller-provided correlation token, echoed back unchanged when supplied. Use this if you want to match the response to your own request ID.

Python

result.clientToken: Optional[str]

PHP

$result->clientToken: ?string

JSON

"clientToken": "string"  // or null

Sub-types

PerspectiveAnalyzeCommentAttributeScore

Google-style attribute score entry returned by analyzeComment.

Python

Field	Type	Description
summaryScore	PerspectiveScore	Overall score for the attribute. This is the main score most integrations use for thresholds and moderation decisions.
spanScores	Optional[List[PerspectiveAnalyzeCommentSpanScore]]	Optional scores for specific spans of text that contributed to the attribute. Returned only when span annotations were requested.

PHP

Field	Type	Description
summaryScore	PerspectiveScore	Overall score for the attribute. This is the main score most integrations use for thresholds and moderation decisions.
spanScores	?array<int, PerspectiveAnalyzeCommentSpanScore>	Optional scores for specific spans of text that contributed to the attribute. Returned only when span annotations were requested.

JSON

{
  "summaryScore": {},
  "spanScores": []
}

PerspectiveAnalyzeCommentSpanScore

Google-style span score for a specific part of the comment.

Python

Field	Type	Description
begin	int	Start offset for the span in UTF-16 code units
end	int	End offset for the span in UTF-16 code units
score	PerspectiveScore	Score for this specific span

PHP

Field	Type	Description
begin	int	Start offset for the span in UTF-16 code units
end	int	End offset for the span in UTF-16 code units
score	PerspectiveScore	Score for this specific span

JSON

{
  "begin": 1,
  "end": 1,
  "score": {}
}

PerspectiveScore

Google-style score object used in Perspective compatibility responses.

Python

Field	Type	Description
value	float # 0.0-1.0	Probability-style score between 0.0 and 1.0. Higher numbers mean more of the requested attribute.
type	str	Score type. Respectify currently returns PROBABILITY, meaning the value is a 0 to 1 probability-style score.

PHP

Field	Type	Description
value	float	Probability-style score between 0.0 and 1.0. Higher numbers mean more of the requested attribute.
type	string	Score type. Respectify currently returns PROBABILITY, meaning the value is a 0 to 1 probability-style score.

JSON

{
  "value": 0.5,
  "type": "string"
}