API Reference

1. Feature Extraction APIs

extractor_init

Function

Creates and initializes a feature extractor.

Function API

feature_extractor *extractor_init(int link_type);

Parameter

• link_type: link layer type, which can be DLT_EN10MB(1) or DLT_LINUX_SLL(113).

Return Values

• A successful API call returns the handle of the feature extractor. A failed API call returns NULL.

extractor_destroy

Function

Destroys a feature extractor.

Function API

void extractor_destroy(feature_extractor *extractor);

Parameter

• extractor: feature extractor handle.

Return Values

• None

process_packet

Function

Processes a single packet by parsing the packet protocol and extracting features of a single flow.

Function API

int process_packet(feature_extractor *extractor, const unsigned char *packet, uint32_t packet_len,
                   const struct timeval *ts, feature_vector *features, int *has_feature);

Parameters

• extractor: feature extractor handle.
• packet: data packet pointer.
• packet_len: data packet length.
• ts: timestamp.
• features: output feature vector (single flow).
• has_feature: output flag, indicating whether feature extraction is performed. Feature extraction is performed only when a flow ends.

Return Values

• Status code: EXTRACTOR_SUCCESS indicates that the operation is successful, and EXTRACTOR_ERROR indicates that the operation fails.

process_packet_with_rawbow

Function

Processes a single data packet by parsing the packet protocol and extracting features of a single flow, including rawbow.

Function API

int process_packet_with_rawbow(feature_extractor *extractor, const unsigned char *packet, uint32_t packet_len,
                               const struct timeval *ts, feature_vector *features, rawbow *rawbows, int *has_feature);

Parameters

• extractor: feature extractor handle.
• packet: data packet pointer.
• packet_len: data packet length.
• ts: timestamp.
• features: output feature vector (single flow).
• rawbows: output rawbow (single flow).
• has_feature: output flag, indicating whether feature extraction is performed. Feature extraction is performed only when a flow ends.

Return Values

• Status code: EXTRACTOR_SUCCESS indicates that the operation is successful, and EXTRACTOR_ERROR indicates that the operation fails.

extractor_finalize

Function

Processes all remaining flows (only feature vectors are extracted).

Function API

int extractor_finalize(feature_extractor *extractor, feature_vector_list *features);

Parameters

• extractor: feature extractor handle.
• features: output feature vector list.

Return Values

• Status code: EXTRACTOR_SUCCESS indicates that the operation is successful, and EXTRACTOR_ERROR indicates that the operation fails.

extractor_finalize_with_rawbow

Function

Processes all remaining flows (feature vectors and rawbow are extracted).

Function API

int extractor_finalize_with_rawbow(feature_extractor *extractor, feature_vector_list *features, rawbow_list *rawbows);

Parameters

• extractor: feature extractor handle.
• features: output feature vector list.
• rawbows: output rawbow list.

Return Values

• Status code: EXTRACTOR_SUCCESS indicates that the operation is successful, and EXTRACTOR_ERROR indicates that the operation fails.

extractor_get_statistics

Function

Obtains statistics.

Function API

int extractor_get_statistics(feature_extractor *extractor, statistics *stats);

Parameters

• extractor: feature extractor handle.
• stats: output statistics.

Return Values

• Status code: EXTRACTOR_SUCCESS indicates that the operation is successful, and EXTRACTOR_ERROR indicates that the operation fails.

extractor_reset_statistics

Function

Resets statistics.

Function API

void extractor_reset_statistics(feature_extractor *extractor);

Parameter

• extractor: feature extractor handle.

Return Values

• None

2. Feature Inference APIs

inference_init

Function

Creates and initializes an inference engine.

Function API

inference_engine *inference_init(const inference_config *config);

Parameter

• config_file: inference configuration (including the configuration file path).

Return Values

• A successful API call returns the inference engine handle. A failed API call returns NULL.

inference_destroy

Function

Destroys an inference engine.

Function API

void inference_destroy(inference_engine *engine);

Parameter

• engine: inference engine handle.

Return Values

• None

inference_predict

Function

Performs batch inference.

Function API

int inference_predict(inference_engine *engine, const feature_vector_list *features, inference_result *result);

Parameters

• engine: inference engine handle.
• features: feature vector list.
• result: output inference result.

Return Values

• Status code: INFERENCE_SUCCESS indicates that the operation is successful, and INFERENCE_ERROR indicates that the operation fails.