### API LLM Basic Completion with Perplexity in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/index_search=

This Rust code example shows how to initialize and use the llm_client with an API-based LLM, specifically Perplexity. It demonstrates setting up a basic completion request, adding a user message with a prompt, and running the completion to get a response.

```rust
let llm_client = LlmClient::perplexity().sonar_large().init();
let mut basic_completion = llm_client.basic_completion();
basic_completion
    .prompt()
    .add_user_message()
    .set_content("Can you help me use the llm_client rust crate? I'm having trouble getting cuda to work.");
let response = basic_completion.run().await?;
```

--------------------------------

### Initialize Logger with Configuration in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.LoggingConfig_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Implements a method to initialize and start the logger based on the current LoggingConfig. It handles log file creation, console output setup, and log rotation (hourly, keeping up to 6 files). Returns an error if log directory or file appender creation fails.

```rust
pub fn load_logger(&mut self) -> Result<(), Error> {
    // If logging is enabled, creates log files and sets up console output.
    // Logs are rotated hourly and up to 6 files are kept.
    // ... implementation details ...
}
```

--------------------------------

### Instruct Prompt Methods

Source: https://docs.rs/llm_client/0.0.7/llm_client/workflows/reason/one_round/struct.ReasonOneRound_search=

Methods for manipulating the instruction prompt, including setting instructions and supporting material, which guide the model's output.

```APIDOC
## Instruct Prompt Methods (via `InstructPromptTrait`)

### Description
These methods, implemented through the `InstructPromptTrait`, allow for the dynamic modification of instructions and supporting materials within the `ReasonOneRound` struct's prompt configuration. This is crucial for tailoring the model's behavior for specific tasks.

### Methods

*   **`instruct_prompt_mut(&mut self) -> &mut InstructPrompt`**: Returns a mutable reference to the `InstructPrompt`.
*   **`set_instructions<T: AsRef<str>>(&mut self, instructions: T) -> &mut Self`**: Sets the primary instructions for the model.
*   **`instructions(&mut self) -> &mut PromptMessage`**: Returns a mutable reference to the instruction `PromptMessage`.
*   **`set_supporting_material<T: AsRef<str>>(&mut self, supporting_material: T) -> &mut Self`**: Sets additional supporting material to guide the model.
*   **`supporting_material(&mut self) -> &mut PromptMessage`**: Returns a mutable reference to the supporting material `PromptMessage`.

### Example Usage (Conceptual)
```rust
let mut reasoner = ReasonOneRound {
    // ... other fields
    primitive: my_primitive_instance,
    instruct_prompt: InstructPrompt {
        // ... initial prompt configuration
    },
    base_req: default_completion_request(),
};

reasoner.set_instructions("Analyze the following text and provide a summary.")
        .set_supporting_material("Ensure the summary is objective and concise.");

// The instruct_prompt is now updated and will be used in subsequent calls.
```
```

--------------------------------

### LoggingConfig::load_logger()

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.LoggingConfig_search=std%3A%3Avec

Initializes and starts the logger based on the current LoggingConfig.

```APIDOC
## LoggingConfig::load_logger()

### Description
Initializes and starts the logger with the current configuration. If logging is enabled, creates log files and sets up console output. Logs are rotated hourly and up to 6 files are kept.

### Method
`load_logger(&mut self)`

### Endpoint
N/A (Method call on a LoggingConfig instance)

### Parameters
None (operates on `self`)

### Request Body
None

### Response
#### Success Response (200)
- **()** - Indicates successful logger initialization.

#### Errors
Returns error if:
  * Log directory creation fails
  * File appender creation fails
```

--------------------------------

### OpenAI Backend Initialization

Source: https://docs.rs/llm_client/0.0.7/llm_client/backend_builders/openai/struct.OpenAiBackendBuilder_search=u32+-%3E+bool

Demonstrates how to initialize the LlmClient with the configured OpenAI backend builder.

```APIDOC
## Initialize LlmClient with OpenAI Backend

### Description
Initializes the `LlmClient` instance after the `OpenAiBackendBuilder` has been configured.

### Method
`init()` method on the builder

### Endpoint
N/A (Client-side initialization)

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
N/A

### Request Example
```rust
use llm_client::backend_builders::openai::OpenAiBackendBuilder;
use llm_client::LlmClient;

// Assuming configuration has been applied previously
let builder = OpenAiBackendBuilder::default()
    .with_api_key("YOUR_API_KEY");

let client: LlmClient = builder.init().expect("Failed to initialize LlmClient");

// Now you can use the client for API calls
```

### Response
#### Success Response (200)
N/A (Initialization, not a direct API call response)

#### Response Example
N/A
```

--------------------------------

### TimingUsage new_from_generic Method (Rust)

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.TimingUsage_search=std%3A%3Avec

A general-purpose constructor for `TimingUsage` that initializes timing statistics with a provided start time. This method is useful when the completion response structure is not specifically `LlamaCppCompletionResponse` or when only a start time is available for initial setup.

```Rust
pub fn new_from_generic(start_time: Instant) -> TimingUsage
```

--------------------------------

### Boolean Outcome Reasoning with LLM in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/index_search=

This Rust example showcases how to use the llm_client to get a boolean outcome from an LLM. It involves setting instructions and supporting material, then calling `return_primitive()` to get the boolean result. This is useful for tasks like spam detection.

```rust
// boolean outcome
let reason_request = llm_client.reason().boolean();
reason_request
    .instructions()
    .set_content("Does this email subject indicate that the email is spam?");
reason_request
    .supporting_material()
    .set_content("You'll never believe these low, low prices 💲💲💲!!!");
let res: bool = reason_request.return_primitive().await.unwrap();
assert_eq!(res, true);
```

--------------------------------

### VZip Implementation

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/instruct_prompt/struct.InstructPrompt

