2732 lines
90 KiB
JSON
2732 lines
90 KiB
JSON
{
|
|
"component": {
|
|
"kind": "component",
|
|
"name": "docling",
|
|
"title": "Docling",
|
|
"description": "Process documents using Docling library for parsing and conversion.",
|
|
"deprecated": false,
|
|
"firstVersion": "3.29.0",
|
|
"label": "transformation,ai",
|
|
"javaType": "org.apache.camel.component.docling.DoclingComponent",
|
|
"supportLevel": "Stable",
|
|
"nativeSupported": true,
|
|
"metadata": {
|
|
"quarkusVersion": "3.36.0",
|
|
"camelArtifact": "org.apache.camel:camel-docling",
|
|
"camelVersion": "4.20.0"
|
|
},
|
|
"groupId": "org.apache.camel.quarkus",
|
|
"artifactId": "camel-quarkus-docling",
|
|
"version": "4.20.0 (CEQ 3.36.0)",
|
|
"scheme": "docling",
|
|
"extendsScheme": "",
|
|
"syntax": "docling:operationId",
|
|
"async": false,
|
|
"api": false,
|
|
"consumerOnly": false,
|
|
"producerOnly": true,
|
|
"lenientProperties": false,
|
|
"browsable": false,
|
|
"remote": true
|
|
},
|
|
"componentProperties": {
|
|
"configuration": {
|
|
"index": 0,
|
|
"kind": "property",
|
|
"displayName": "Configuration",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "object",
|
|
"javaType": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"deprecated": false,
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "The configuration for the Docling Endpoint"
|
|
},
|
|
"contentInBody": {
|
|
"index": 1,
|
|
"kind": "property",
|
|
"displayName": "Content In Body",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Include the content of the output file in the exchange body and delete the output file"
|
|
},
|
|
"doclingServeUrl": {
|
|
"index": 2,
|
|
"kind": "property",
|
|
"displayName": "Docling Serve Url",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": "http://localhost:5001",
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Docling-serve API URL (e.g., http://localhost:5001)"
|
|
},
|
|
"enableOCR": {
|
|
"index": 3,
|
|
"kind": "property",
|
|
"displayName": "Enable OCR",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": true,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable OCR processing for scanned documents"
|
|
},
|
|
"includeLayoutInfo": {
|
|
"index": 4,
|
|
"kind": "property",
|
|
"displayName": "Include Layout Info",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Show layout information with bounding boxes"
|
|
},
|
|
"lazyStartProducer": {
|
|
"index": 5,
|
|
"kind": "property",
|
|
"displayName": "Lazy Start Producer",
|
|
"group": "producer",
|
|
"label": "producer",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"description": "Whether the producer should be started lazy (on the first message). By starting lazy you can use this to allow CamelContext and routes to startup in situations where a producer may otherwise fail during starting and cause the route to fail being started. By deferring this startup to be lazy then the startup failure can be handled during routing messages via Camel's routing error handlers. Beware that when the first message is processed then creating and starting the producer may take a little time and prolong the total processing time of the processing."
|
|
},
|
|
"ocrLanguage": {
|
|
"index": 6,
|
|
"kind": "property",
|
|
"displayName": "Ocr Language",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": "en",
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Language code for OCR processing"
|
|
},
|
|
"operation": {
|
|
"index": 7,
|
|
"kind": "property",
|
|
"displayName": "Operation",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": true,
|
|
"type": "enum",
|
|
"javaType": "org.apache.camel.component.docling.DoclingOperations",
|
|
"enum": [
|
|
"CONVERT_TO_MARKDOWN",
|
|
"CONVERT_TO_HTML",
|
|
"CONVERT_TO_JSON",
|
|
"EXTRACT_TEXT",
|
|
"EXTRACT_STRUCTURED_DATA",
|
|
"SUBMIT_ASYNC_CONVERSION",
|
|
"CHECK_CONVERSION_STATUS",
|
|
"BATCH_CONVERT_TO_MARKDOWN",
|
|
"BATCH_CONVERT_TO_HTML",
|
|
"BATCH_CONVERT_TO_JSON",
|
|
"BATCH_EXTRACT_TEXT",
|
|
"BATCH_EXTRACT_STRUCTURED_DATA",
|
|
"EXTRACT_METADATA",
|
|
"CHUNK_HYBRID",
|
|
"CHUNK_HIERARCHICAL"
|
|
],
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": "CONVERT_TO_MARKDOWN",
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "The operation to perform"
|
|
},
|
|
"outputFormat": {
|
|
"index": 8,
|
|
"kind": "property",
|
|
"displayName": "Output Format",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": "markdown",
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Output format for document conversion"
|
|
},
|
|
"useDoclingServe": {
|
|
"index": 9,
|
|
"kind": "property",
|
|
"displayName": "Use Docling Serve",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Use docling-serve API instead of CLI command"
|
|
},
|
|
"abortOnError": {
|
|
"index": 10,
|
|
"kind": "property",
|
|
"displayName": "Abort On Error",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Abort processing on error"
|
|
},
|
|
"asyncPollInterval": {
|
|
"index": 11,
|
|
"kind": "property",
|
|
"displayName": "Async Poll Interval",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": 2000,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Polling interval for async conversion status in milliseconds"
|
|
},
|
|
"asyncTimeout": {
|
|
"index": 12,
|
|
"kind": "property",
|
|
"displayName": "Async Timeout",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": 300000,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Maximum time to wait for async conversion completion in milliseconds"
|
|
},
|
|
"autowiredEnabled": {
|
|
"index": 13,
|
|
"kind": "property",
|
|
"displayName": "Autowired Enabled",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": true,
|
|
"description": "Whether autowiring is enabled. This is used for automatic autowiring options (the option must be marked as autowired) by looking up in the registry to find if there is a single instance of matching type, which then gets configured on the component. This can be used for automatic configuring JDBC data sources, JMS connection factories, AWS Clients, etc."
