Struct TextAccumulator 

pub struct TextAccumulator { /* private fields */ }

Accumulates countable text strings from structured Gemini API objects.

This is a faithful port of the Python SDK’s _TextsAccumulator class. It traverses Content, Tool, FunctionCall, FunctionResponse, and Schema objects, extracting all text that Google counts when computing token counts.

How text is extracted

• Text parts: the text string itself
• Function calls: the function name, plus all dict keys and string values from the args
• Function responses: the function name, plus all dict keys and string values from the response
• Function declarations (tools): the name, description, plus recursive schema traversal of parameters and response
• Schemas: format, description, enum values, required field names, property keys, and recursive traversal of nested schemas

Implementations

impl TextAccumulator

pub fn new() -> Self

Creates a new empty accumulator.

pub fn get_texts(&self) -> &[String]

Returns the accumulated text segments.

pub fn into_texts(self) -> Vec<String>

Consumes the accumulator and returns the accumulated text segments.

pub fn add_contents(&mut self, contents: &[Content])

Adds all text from multiple content objects.

pub fn add_content(&mut self, content: &Content)

Adds all countable text from a single content object.

Processes each part in the content by delegating to add_part.

pub fn add_part(&mut self, part: &Part)

Adds countable text from a single content part.

Processes the part’s fields:

• Text: appends the text directly
• Function calls: delegates to add_function_call
• Function responses: delegates to add_function_response

pub fn add_function_call(&mut self, function_call: &FunctionCall)

Adds countable text from a function call.

Extracts the function name and traverses the args dictionary, collecting all keys and string values.

pub fn add_tools(&mut self, tools: &[Tool])

Adds countable text from multiple tools.

pub fn add_tool(&mut self, tool: &Tool)

Adds countable text from a single tool definition.

Processes each function declaration in the tool.

pub fn add_function_responses(&mut self, responses: &[FunctionResponse])

Adds countable text from multiple function responses.

pub fn add_function_response(&mut self, function_response: &FunctionResponse)

Adds countable text from a function response.

Extracts the function name and traverses the response dictionary, collecting all keys and string values.

pub fn add_schema(&mut self, schema: &Schema)

Adds countable text from a schema definition.

Extracts format, description, enum values, required field names, property keys, and recursively processes nested schemas (items, properties, examples).

Trait Implementations

impl Default for TextAccumulator

fn default() -> Self

Returns the “default value” for a type.