Documentation for the `VZip` implementation, used for zipping operations.

```APIDOC
## VZip API

### Description
Provides functionality for zipping operations using the `VZip` trait.

### `impl<V, T> VZip<V> for T`

#### Methods

- **`fn vzip(self) -> V`**: Performs the zipping operation.

### Type Parameters
- **`V`**: The type of the zipped result.
- **`T`**: The type on which the `VZip` trait is implemented.
```

--------------------------------

### Get Base Completion Request in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.LlmClient_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Retrieves a base `CompletionRequest` object. This can be used as a starting point for constructing more complex requests to the language model.

```rust
pub fn base_request(&self) -> CompletionRequest
```

--------------------------------

### InstructPrompt Methods in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/instruct_prompt/struct.InstructPrompt_search=std%3A%3Avec

Provides essential methods for the `InstructPrompt` struct, including creating a new instance, resetting its prompt, and building instructions, supporting material, or the complete prompt string. The `build_instruct_prompt` method allows for flexible ordering of supporting material.

```rust
pub fn new() -> Self

pub fn reset_instruct_prompt(&mut self)

pub fn build_instructions(&self) -> Option<String>

pub fn build_supporting_material(&self) -> Option<String>

pub fn build_instruct_prompt( &self,
 supporting_material_first: bool,
 ) -> Result<String>
```

--------------------------------

### Ownership and Conversion

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/instruct_prompt/struct.InstructPrompt

Documentation for methods related to type ownership and conversion, including cloning and `TryFrom`/`TryInto` implementations.

```APIDOC
## Ownership and Conversion API

### Description
This section covers methods for managing type ownership and performing conversions.

### `impl<T> ToOwned for T`

#### Methods

- **`type Owned = T`**: The resulting type after obtaining ownership.
- **`fn to_owned(&self) -> T`**: Creates owned data from borrowed data, usually by cloning.
- **`fn clone_into(&self, target: &mut T)`**: Uses borrowed data to replace owned data, usually by cloning.

### `impl<T, U> TryFrom<U> for T`

#### Methods

- **`type Error = Infallible`**: The type returned in the event of a conversion error.
- **`fn try_from(value: U) -> Result<T, <T as TryFrom<U>>::Error>`**: Performs the conversion.

### `impl<T, U> TryInto<U> for T`

#### Methods

- **`type Error = <U as TryFrom<T>>::Error`**: The type returned in the event of a conversion error.
- **`fn try_into(self) -> Result<U, <U as TryFrom<T>>::Error>`**: Performs the conversion.

### Type Parameters
- **`T`**: The target type for conversion or ownership.
- **`U`**: The source type for conversion.
```

--------------------------------

### LLM Reasoning: Boolean Outcome

Source: https://docs.rs/llm_client/0.0.7/llm_client/index

This Rust example demonstrates using the llm_client for reasoning tasks that result in a boolean outcome. It shows how to set instructions and supporting material to guide the LLM in determining if an email subject indicates spam.

```rust
// boolean outcome
let reason_request = llm_client.reason().boolean();
reason_request
    .instructions()
    .set_content("Does this email subject indicate that the email is spam?");
reason_request
    .supporting_material()
    .set_content("You'll never believe these low, low prices 💲💲💲!!!");
let res: bool = reason_request.return_primitive().await.unwrap();
assert_eq!(res, true);

```

--------------------------------

### InstructPrompt Prompt Building Methods in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/instruct_prompt/struct.InstructPrompt_search=u32+-%3E+bool

Offers methods to build the instruction and supporting material strings from the respective fields within an InstructPrompt. These methods are essential for preparing prompt content before final assembly.

```rust
pub fn build_instructions(&self) -> Option<String>
```

```rust
pub fn build_supporting_material(&self) -> Option<String>
```

--------------------------------

### OpenAiBackendBuilder Initialization

Source: https://docs.rs/llm_client/0.0.7/llm_client/backend_builders/openai/struct.OpenAiBackendBuilder

Provides methods to configure and initialize the OpenAI backend, including setting API keys, hosts, ports, and selecting specific OpenAI models.

```APIDOC
## OpenAiBackendBuilder

`OpenAiBackendBuilder` is used to configure and build an OpenAI backend for the `llm_client`.

### Struct Definition

```rust
pub struct OpenAiBackendBuilder {
    pub config: OpenAiConfig,
    pub model: ApiLlmModel,
}
```

### Methods

#### `fn init(self) -> Result<LlmClient>`

Initializes the `LlmClient` with the configured OpenAI backend.

### Configuration Methods (Implementing `LlmApiConfigTrait`)

- `fn with_api_host<S>(self, host: S) -> Self` where `S: AsRef<str>`
  Sets the API host.
- `fn with_api_port<S>(self, port: S) -> Self` where `S: AsRef<str>`
  Sets the API port.
- `fn with_api_key<S>(self, api_key: S) -> Self` where `S: Into<String>`
  Sets the API key.
- `fn with_api_key_env_var<S>(self, api_key_env_var: S) -> Self` where `S: Into<String>`
  Sets the environment variable name for the API key.

### Logging Configuration Methods (Implementing `LoggingConfigTrait`)

- `fn logging_enabled(self, enabled: bool) -> Self`
  Enables or disables logging.
- `fn logger_name<S>(self, logger_name: S) -> Self` where `S: Into<String>`
  Sets the name of the logger.
- `fn log_path<P>(self, path: P) -> Self` where `P: AsRef<Path>`
  Sets the path for log files.
- `fn log_level_trace(self) -> Self`
  Sets the log level to TRACE.
- `fn log_level_debug(self) -> Self`
  Sets the log level to DEBUG.
- `fn log_level_info(self) -> Self`
  Sets the log level to INFO.
- `fn log_level_warn(self) -> Self`
  Sets the log level to WARN.
- `fn log_level_error(self) -> Self`
  Sets the log level to ERROR.

### Model Selection Methods (Implementing `OpenAiModelTrait`)

- `fn model(&mut self) -> &mut ApiLlmModel`
  Gets a mutable reference to the current model.
- `fn model_id_str(self, model_id: &str) -> Self`
  Sets the model using a model ID string.
- `fn gpt_4(self) -> Self`
  Uses `gpt-4` as the model.
- `fn gpt_4_32k(self) -> Self`
  Uses `gpt-4-32k` as the model.
- `fn gpt_3_5_turbo(self) -> Self`
  Uses `gpt-3.5-turbo` as the model.
- `fn gpt_4_turbo(self) -> Self`
  Uses `gpt-4-turbo` as the model.
- `fn gpt_4_o(self) -> Self`
  Uses `gpt-4-o` as the model.
- `fn gpt_4_o_mini(self) -> Self`
  Uses `gpt-4o-mini` as the model.
- `fn o1(self) -> Self`
  Uses `o1` as the model.
- `fn o1_mini(self) -> Self`
  Uses `o1-mini` as the model.

### Default Implementation

- `fn default() -> Self`
  Returns the default `OpenAiBackendBuilder`.
```