|
|
},
|
|
"doclingCommand": {
|
|
"index": 14,
|
|
"kind": "property",
|
|
"displayName": "Docling Command",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Path to Docling Python executable or command"
|
|
},
|
|
"doCodeEnrichment": {
|
|
"index": 15,
|
|
"kind": "property",
|
|
"displayName": "Do Code Enrichment",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable code enrichment in document processing"
|
|
},
|
|
"documentTimeout": {
|
|
"index": 16,
|
|
"kind": "property",
|
|
"displayName": "Document Timeout",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "java.lang.Long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Document processing timeout in seconds"
|
|
},
|
|
"doFormulaEnrichment": {
|
|
"index": 17,
|
|
"kind": "property",
|
|
"displayName": "Do Formula Enrichment",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable formula enrichment in document processing"
|
|
},
|
|
"doOcr": {
|
|
"index": 18,
|
|
"kind": "property",
|
|
"displayName": "Do Ocr",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable OCR processing in docling-serve API mode. When not set, the server uses its own defaults. Set enableOCR to false to explicitly disable OCR."
|
|
},
|
|
"doPictureClassification": {
|
|
"index": 19,
|
|
"kind": "property",
|
|
"displayName": "Do Picture Classification",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable picture classification in document processing"
|
|
},
|
|
"doPictureDescription": {
|
|
"index": 20,
|
|
"kind": "property",
|
|
"displayName": "Do Picture Description",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable picture description generation in document processing"
|
|
},
|
|
"doTableStructure": {
|
|
"index": 21,
|
|
"kind": "property",
|
|
"displayName": "Do Table Structure",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable table structure recognition"
|
|
},
|
|
"forceOcr": {
|
|
"index": 22,
|
|
"kind": "property",
|
|
"displayName": "Force Ocr",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Force OCR processing even for digital documents"
|
|
},
|
|
"imageExportMode": {
|
|
"index": 23,
|
|
"kind": "property",
|
|
"displayName": "Image Export Mode",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Image export mode for referenced images"
|
|
},
|
|
"imagesScale": {
|
|
"index": 24,
|
|
"kind": "property",
|
|
"displayName": "Images Scale",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "number",
|
|
"javaType": "java.lang.Double",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Scale factor for exported images"
|
|
},
|
|
"includeImages": {
|
|
"index": 25,
|
|
"kind": "property",
|
|
"displayName": "Include Images",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Include images in the conversion output"
|
|
},
|
|
"mdPageBreakPlaceholder": {
|
|
"index": 26,
|
|
"kind": "property",
|
|
"displayName": "Md Page Break Placeholder",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Placeholder string for page breaks in markdown output"
|
|
},
|
|
"ocrEngine": {
|
|
"index": 27,
|
|
"kind": "property",
|
|
"displayName": "Ocr Engine",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "OCR engine to use"
|
|
},
|
|
"pdfBackend": {
|
|
"index": 28,
|
|
"kind": "property",
|
|
"displayName": "Pdf Backend",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "PDF parsing backend"
|
|
},
|
|
"pipeline": {
|
|
"index": 29,
|
|
"kind": "property",
|
|
"displayName": "Pipeline",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Processing pipeline to use"
|
|
},
|
|
"processTimeout": {
|
|
"index": 30,
|
|
"kind": "property",
|
|
"displayName": "Process Timeout",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": 30000,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Timeout for Docling process execution in milliseconds"
|
|
},
|
|
"tableCellMatching": {
|
|
"index": 31,
|
|
"kind": "property",
|
|
"displayName": "Table Cell Matching",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable table cell matching post-processing"
|
|
},
|
|
"tableMode": {
|
|
"index": 32,
|
|
"kind": "property",
|
|
"displayName": "Table Mode",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Table structure recognition mode"
|
|
},
|
|
"useAsyncMode": {
|
|
"index": 33,
|
|
"kind": "property",
|
|
"displayName": "Use Async Mode",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Use asynchronous conversion mode (docling-serve API only)"
|
|
},
|
|
"workingDirectory": {
|
|
"index": 34,
|
|
"kind": "property",
|
|
"displayName": "Working Directory",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Working directory for Docling execution"
|
|
},
|
|
"batchFailOnFirstError": {
|
|
"index": 35,
|
|
"kind": "property",
|
|
"displayName": "Batch Fail On First Error",
|
|
"group": "batch",
|
|
"label": "batch",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": true,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Fail entire batch on first error (true) or continue processing remaining documents (false)"
|
|
},
|
|
"batchParallelism": {
|
|
"index": 36,
|
|
"kind": "property",
|
|
"displayName": "Batch Parallelism",
|
|
"group": "batch",
|
|
"label": "batch",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "int",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": 4,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Number of parallel threads for batch processing"
|
|
},
|
|
"batchSize": {
|
|
"index": 37,
|
|
"kind": "property",
|
|
"displayName": "Batch Size",
|
|
"group": "batch",
|
|
"label": "batch",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "int",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": 10,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Number of documents to submit per sub-batch. Documents are partitioned into sub-batches of this size and each sub-batch is processed before starting the next one. Within each sub-batch, up to batchParallelism threads are used concurrently. This controls memory usage and back-pressure when processing large document sets."
