ainsoft-rag-spring-boot-autoconfigure/com.ainsoft.rag.spring/RagProperties

RagProperties

@ConfigurationProperties(prefix = "rag")

data class RagProperties(val indexPath: String = "./rag-index", val storeChunkText: Boolean = true, val uploadMaxBytes: Long = 10 * 1024 * 1024, val uploadAllowedContentTypes: List<String> = listOf( "text/plain", "application/pdf", "application/vnd.openxmlformats-officedocument.wordprocessingml.document", "application/vnd.openxmlformats-officedocument.presentationml.presentation" ), val sourceLoadTimeoutMillis: Long, val sourceLoadAllowHosts: List<String> = emptyList(), val sourceLoadAuthHeaders: Map<String, String> = emptyMap(), val sourceLoadInsecureSkipTlsVerify: Boolean = false, val sourceLoadCustomCaCertPath: String? = null, val sourceLoadDefaultProfile: String? = null, val sourceLoadProfiles: Map<String, SourceLoadProfile> = emptyMap(), val statsCacheTtlMillis: Long, val statsCacheMaxEntries: Int = 32, val statsCacheMaxEntriesPerTenant: Int = 4, val contextualRetrievalEnabled: Boolean = true, val contextualIncludeDocId: Boolean = true, val contextualIncludeMetadataContext: Boolean = true, val contextualMaxDocumentSummaryChars: Int = 240, val rerankerEnabled: Boolean = true, val rerankerType: String = "heuristic", val rerankerTopN: Int = 24, val rerankerAlpha: Double = 0.65, val rerankerApiBaseUrl: String = "https://api.cohere.com", val rerankerApiKey: String? = null, val rerankerModel: String = "rerank-v3.5", val rerankerRequestTimeoutMillis: Long, val rerankerOnnxQueryInputName: String? = null, val rerankerOnnxDocumentInputName: String? = null, val rerankerOnnxOutputName: String? = null, val rerankerOnnxExpectedDimensions: Int = 0, val rerankerOnnxExpectedInputSchema: String? = null, val rerankerOnnxExpectedTokenizer: String? = null, val rerankerOnnxExpectedScoreSemantics: String? = null, val rerankerOnnxExpectedTokenizerVocabChecksum: String? = null, val rerankerOnnxTokenizerSchemaChecksum: String? = null, val rerankerOnnxModelContractChecksum: String? = null, val correctiveRetrievalEnabled: Boolean = true, val correctiveMinConfidence: Double = 0.08, val correctiveMinResultsBeforeSkip: Int = 3, val correctiveExpandedCandidateMultiplier: Int = 2, val queryRewriteEnabled: Boolean = true, val queryRewriterType: String = "heuristic", val queryRewriteApiBaseUrl: String = "https://api.openai.com/v1", val queryRewriteApiKey: String? = null, val queryRewriteModel: String = "gpt-4o-mini", val queryRewriteRequestTimeoutMillis: Long, val hierarchicalSummariesEnabled: Boolean = true, val hierarchicalMaxSectionSummaries: Int = 4, val hierarchicalMaxSummaryChars: Int = 280, val hierarchicalTargetChunksPerSection: Int = 3, val summarizerType: String = "rule-based", val summarizerApiBaseUrl: String = "https://api.openai.com/v1", val summarizerApiKey: String? = null, val summarizerModel: String = "gpt-4o-mini", val summarizerRequestTimeoutMillis: Long, val statsCacheStoreType: String = "memory", val statsCacheFilePath: String? = null, val statsCacheFileMaxBytes: Long = 0, val statsCacheFileRotateCount: Int = 2, val statsCacheFileCleanupOnStart: Boolean = false, val providerHealthAutoExportPath: String? = null, val providerHealthAutoExportIntervalMillis: Long = 0, val providerHealthAutoExportWindowMillis: Long? = null, val providerHealthAutoExportFormat: String = "json", val providerHealthAutoExportRetainCount: Int = 5, val providerHealthAutoExportIncludeScopeSuffix: Boolean = false, val providerHealthAutoExportPushUrl: String? = null, val providerHealthAutoExportPushFormat: String = "json", val providerHealthAutoExportPushTimeoutMillis: Long, val providerHealthAutoExportPushHeaders: Map<String, String> = emptyMap(), val providerHealthAutoExportPushMaxRetries: Int = 2, val providerHealthAutoExportPushRetryBackoffMillis: Long = 250, val providerHealthAutoExportPushDeadLetterPath: String? = null, val providerHealthAutoExportPushDeadLetterRetainCount: Int = 5, val providerHealthAutoExportPushAsyncEnabled: Boolean = false, val providerHealthAutoExportPushQueueCapacity: Int = 32, val providerHealthAutoExportPushDropOldestOnOverflow: Boolean = true, val providerHealthAutoExportPushHmacSecret: String? = null, val providerHealthAutoExportPushHmacHeaderName: String = "X-Rag-Signature", val providerHealthAutoExportPushTimestampHeaderName: String = "X-Rag-Timestamp", val chunkerType: String = "basic", val basicMaxChars: Int = 1800, val basicOverlapChars: Int = 200, val slidingWindowSize: Int = 240, val slidingOverlap: Int = 40, val adaptiveSectionRegex: String = com.ainsoft.rag.chunking.Chunkers.DEFAULT_ADAPTIVE_SECTION_REGEX, val adaptiveMinChunkSize: Int = 200, val adaptiveMaxChunkSize: Int = 1000, val regexSplitPattern: String = "(?m)^##?", val regexGroupByPattern: String? = null, val embeddingDimensions: Int = 256, val embeddingProvider: String = "hash", val openAiApiKey: String? = null, val openAiModel: String = "text-embedding-3-small", val openAiBaseUrl: String = "https://api.openai.com/v1")