--------------------------------

### Exact String Outcome Reasoning with LLM in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/index_search=

This Rust example shows how to get an exact string outcome from an LLM, constrained to a list of allowed options. It involves setting instructions, supporting material, and adding allowed strings to the primitive's configuration before calling `return_primitive()`.

```rust
// string from a list of options outcome
let mut reason_request = llm_client.reason().exact_string();
reason_request
    .instructions()
    .set_content("Based on this readme, what is the name of the creator of this project?");
reason_request
    .supporting_material()
    .set_content(llm_client_readme);
reason_request
    .primitive
    .add_strings_to_allowed(&["shelby", "jack", "camacho", "john"]);
let response: String = reason_request.return_primitive().await.unwrap();
assert_eq!(res, "shelby");
```

--------------------------------

### Build Full Instruct Prompt in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/instruct_prompt/struct.InstructPrompt_search=u32+-%3E+bool

Constructs the complete instruct prompt string by combining instructions and supporting material, with an option to place supporting material first. This method returns a Result, indicating potential errors during prompt building.

```rust
pub fn build_instruct_prompt( &self,
 supporting_material_first: bool,
) -> Result<String>
```

--------------------------------

### Load Logger with LoggingConfig in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.LoggingConfig_search=std%3A%3Avec

Implements the `load_logger` method for `LoggingConfig` to initialize and start the logging system based on the struct's configuration. This method handles log file creation, console output setup, and log rotation. It returns a `Result` to indicate success or failure, with errors potentially arising from directory or file appender creation issues.

```rust
pub fn load_logger(&mut self) -> Result<(), Error> {
    // Initializes and starts the logger with the current configuration.
    // If logging is enabled, creates log files and sets up console output.
    // Logs are rotated hourly and up to 6 files are kept.
    // Errors: Returns error if:
    // * Log directory creation fails
    // * File appender creation fails
}
```

--------------------------------

### AnthropicBackendBuilder Initialization

Source: https://docs.rs/llm_client/0.0.7/llm_client/backend_builders/anthropic/struct.AnthropicBackendBuilder_search=

This section details the initialization and configuration of the AnthropicBackendBuilder.

```APIDOC
## Struct AnthropicBackendBuilder

### Description
Builder for configuring and initializing the Anthropic LLM client.

### Fields
- **config** (AnthropicConfig) - Configuration for the Anthropic API.
- **model** (ApiLlmModel) - The specific LLM model to use.

### Methods
#### `init(self) -> Result<LlmClient>`
Initializes the LlmClient with the configured Anthropic backend.

### Model Selection
#### `model_id_str(self, model_id: &str) -> Self`
Sets the model using a model ID string.

#### `claude_3_opus(self) -> Self`
Uses the Claude 3 Opus model.

#### `claude_3_sonnet(self) -> Self`
Uses the Claude 3 Sonnet model.

#### `claude_3_haiku(self) -> Self`
Uses the Claude 3 Haiku model.

#### `claude_3_5_sonnet(self) -> Self`
Uses the Claude 3.5 Sonnet model.

#### `claude_3_5_haiku(self) -> Self`
Uses the Claude 3.5 Haiku model.

### API Configuration
#### `with_api_host<S>(self, host: S) -> Self`
Sets the API host.

#### `with_api_port<S>(self, port: S) -> Self`
Sets the API port.

#### `with_api_key<S>(self, api_key: S) -> Self`
Sets the API key.

#### `with_api_key_env_var<S>(self, api_key_env_var: S) -> Self`
Sets the environment variable name for the API key.

### Logging Configuration
#### `logging_enabled(self, enabled: bool) -> Self`
Enables or disables logging.

#### `logger_name<S>(self, logger_name: S) -> Self`
Sets the name of the logger.

#### `log_path<P>(self, path: P) -> Self`
Sets the path for log files.

#### `log_level_trace(self) -> Self`
Sets the log level to TRACE.

#### `log_level_debug(self) -> Self`
Sets the log level to DEBUG.

#### `log_level_info(self) -> Self`
Sets the log level to INFO.

#### `log_level_warn(self) -> Self`
Sets the log level to WARN.

#### `log_level_error(self) -> Self`
Sets the log level to ERROR.

### Default Implementation
#### `default() -> Self`
Returns the default `AnthropicBackendBuilder`.

### Request Example (Initialization)
```rust
use llm_client::backend_builders::anthropic::AnthropicBackendBuilder;
use llm_client::LlmClient;

let client: LlmClient = AnthropicBackendBuilder::default()
    .claude_3_sonnet()
    .with_api_key("YOUR_ANTHROPIC_API_KEY")
    .init()
    .expect("Failed to initialize LlmClient");
```

### Response Example (Success)
Upon successful initialization, `init()` returns a `LlmClient` instance.
```rust
// LlmClient instance
```
```

