ReasoningUtils class
Utilities for handling reasoning/thinking content
Constructors
Properties
- hashCode → int
-
The hash code for this object.
no setterinherited
- runtimeType → Type
-
A representation of the runtime type of the object.
no setterinherited
Methods
-
noSuchMethod(
Invocation invocation) → dynamic -
Invoked when a nonexistent method or property is accessed.
inherited
-
toString(
) → String -
A string representation of this object.
inherited
Operators
-
operator ==(
Object other) → bool -
The equality operator.
inherited
Static Methods
-
checkReasoningStatus(
{required Map< String, dynamic> ? delta, required bool hasReasoningContent, required String lastChunk}) → ReasoningDetectionResult - Check if reasoning just finished based on delta content This matches the TypeScript isReasoningJustDone function
-
containsThinkingTags(
String content) → bool - Check if content contains thinking tags
-
extractContentWithoutThinking(
String content) → String - Extract content without thinking tags
-
extractReasoningContent(
Map< String, dynamic> ? delta) → String? - Extract reasoning content from delta
-
filterThinkingContent(
String content) → String -
Filter thinking content from text for display purposes
Removes
<think>...</think>tags and their content -
getMaxTokensParams(
{required String model, int? maxTokens}) → Map< String, dynamic> - Get appropriate max tokens parameter for reasoning models
-
getReasoningEffortParams(
{required String providerId, required String model, ReasoningEffort? reasoningEffort, int? maxTokens}) → Map< String, dynamic> - Get reasoning effort parameter for different providers
-
hasReasoningContent(
Map< String, dynamic> ? delta) → bool - Check if delta contains reasoning content
-
isKnownReasoningModel(
String model) → bool - Check if the model is known to support reasoning (broader check) This is a hint for UI behavior, but actual reasoning detection should be based on response content, not model name
-
isOpenAIReasoningModel(
String model) → bool - Check if the model is an OpenAI reasoning model (o1, o3, o4 series) This is used for specific OpenAI API parameter handling
-
parseReasoningMetrics(
{required DateTime startTime, DateTime? firstTokenTime, DateTime? firstContentTime, int? completionTokens}) → Map< String, dynamic> - Parse reasoning metrics from response
-
shouldDisableTemperature(
String model) → bool - Check if temperature should be disabled for reasoning models
-
shouldDisableTopP(
String model) → bool - Check if top_p should be disabled for reasoning models