kcenon::pacs::ai::ai_service_connector Class Reference

Connector for external AI inference services. More...

#include <ai_service_connector.h>

Collaboration diagram for kcenon::pacs::ai::ai_service_connector:

Public Types
using	status_callback = std::function<void(const inference_status&)>
	Callback type for status updates.
using	completion_callback
	Callback type for completion notification.

Static Public Member Functions
static auto	initialize (const ai_service_config &config) -> Result< std::monostate >
	Initialize the AI service connector.
static void	shutdown ()
	Shutdown the AI service connector.
static auto	is_initialized () noexcept -> bool
	Check if the connector is initialized.
static auto	request_inference (const inference_request &request) -> Result< std::string >
	Request AI inference for a study.
static auto	check_status (const std::string &job_id) -> Result< inference_status >
	Check the status of an inference job.
static auto	cancel (const std::string &job_id) -> Result< std::monostate >
	Cancel an inference job.
static auto	wait_for_completion (const std::string &job_id, std::chrono::milliseconds timeout=std::chrono::minutes{30}, status_callback callback=nullptr) -> Result< inference_status >
	Wait for a job to complete.
static auto	list_active_jobs () -> Result< std::vector< inference_status > >
	List active inference jobs.
static auto	list_models () -> Result< std::vector< model_info > >
	List available AI models.
static auto	get_model_info (const std::string &model_id) -> Result< model_info >
	Get information about a specific model.
static auto	check_health () -> bool
	Check AI service health.
static auto	get_latency () -> std::optional< std::chrono::milliseconds >
	Get current latency to the AI service.
static auto	get_config () -> const ai_service_config &
	Get the current configuration.
static auto	update_credentials (authentication_type auth_type, const std::string &credentials) -> Result< std::monostate >
	Update authentication credentials.

Private Member Functions
	ai_service_connector ()=delete
	~ai_service_connector ()=delete
	ai_service_connector (const ai_service_connector &)=delete
ai_service_connector &	operator= (const ai_service_connector &)=delete

Static Private Attributes
static std::unique_ptr< impl >	pimpl_

Detailed Description

Connector for external AI inference services.

This class provides a unified interface for interacting with external AI inference services, including:

• Sending DICOM studies for AI processing
• Tracking inference job status
• Cancelling running jobs
• Listing available AI models

The connector uses network_system for HTTP communication and integrates with logger_system for audit logging and monitoring_system for metrics.

Thread Safety: All methods are thread-safe.

Example:

// Initialize the connector
ai_service_config config;
config.base_url = "https://ai.hospital.org/api/v1";
config.auth_type = authentication_type::api_key;
config.api_key = "your-api-key";
ai_service_connector::initialize(config);
 
// Request inference
inference_request request;
request.study_instance_uid = "1.2.3.4.5.6.7.8.9";
request.model_id = "chest-xray-detector-v2";
 
auto result = ai_service_connector::request_inference(request);
if (result.is_ok()) {
    std::string job_id = result.value();
 
    // Poll for status
    auto status_result = ai_service_connector::check_status(job_id);
    if (status_result.is_ok()) {
        auto& status = status_result.value();
        if (status.status == inference_status_code::completed) {
            // Process results
        }
    }
}
 
// Shutdown
ai_service_connector::shutdown();

Definition at line 311 of file ai_service_connector.h.

Member Typedef Documentation

completion_callback

using kcenon::pacs::ai::ai_service_connector::completion_callback

Initial value:

 std::function<void(const std::string& job_id,
                                                   bool success,
                                                   const std::vector<std::string>& result_uids)>

Callback type for completion notification.

Definition at line 321 of file ai_service_connector.h.

status_callback

using kcenon::pacs::ai::ai_service_connector::status_callback = std::function<void(const inference_status&)>

Callback type for status updates.

Definition at line 318 of file ai_service_connector.h.

Constructor & Destructor Documentation

ai_service_connector() [1/2]

kcenon::pacs::ai::ai_service_connector::ai_service_connector ( )

privatedelete

~ai_service_connector()

kcenon::pacs::ai::ai_service_connector::~ai_service_connector ( )

privatedelete

ai_service_connector() [2/2]

kcenon::pacs::ai::ai_service_connector::ai_service_connector ( const ai_service_connector & )

privatedelete

Member Function Documentation

cancel()

static auto kcenon::pacs::ai::ai_service_connector::cancel ( const std::string & job_id ) -> Result< std::monostate >

staticnodiscard

Cancel an inference job.