--------------------------------

### Rust: Pointable trait for pointer management

Source: https://docs.rs/llm_client/0.0.7/llm_client/workflows/reason/decision/struct.Decision_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Defines methods for managing raw pointers, including getting the pointer's alignment, initializing an object at a given memory address, dereferencing the pointer to get a reference (mutable or immutable), and dropping the object pointed to.

```rust
trait Pointable {
    type Init = T;
    const ALIGN: usize;
    unsafe fn init(init: Self::Init) -> usize;
    unsafe fn deref<'a>(ptr: usize) -> &'a Self;
    unsafe fn deref_mut<'a>(ptr: usize) -> &'a mut Self;
    unsafe fn drop(ptr: usize);
}
```

--------------------------------

### InstructPromptTrait Provided Method: set_instructions

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/instruct_prompt/trait.InstructPromptTrait_search=

The `set_instructions` method is a provided implementation within the `InstructPromptTrait`. It allows setting the instructional text for the prompt, accepting any type that can be referenced as a string slice.

```rust
fn set_instructions<T: AsRef<str>>(&mut self, instructions: T) -> &mut Self { ... }
```

--------------------------------

### ApiPrompt - Constructor

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.ApiPrompt

Initializes a new ApiPrompt with optional token overhead configurations.

```APIDOC
## POST /llm_client/ApiPrompt/new

### Description
Initializes a new `ApiPrompt` instance. This constructor allows setting up the prompt formatter with a specific tokenizer and defining overhead token counts for messages and names.

### Method
POST

### Endpoint
/llm_client/ApiPrompt/new

### Parameters
#### Query Parameters
- **tokenizer** (Arc<dyn PromptTokenizer>) - Required - The tokenizer to use for counting tokens.
- **tokens_per_message** (Option<u32>) - Optional - The number of tokens to add as overhead for each message.
- **tokens_per_name** (Option<i32>) - Optional - The number of tokens to add as overhead for each message name.

### Request Example
```json
{
  "tokenizer": "<tokenizer_implementation>",
  "tokens_per_message": 5,
  "tokens_per_name": 1
}
```

### Response
#### Success Response (200)
- **ApiPrompt** - An instance of the ApiPrompt struct.

#### Response Example
```json
{
  "ApiPrompt": "<initialized_prompt_formatter>"
}
```
```

--------------------------------

### TypeId Trait

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.DeviceConfig_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Provides methods to get the TypeId of a value.

```APIDOC
## fn type_id(&self) -> TypeId

### Description
Gets the `TypeId` of `self`.

### Method
`type_id`

### Parameters
None

### Response
#### Success Response (200)
- **TypeId** (TypeId) - The type identifier of the object.

### Response Example
```json
{
  "TypeId": "some_type_id_value"
}
```
```

--------------------------------

### InstructPromptTrait Provided Method: instructions

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/instruct_prompt/trait.InstructPromptTrait_search=

The `instructions` method is a provided implementation within the `InstructPromptTrait`. It returns a mutable reference to the `PromptMessage` containing the current instructions, allowing direct manipulation.

```rust
fn instructions(&mut self) -> &mut PromptMessage { ... }
```

--------------------------------

### Rust - PrimitiveTrait Implementations for WordsPrimitive

Source: https://docs.rs/llm_client/0.0.7/llm_client/primitives/trait.PrimitiveTrait_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Example implementation of PrimitiveTrait for WordsPrimitive, specifying that its PrimitiveResult is a String.

```rust
impl PrimitiveTrait for WordsPrimitive {
    type PrimitiveResult = String;
}
```

--------------------------------

### InstructPrompt Build Instructions

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/instruct_prompt/struct.InstructPrompt

Builds a string representation of the instructions from the InstructPrompt. Returns an Option<String> which is None if no instructions are present.

```rust
pub fn build_instructions(&self) -> Option<String>
```

--------------------------------

### TryFrom and TryInto Implementations

Source: https://docs.rs/llm_client/0.0.7/llm_client/backend_builders/openai/struct.OpenAiBackendBuilder_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Documentation for `TryFrom` and `TryInto` trait implementations for type conversions.

```APIDOC
## TryFrom and TryInto Implementations

### Description
Provides functionality for fallible type conversions.

#### `impl<T, U> TryFrom<U> for T`
- **Type Alias `Error`**: The type returned in the event of a conversion error (`Infallible`).
- **`fn try_from(value: U) -> Result<T, <T as TryFrom<U>>::Error>`**: Performs the conversion.

#### `impl<T, U> TryInto<U> for T`
- **Type Alias `Error`**: The type returned in the event of a conversion error (`<U as TryFrom<T>>::Error`).
- **`fn try_into(self) -> Result<U, <U as TryFrom<T>>::Error>`**: Performs the conversion.
```

--------------------------------

### Rust - PrimitiveTrait Implementations for TextListPrimitive

Source: https://docs.rs/llm_client/0.0.7/llm_client/primitives/trait.PrimitiveTrait_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Example implementation of PrimitiveTrait for TextListPrimitive, specifying that its PrimitiveResult is TextListType.

```rust
impl PrimitiveTrait for TextListPrimitive {
    type PrimitiveResult = TextListType;
}
```

--------------------------------

### Rust - PrimitiveTrait Implementations for TextPrimitive

Source: https://docs.rs/llm_client/0.0.7/llm_client/primitives/trait.PrimitiveTrait_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Example implementation of PrimitiveTrait for TextPrimitive, specifying that its PrimitiveResult is a String.

```rust
impl PrimitiveTrait for TextPrimitive {
    type PrimitiveResult = String;
}
```

--------------------------------

### NVML Initialization API

Source: https://docs.rs/llm_client/0.0.7/llm_client/fn.init_nvml_wrapper_search=u32+-%3E+bool