Constructors

RagProperties

constructor(indexPath: String = "./rag-index", storeChunkText: Boolean = true, uploadMaxBytes: Long = 10 * 1024 * 1024, uploadAllowedContentTypes: List<String> = listOf( "text/plain", "application/pdf", "application/vnd.openxmlformats-officedocument.wordprocessingml.document", "application/vnd.openxmlformats-officedocument.presentationml.presentation" ), sourceLoadTimeoutMillis: Long, sourceLoadAllowHosts: List<String> = emptyList(), sourceLoadAuthHeaders: Map<String, String> = emptyMap(), sourceLoadInsecureSkipTlsVerify: Boolean = false, sourceLoadCustomCaCertPath: String? = null, sourceLoadDefaultProfile: String? = null, sourceLoadProfiles: Map<String, SourceLoadProfile> = emptyMap(), statsCacheTtlMillis: Long, statsCacheMaxEntries: Int = 32, statsCacheMaxEntriesPerTenant: Int = 4, contextualRetrievalEnabled: Boolean = true, contextualIncludeDocId: Boolean = true, contextualIncludeMetadataContext: Boolean = true, contextualMaxDocumentSummaryChars: Int = 240, rerankerEnabled: Boolean = true, rerankerType: String = "heuristic", rerankerTopN: Int = 24, rerankerAlpha: Double = 0.65, rerankerApiBaseUrl: String = "https://api.cohere.com", rerankerApiKey: String? = null, rerankerModel: String = "rerank-v3.5", rerankerRequestTimeoutMillis: Long, rerankerOnnxQueryInputName: String? = null, rerankerOnnxDocumentInputName: String? = null, rerankerOnnxOutputName: String? = null, rerankerOnnxExpectedDimensions: Int = 0, rerankerOnnxExpectedInputSchema: String? = null, rerankerOnnxExpectedTokenizer: String? = null, rerankerOnnxExpectedScoreSemantics: String? = null, rerankerOnnxExpectedTokenizerVocabChecksum: String? = null, rerankerOnnxTokenizerSchemaChecksum: String? = null, rerankerOnnxModelContractChecksum: String? = null, correctiveRetrievalEnabled: Boolean = true, correctiveMinConfidence: Double = 0.08, correctiveMinResultsBeforeSkip: Int = 3, correctiveExpandedCandidateMultiplier: Int = 2, queryRewriteEnabled: Boolean = true, queryRewriterType: String = "heuristic", queryRewriteApiBaseUrl: String = "https://api.openai.com/v1", queryRewriteApiKey: String? = null, queryRewriteModel: String = "gpt-4o-mini", queryRewriteRequestTimeoutMillis: Long, hierarchicalSummariesEnabled: Boolean = true, hierarchicalMaxSectionSummaries: Int = 4, hierarchicalMaxSummaryChars: Int = 280, hierarchicalTargetChunksPerSection: Int = 3, summarizerType: String = "rule-based", summarizerApiBaseUrl: String = "https://api.openai.com/v1", summarizerApiKey: String? = null, summarizerModel: String = "gpt-4o-mini", summarizerRequestTimeoutMillis: Long, statsCacheStoreType: String = "memory", statsCacheFilePath: String? = null, statsCacheFileMaxBytes: Long = 0, statsCacheFileRotateCount: Int = 2, statsCacheFileCleanupOnStart: Boolean = false, providerHealthAutoExportPath: String? = null, providerHealthAutoExportIntervalMillis: Long = 0, providerHealthAutoExportWindowMillis: Long? = null, providerHealthAutoExportFormat: String = "json", providerHealthAutoExportRetainCount: Int = 5, providerHealthAutoExportIncludeScopeSuffix: Boolean = false, providerHealthAutoExportPushUrl: String? = null, providerHealthAutoExportPushFormat: String = "json", providerHealthAutoExportPushTimeoutMillis: Long, providerHealthAutoExportPushHeaders: Map<String, String> = emptyMap(), providerHealthAutoExportPushMaxRetries: Int = 2, providerHealthAutoExportPushRetryBackoffMillis: Long = 250, providerHealthAutoExportPushDeadLetterPath: String? = null, providerHealthAutoExportPushDeadLetterRetainCount: Int = 5, providerHealthAutoExportPushAsyncEnabled: Boolean = false, providerHealthAutoExportPushQueueCapacity: Int = 32, providerHealthAutoExportPushDropOldestOnOverflow: Boolean = true, providerHealthAutoExportPushHmacSecret: String? = null, providerHealthAutoExportPushHmacHeaderName: String = "X-Rag-Signature", providerHealthAutoExportPushTimestampHeaderName: String = "X-Rag-Timestamp", chunkerType: String = "basic", basicMaxChars: Int = 1800, basicOverlapChars: Int = 200, slidingWindowSize: Int = 240, slidingOverlap: Int = 40, adaptiveSectionRegex: String = com.ainsoft.rag.chunking.Chunkers.DEFAULT_ADAPTIVE_SECTION_REGEX, adaptiveMinChunkSize: Int = 200, adaptiveMaxChunkSize: Int = 1000, regexSplitPattern: String = "(?m)^##?", regexGroupByPattern: String? = null, embeddingDimensions: Int = 256, embeddingProvider: String = "hash", openAiApiKey: String? = null, openAiModel: String = "text-embedding-3-small", openAiBaseUrl: String = "https://api.openai.com/v1")

RagProperties

Constructors

Properties

Functions