Attempts to cancel a pending or running job. Jobs that have already completed cannot be cancelled.

Parameters

job_id

The job identifier to cancel

Returns

Result indicating success or failure

check_health()

static auto kcenon::pacs::ai::ai_service_connector::check_health ( ) -> bool

staticnodiscard

Check AI service health.

Returns

true if the service is healthy and accessible

check_status()

static auto kcenon::pacs::ai::ai_service_connector::check_status ( const std::string & job_id ) -> Result< inference_status >

staticnodiscard

Check the status of an inference job.

Parameters

job_id

The job identifier returned from request_inference

Returns

Result containing current status on success

get_config()

static auto kcenon::pacs::ai::ai_service_connector::get_config ( ) -> const ai_service_config &

staticnodiscard

Get the current configuration.

Returns

Current AI service configuration

get_latency()

static auto kcenon::pacs::ai::ai_service_connector::get_latency ( ) -> std::optional< std::chrono::milliseconds >

staticnodiscard

Get current latency to the AI service.

Returns

Round-trip time to the service, or nullopt if unavailable

get_model_info()

static auto kcenon::pacs::ai::ai_service_connector::get_model_info ( const std::string & model_id ) -> Result< model_info >

staticnodiscard

Get information about a specific model.

Parameters

model_id

The model identifier

Returns

Result containing model information

initialize()

static auto kcenon::pacs::ai::ai_service_connector::initialize ( const ai_service_config & config ) -> Result< std::monostate >

staticnodiscard

Initialize the AI service connector.

Must be called before any other operations. Sets up HTTP client, configures authentication, and validates connection.

Parameters

config

Configuration options

Returns

Result indicating success or initialization error

is_initialized()

static auto kcenon::pacs::ai::ai_service_connector::is_initialized ( ) -> bool

staticnodiscardnoexcept

Check if the connector is initialized.

Returns

true if initialized, false otherwise

list_active_jobs()

static auto kcenon::pacs::ai::ai_service_connector::list_active_jobs ( ) -> Result< std::vector< inference_status > >

staticnodiscard

List active inference jobs.

Returns all jobs that are currently pending or running.

Returns

Result containing list of active job statuses

list_models()

static auto kcenon::pacs::ai::ai_service_connector::list_models ( ) -> Result< std::vector< model_info > >

staticnodiscard

List available AI models.

Returns

Result containing list of available models

operator=()

ai_service_connector & kcenon::pacs::ai::ai_service_connector::operator= ( const ai_service_connector & )

privatedelete

request_inference()

static auto kcenon::pacs::ai::ai_service_connector::request_inference ( const inference_request & request ) -> Result< std::string >

staticnodiscard

Request AI inference for a study.

Submits a study for AI processing and returns a job ID for tracking.

Parameters

request

Inference request parameters

Returns

Result containing job ID on success, or error on failure

Note

The study must be accessible to the AI service (either via DICOM C-MOVE or DICOMweb WADO-RS)

shutdown()

static void kcenon::pacs::ai::ai_service_connector::shutdown ( )

static

Shutdown the AI service connector.

Cancels pending requests and releases resources.

update_credentials()

static auto kcenon::pacs::ai::ai_service_connector::update_credentials ( authentication_type auth_type, const std::string & credentials ) -> Result< std::monostate >

staticnodiscard

Update authentication credentials.

Parameters

auth_type	New authentication type
credentials	New credentials (API key, token, or username:password)

Returns

Result indicating success or failure

wait_for_completion()

static auto kcenon::pacs::ai::ai_service_connector::wait_for_completion ( const std::string & job_id, std::chrono::milliseconds timeout = std::chrono::minutes{30}, status_callback callback = nullptr ) -> Result< inference_status >

staticnodiscard

Wait for a job to complete.

Blocks until the job completes, fails, or times out.

Parameters

job_id	The job identifier to wait for
timeout	Maximum time to wait
callback	Optional status_callback for status updates

Returns

Result containing final status on completion

Member Data Documentation

pimpl_

std::unique_ptr<impl> kcenon::pacs::ai::ai_service_connector::pimpl_

staticprivate

Definition at line 490 of file ai_service_connector.h.

---

The documentation for this class was generated from the following file:

• include/kcenon/pacs/ai/ai_service_connector.h