Provides functionality to initialize the NVIDIA Management Library (NVML).

```APIDOC
## init_nvml_wrapper

### Description
Initializes NVIDIA Management Library (NVML). Attempts to load the NVML library from common paths on Linux, WSL, and Windows.

### Method
`pub fn init_nvml_wrapper() -> Result<Nvml, Error>`

### Endpoint
N/A (This is a function call within the Rust library, not a REST endpoint)

### Parameters
None

### Request Example
N/A

### Response
#### Success Response
- **Nvml** (type) - An object representing the initialized NVML library.

#### Error Response
- **Error** (type) - Returned if NVML initialization fails on all attempted paths.

#### Response Example
```rust
// Success
let nvml_handle = llm_client::init_nvml_wrapper()?; 

// Error
// Returns an Error type if initialization fails.
```
```

--------------------------------

### InstructPrompt Initialization and Reset Methods in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/instruct_prompt/struct.InstructPrompt_search=u32+-%3E+bool

Provides methods for creating a new InstructPrompt instance and resetting its internal state. The `new` function initializes the struct, while `reset_instruct_prompt` clears existing prompt components.

```rust
pub fn new() -> Self
```

```rust
pub fn reset_instruct_prompt(&mut self)
```

--------------------------------

### Rust - PrimitiveTrait Implementations for SentencesPrimitive

Source: https://docs.rs/llm_client/0.0.7/llm_client/primitives/trait.PrimitiveTrait_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Example implementation of PrimitiveTrait for SentencesPrimitive, specifying that its PrimitiveResult is a String.

```rust
impl PrimitiveTrait for SentencesPrimitive {
    type PrimitiveResult = String;
}
```

--------------------------------

### Rust - PrimitiveTrait Implementations for IntegerPrimitive

Source: https://docs.rs/llm_client/0.0.7/llm_client/primitives/trait.PrimitiveTrait_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Example implementation of PrimitiveTrait for IntegerPrimitive, specifying that its PrimitiveResult is a u32.

```rust
impl PrimitiveTrait for IntegerPrimitive {
    type PrimitiveResult = u32;
}
```

--------------------------------

### Rust - PrimitiveTrait Implementations for ExactStringPrimitive

Source: https://docs.rs/llm_client/0.0.7/llm_client/primitives/trait.PrimitiveTrait_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Example implementation of PrimitiveTrait for ExactStringPrimitive, specifying that its PrimitiveResult is a String.

```rust
impl PrimitiveTrait for ExactStringPrimitive {
    type PrimitiveResult = String;
}
```

--------------------------------

### NVML Wrapper Initialization

Source: https://docs.rs/llm_client/0.0.7/llm_client/fn.init_nvml_wrapper_search=std%3A%3Avec

Initializes the NVIDIA Management Library (NVML) and returns an Nvml object on success.

```APIDOC
## init_nvml_wrapper

### Description
Initializes NVIDIA Management Library (NVML). Attempts to load the NVML library from common paths on Linux, WSL, and Windows.

### Method
Rust function call (equivalent to a backend operation)

### Endpoint
N/A (Rust function)

### Parameters
None

### Request Example
N/A (Rust function)

### Response
#### Success Response
- **Nvml** (Nvml) - An initialized NVML object if successful.

#### Response Example
```rust
// Successful initialization returns Ok(Nvml)
```

#### Error Response
- **Error** (Error) - Returned if NVML initialization fails on all attempted paths.

#### Error Response Example
```rust
// Failed initialization returns Err(Error)
```
```

--------------------------------

### InstructPrompt Build Supporting Material

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/instruct_prompt/struct.InstructPrompt

Builds a string representation of the supporting material from the InstructPrompt. Returns an Option<String> which is None if no supporting material is present.

```rust
pub fn build_supporting_material(&self) -> Option<String>
```

--------------------------------

### Rust - PrimitiveTrait Implementations for BooleanPrimitive

Source: https://docs.rs/llm_client/0.0.7/llm_client/primitives/trait.PrimitiveTrait_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Example implementation of PrimitiveTrait for BooleanPrimitive, specifying that its PrimitiveResult is a boolean.

```rust
impl PrimitiveTrait for BooleanPrimitive {
    type PrimitiveResult = bool;
}
```

--------------------------------

### Get Type ID in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/backend_builders/anthropic/struct.AnthropicBackendBuilder_search=u32+-%3E+bool

Returns the `TypeId` of the current type. This is a fundamental method for runtime type identification in Rust.

```rust
fn type_id(&self) -> TypeId
```

--------------------------------

### LlmClient Initialization

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.LlmClient

Methods for creating and initializing an LlmClient instance using different backend builders.

```APIDOC
## LlmClient Initialization Methods

### `LlmClient::new(backend: Arc<LlmBackend>) -> Self`

Creates a new instance of `LlmClient` with a provided backend.

### `LlmClient::llama_cpp() -> LlamaCppBackendBuilder`

Creates a builder for the Llama.cpp backend. Use the `init` method on the builder to create an `LlmClient`.

### `LlmClient::openai() -> OpenAiBackendBuilder`

Creates a builder for the OpenAI backend. Use the `init` method on the builder to create an `LlmClient`.

### `LlmClient::anthropic() -> AnthropicBackendBuilder`

Creates a builder for the Anthropic backend. Use the `init` method on the builder to create an `LlmClient`.

### `LlmClient::perplexity() -> PerplexityBackendBuilder`

Creates a builder for the Perplexity backend. Use the `init` method on the builder to create an `LlmClient`.
```

--------------------------------

### Initialize ApiPrompt with Tokenizer and Overhead Settings

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.ApiPrompt

Creates a new instance of `ApiPrompt`. It requires a tokenizer, and optionally accepts token counts for messages and names to accurately calculate prompt token usage.

```rust
pub fn new( tokenizer: Arc<dyn PromptTokenizer>, tokens_per_message: Option<u32>, tokens_per_name: Option<i32>, ) -> ApiPrompt
```

--------------------------------

### Get Mutable InstructPrompt in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/workflows/nlp/extract/urls/struct.ExtractUrls_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Provides mutable access to the `InstructPrompt` associated with the `ExtractUrls` struct. This allows modification of the prompt configuration.

```rust
fn instruct_prompt_mut(&mut self) -> &mut InstructPrompt
```

--------------------------------

### DeviceConfig Initialization

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.DeviceConfig

Initializes the device configuration, detecting hardware capabilities and setting up appropriate configurations.

```APIDOC
## POST /deviceconfig/initialize

### Description
Initializes the device configuration, detecting hardware capabilities and setting up appropriate configurations.

### Method
POST

### Endpoint
/deviceconfig/initialize

### Parameters
#### Request Body
- **cpu_config** (CpuConfig) - Required - CPU configuration for thread count.
- **ram_config** (RamConfig) - Required - RAM configuration for non-GPU inference on Windows and Unix.
- **use_gpu** (bool) - Required - Indicates whether to use any available GPUs for inference.
- **cuda_config** (Option<CudaConfig>) - Optional - CUDA configuration for GPU inference on non-macOS platforms.
- **error_on_config_issue** (bool) - Optional - Determines error handling behavior for configuration issues.
- **layer_count** (Option<u64>) - Optional - The number of layers in the model.
- **average_layer_size_bytes** (Option<u64>) - Optional - The average size of a layer in bytes.
- **local_model_path** (String) - Required - The file system path to the local model.

### Response
#### Success Response (200)
- **Result** (Result<(), Error>) - Indicates success or failure of the initialization.

#### Error Response (e.g., 400, 500)
- **Error** - Details about the configuration or platform issue.
```

--------------------------------

### Get Target Directory

Source: https://docs.rs/llm_client/0.0.7/llm_client/fn.get_target_directory_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Resolves the Cargo target directory path. This function attempts to find the target directory through multiple methods.

```APIDOC
## get_target_directory

### Description
Resolves the Cargo target directory path. This function attempts to find the target directory through multiple methods:
  1. Checks the `CARGO_TARGET_DIR` environment variable
  2. Traverses up from the `OUT_DIR` to find a ‘target’ directory
  3. Traverses up from `CARGO_MANIFEST_DIR` to find a ‘target’ directory
  4. Falls back to using the compile-time `CARGO_MANIFEST_DIR`

This helps resolve issues with cargo workspace target directory resolution.

### Method
GET (Conceptual - this is a function call, not an HTTP endpoint)

### Endpoint
N/A (Function within the llm_client crate)

### Parameters
None

### Request Example
```rust
use llm_client::get_target_directory;

match get_target_directory() {
    Ok(path) => println!("Target directory: {}", path.display()),
    Err(e) => eprintln!("Error resolving target directory: {}", e),
}
```

### Response
#### Success Response (Ok)
- **PathBuf** - Path to the cargo target directory if found

#### Response Example
```json
{
  "path": "/path/to/your/project/target"
}
```

#### Error Response (Err)
- **Error** - An error indicating that the target directory could not be located.

#### Error Example
```json
{
  "error": "Failed to resolve target directory."
}
```
```

--------------------------------

### Get Target Directory

Source: https://docs.rs/llm_client/0.0.7/llm_client/fn.get_target_directory_search=

Resolves the Cargo target directory path using various methods including environment variables and directory traversal.

```APIDOC
## get_target_directory

### Description
Resolves the Cargo target directory path. This function attempts to find the target directory through multiple methods:
  1. Checks the `CARGO_TARGET_DIR` environment variable
  2. Traverses up from the `OUT_DIR` to find a ‘target’ directory
  3. Traverses up from `CARGO_MANIFEST_DIR` to find a ‘target’ directory
  4. Falls back to using the compile-time `CARGO_MANIFEST_DIR`

This helps resolve issues with cargo workspace target directory resolution.

### Method
GET

### Endpoint
/llm_client/get_target_directory

### Parameters
None

### Request Body
None

### Request Example
None

### Response
#### Success Response (200)
- **PathBuf**: Path to the cargo target directory if found

#### Response Example
{
  "path": "/path/to/cargo/target"
}

### Errors
Returns an error if unable to locate the target directory through any method.
```

--------------------------------

### Policy Composition: Or

Source: https://docs.rs/llm_client/0.0.7/llm_client/basic_completion/struct.BasicCompletion_search=std%3A%3Avec

Creates a new Policy that returns Action::Follow if either self or other policies return Action::Follow.

```APIDOC
## POST /policy/or

### Description
Creates a new Policy that returns Action::Follow if either self or other policies return Action::Follow.

### Method
POST

### Endpoint
/policy/or

### Parameters

#### Request Body
- **self_policy** (Policy<B, E>) - Required - The first policy.
- **other_policy** (Policy<B, E>) - Required - The second policy.

### Request Example
```json
{
  "self_policy": {"name": "policy1"},
  "other_policy": {"name": "policy2"}
}
```

### Response
#### Success Response (200)
- **new_policy** (Or<T, P>) - The combined policy.

#### Response Example
```json
{
  "new_policy": {"name": "combined_or_policy"}
}
```
```

--------------------------------

### TryInto<U>

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.DeviceConfig

Provides a fallible way to attempt conversion into another type.

```APIDOC
## POST /std/try_into

### Description
Provides a fallible way to attempt conversion from the current type into another type (`U`). Returns a `Result`.

### Method
POST

### Endpoint
/std/try_into

### Parameters
#### Request Body
- **from_type** (string) - Required - The source type.
- **to_type** (string) - Required - The target type.
- **value** (any) - Required - The value to attempt conversion on.

### Request Example
```json
{
  "from_type": "u32",
  "to_type": "i32",
  "value": 100
}
```

### Response
#### Success Response (200)
- **result** (string) - "Ok(converted_value)" on success, or "Err(error_details)" on failure.

#### Response Example
```json
{
  "result": "Ok(100)"
}
```
```

--------------------------------

### Get owned type from ToOwned implementation

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/instruct_prompt/struct.InstructPrompt_search=

Defines the associated type `Owned` which represents the owned version of the type `T` when implementing the `ToOwned` trait.

```rust
type Owned = T;
```

--------------------------------

### Get Mutable Reference to Last CascadeStep in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/cascade/round/struct.CascadeRound_search=

Provides mutable access to the last `CascadeStep` within the `CascadeRound`. This allows for modification of the most recent step.

```rust
pub fn last_step(&mut self) -> Result<&mut CascadeStep>
```

--------------------------------

### TryFrom<U>

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.DeviceConfig

Provides a fallible way to convert one type into another.

```APIDOC
## POST /std/try_from

### Description
Provides a fallible way to attempt conversion from one type (`U`) into another (`T`). Returns a `Result`.

### Method
POST

### Endpoint
/std/try_from

### Parameters
#### Request Body
- **from_type** (string) - Required - The source type.
- **to_type** (string) - Required - The target type.
- **value** (any) - Required - The value to attempt conversion on.

### Request Example
```json
{
  "from_type": "i32",
  "to_type": "u32",
  "value": 100
}
```

### Response
#### Success Response (200)
- **result** (string) - "Ok(converted_value)" on success, or "Err(error_details)" on failure.

#### Response Example
```json
{
  "result": "Ok(100)"
}
```
```

--------------------------------

### Get Generation Prefix in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/cascade/round/struct.CascadeRound_search=std%3A%3Avec

Retrieves the generation prefix for a given step within the `CascadeRound`. It returns a `Result` that may contain an `Option<String>`.

```Rust
pub fn generation_prefix( &self, current_step: &CascadeStep, ) -> Result<Option<String>>
```

--------------------------------

### TryFrom and TryInto Implementations

Source: https://docs.rs/llm_client/0.0.7/llm_client/backend_builders/openai/struct.OpenAiBackendBuilder_search=

Implementations for attempting conversions between types, with error handling for potential failures.

```APIDOC
## `TryFrom<U>` for `T`

### Description
Provides a mechanism to attempt conversion from type `U` into type `T`.

#### `Error` Type

*   **Description**: The type returned in the event of a conversion error.
*   **Type**: `Infallible`

#### `try_from` Method

### Description
Performs the conversion.

### Method
`fn try_from(value: U) -> Result<T, <T as TryFrom<U>>::Error>`

## `TryInto<U>` for `T`

### Description
Provides a mechanism to attempt conversion from type `T` into type `U`.

#### `Error` Type

*   **Description**: The type returned in the event of a conversion error.
*   **Type**: `<U as TryFrom<T>>::Error`

#### `try_into` Method

### Description
Performs the conversion.

### Method
`fn try_into(self) -> Result<U, <U as TryFrom<T>>::Error>`
```

--------------------------------

### Get Mutable Logging Configuration in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/backend_builders/anthropic/struct.AnthropicBackendBuilder_search=u32+-%3E+bool

Retrieves a mutable reference to the logging configuration. This allows detailed customization of logging behavior for the Anthropic client.

```rust
fn logging_config_mut(&mut self) -> &mut LoggingConfig
```

--------------------------------

### InstructPrompt Build Prompt

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/instruct_prompt/struct.InstructPrompt

Constructs the final prompt string by combining instructions and supporting material, with an option to place supporting material first. Returns a Result<String> as the operation can fail.

```rust
pub fn build_instruct_prompt( &self, supporting_material_first: bool, ) -> Result<String>
```

--------------------------------

### Rust DeviceConfig Initialization and Error Handling

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.DeviceConfig_search=

Initializes the device configuration by detecting hardware capabilities. It may return an error if hardware detection fails and 'error_on_config_issue' is true, or if the platform is unsupported.

```rust
pub fn initialize(&mut self) -> Result<(), Error>
```

--------------------------------

### Get Default AnthropicBackendBuilder in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/backend_builders/anthropic/struct.AnthropicBackendBuilder_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Implements the Default trait for AnthropicBackendBuilder, providing a default constructor. This allows creating a builder with standard initial settings.

```rust
fn default() -> Self
```

--------------------------------

### InstructPromptTrait Methods

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/instruct_prompt/trait.InstructPromptTrait_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

This section details the methods available within the InstructPromptTrait, including required and provided methods for prompt management.

```APIDOC
## Trait InstructPromptTrait

### Description
Provides methods to manage instructional prompts and supporting material for language models.

### Required Methods

#### `fn instruct_prompt_mut(&mut self) -> &mut InstructPrompt`

- **Description**: Returns a mutable reference to the underlying `InstructPrompt`.
- **Method**: `GET` (implicit, as it returns a mutable reference)
- **Endpoint**: N/A (trait method)

### Provided Methods

#### `fn set_instructions<T: AsRef<str>>(&mut self, instructions: T) -> &mut Self`

- **Description**: Sets the instructions for the prompt.
- **Method**: `SET` (implicit, as it modifies the object)
- **Endpoint**: N/A (trait method)
- **Parameters**:
  - `instructions` (string slice) - Required - The instructions to set.

#### `fn instructions(&mut self) -> &mut PromptMessage`

- **Description**: Returns a mutable reference to the prompt's instructions.
- **Method**: `GET` (implicit, as it returns a mutable reference)
- **Endpoint**: N/A (trait method)

#### `fn set_supporting_material<T: AsRef<str>>(&mut self, supporting_material: T) -> &mut Self`

- **Description**: Sets the supporting material for the prompt.
- **Method**: `SET` (implicit, as it modifies the object)
- **Endpoint**: N/A (trait method)
- **Parameters**:
  - `supporting_material` (string slice) - Required - The supporting material to set.

#### `fn supporting_material(&mut self) -> &mut PromptMessage`

- **Description**: Returns a mutable reference to the prompt's supporting material.
- **Method**: `GET` (implicit, as it returns a mutable reference)
- **Endpoint**: N/A (trait method)

### Dyn Compatibility

This trait is **not** dyn compatible.
```

--------------------------------

### Implement Default for LoggingConfig in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.LoggingConfig_search=

Allows LoggingConfig to be created using the default() method, providing a convenient way to get a LoggingConfig instance with its default settings.

```rust
fn default() -> LoggingConfig
```

--------------------------------

### CloneToUninit

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.DeviceConfig

Nightly-only experimental API for copy-assignment to uninitialized memory.

```APIDOC
## POST /std/clone_to_uninit

### Description
Nightly-only experimental API for copy-assignment to uninitialized memory.

### Method
POST

### Endpoint
/std/clone_to_uninit

### Parameters
#### Request Body
- **self** (any) - Required - The value to be copied.
- **dest** (pointer) - Required - The destination pointer in uninitialized memory.

### Request Example
```json
{
  "self": "value_to_copy",
  "dest": "0x1234567890ab"
}
```

### Response
#### Success Response (200)
- **status** (string) - Indicates successful operation.

#### Response Example
```json
{
  "status": "success"
}
```
```

--------------------------------

### Get Model Layer Count (Rust)

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.DeviceConfig_search=u32+-%3E+bool

Returns the total number of layers in the model. This value is set at runtime. An error is returned if `layer_count` has not been set.

```rust
pub fn layer_count(&self) -> Result<u64, Error>
```

--------------------------------

### Get Generation Prefix for CascadeStep in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/cascade/round/struct.CascadeRound_search=

Retrieves the generation prefix for a given `CascadeStep` within the `CascadeRound`. This operation may return an error or an optional String.

```rust
pub fn generation_prefix(
    &self,
    current_step: &CascadeStep,
) -> Result<Option<String>>
```

--------------------------------

### Policy Composition: And

Source: https://docs.rs/llm_client/0.0.7/llm_client/basic_completion/struct.BasicCompletion_search=std%3A%3Avec

Creates a new Policy that returns Action::Follow only if both self and other policies return Action::Follow.

```APIDOC
## POST /policy/and

### Description
Creates a new Policy that returns Action::Follow only if both self and other policies return Action::Follow.

### Method
POST

### Endpoint
/policy/and

### Parameters

#### Request Body
- **self_policy** (Policy<B, E>) - Required - The first policy.
- **other_policy** (Policy<B, E>) - Required - The second policy.

### Request Example
```json
{
  "self_policy": {"name": "policy1"},
  "other_policy": {"name": "policy2"}
}
```

### Response
#### Success Response (200)
- **new_policy** (And<T, P>) - The combined policy.

#### Response Example
```json
{
  "new_policy": {"name": "combined_and_policy"}
}
```
```

--------------------------------

### Get API Configuration in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/backend_builders/anthropic/struct.AnthropicBackendBuilder_search=u32+-%3E+bool

Retrieves an immutable reference to the API configuration for the Anthropic client. Used for inspecting current API settings without modification.

```rust
fn api_config(&self) -> &ApiConfig
```

--------------------------------

### InstructPromptTrait API

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/instruct_prompt/trait.InstructPromptTrait_search=u32+-%3E+bool

Details the methods available on the InstructPromptTrait, including required and provided methods.

```APIDOC
## Trait InstructPromptTrait

### Description
This trait provides methods for interacting with and modifying instruct prompts, including setting and retrieving instructions and supporting material.

### Required Methods

#### `fn instruct_prompt_mut(&mut self) -> &mut InstructPrompt`

This is a required method that must be implemented by any type that uses this trait. It returns a mutable reference to an `InstructPrompt`.

### Provided Methods

#### `fn set_instructions<T: AsRef<str>>(&mut self, instructions: T) -> &mut Self`

Sets the instructions for the prompt. Accepts any type that can be referenced as a string slice.

#### `fn instructions(&mut self) -> &mut PromptMessage`

Returns a mutable reference to the prompt's instructions, represented as a `PromptMessage`.

#### `fn set_supporting_material<T: AsRef<str>>(&mut self, supporting_material: T) -> &mut Self`

Sets the supporting material for the prompt. Accepts any type that can be referenced as a string slice.

#### `fn supporting_material(&mut self) -> &mut PromptMessage`

Returns a mutable reference to the prompt's supporting material, represented as a `PromptMessage`.

### Dyn Compatibility

This trait is **not** dyn compatible, meaning it cannot be used with dynamic dispatch (e.g., `dyn InstructPromptTrait`).
```

--------------------------------

### CascadeFlow: Open Cascade Operation

Source: https://docs.rs/llm_client/0.0.7/llm_client/components/cascade/struct.CascadeFlow

Marks the beginning of a cascade operation. This method likely performs setup or initialization tasks required before the cascade execution begins.

```rust
pub fn open_cascade(&mut self)
```

--------------------------------

### LlamaCppBackendBuilder Initialization

Source: https://docs.rs/llm_client/0.0.7/llm_client/backend_builders/llama_cpp/struct.LlamaCppBackendBuilder_search=u32+-%3E+bool

The `init` method is an asynchronous function that consumes the builder and returns a configured `LlmClient` instance, assuming successful initialization.

```APIDOC
## `init` Method

### Description
Initializes the Llama.cpp backend with the current configuration and returns a `LlmClient`.

### Method
`pub async fn init(self) -> Result<LlmClient>`

### Parameters
This method takes ownership of the `LlamaCppBackendBuilder` instance.
```

--------------------------------

### Get Extract URLs in Rust

Source: https://docs.rs/llm_client/0.0.7/llm_client/workflows/nlp/extract/struct.Extract_search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

A method `urls` on the `Extract` struct that consumes `self` and returns an `ExtractUrls` type, likely used to retrieve associated URLs.

```rust
pub fn urls(self) -> ExtractUrls
```

--------------------------------

### Get Built PromptMessage Content

Source: https://docs.rs/llm_client/0.0.7/llm_client/struct.PromptMessage

Retrieves the final, concatenated message content. This method returns the complete message as a string, assembled according to the rules defined by the TextConcatenator.

```rust
pub fn get_built_prompt_message(&self) -> Result<String, Error>
```