|
|
},
|
|
"batchTimeout": {
|
|
"index": 38,
|
|
"kind": "property",
|
|
"displayName": "Batch Timeout",
|
|
"group": "batch",
|
|
"label": "batch",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": 300000,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Maximum time to wait for batch completion in milliseconds"
|
|
},
|
|
"splitBatchResults": {
|
|
"index": 39,
|
|
"kind": "property",
|
|
"displayName": "Split Batch Results",
|
|
"group": "batch",
|
|
"label": "batch",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Split batch results into individual exchanges (one per document) instead of single BatchProcessingResults"
|
|
},
|
|
"chunkingIncludeRawText": {
|
|
"index": 40,
|
|
"kind": "property",
|
|
"displayName": "Chunking Include Raw Text",
|
|
"group": "chunking",
|
|
"label": "chunking",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Include raw text in chunk output"
|
|
},
|
|
"chunkingMaxTokens": {
|
|
"index": 41,
|
|
"kind": "property",
|
|
"displayName": "Chunking Max Tokens",
|
|
"group": "chunking",
|
|
"label": "chunking",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "java.lang.Integer",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Maximum number of tokens per chunk for hybrid chunking"
|
|
},
|
|
"chunkingMergePeers": {
|
|
"index": 42,
|
|
"kind": "property",
|
|
"displayName": "Chunking Merge Peers",
|
|
"group": "chunking",
|
|
"label": "chunking",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": true,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Whether to merge peer chunks in hybrid chunking"
|
|
},
|
|
"chunkingTokenizer": {
|
|
"index": 43,
|
|
"kind": "property",
|
|
"displayName": "Chunking Tokenizer",
|
|
"group": "chunking",
|
|
"label": "chunking",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Tokenizer model for hybrid chunking (e.g. sentence-transformers/all-MiniLM-L6-v2)"
|
|
},
|
|
"chunkingUseMarkdownTables": {
|
|
"index": 44,
|
|
"kind": "property",
|
|
"displayName": "Chunking Use Markdown Tables",
|
|
"group": "chunking",
|
|
"label": "chunking",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Use markdown format for tables in chunk output"
|
|
},
|
|
"includeMetadataInHeaders": {
|
|
"index": 45,
|
|
"kind": "property",
|
|
"displayName": "Include Metadata In Headers",
|
|
"group": "metadata",
|
|
"label": "metadata",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": true,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Include metadata in message headers when extracting metadata"
|
|
},
|
|
"includeRawMetadata": {
|
|
"index": 46,
|
|
"kind": "property",
|
|
"displayName": "Include Raw Metadata",
|
|
"group": "metadata",
|
|
"label": "metadata",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Include raw metadata as returned by the parser"
|
|
},
|
|
"apiKeyHeader": {
|
|
"index": 47,
|
|
"kind": "property",
|
|
"displayName": "Api Key Header",
|
|
"group": "security",
|
|
"label": "security",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": "X-API-Key",
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Header name for API key authentication"
|
|
},
|
|
"authenticationScheme": {
|
|
"index": 48,
|
|
"kind": "property",
|
|
"displayName": "Authentication Scheme",
|
|
"group": "security",
|
|
"label": "security",
|
|
"required": false,
|
|
"type": "enum",
|
|
"javaType": "org.apache.camel.component.docling.AuthenticationScheme",
|
|
"enum": [
|
|
"NONE",
|
|
"BEARER",
|
|
"API_KEY"
|
|
],
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": "NONE",
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Authentication scheme (BEARER, API_KEY, NONE)"
|
|
},
|
|
"authenticationToken": {
|
|
"index": 49,
|
|
"kind": "property",
|
|
"displayName": "Authentication Token",
|
|
"group": "security",
|
|
"label": "security",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": true,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Authentication token for docling-serve API (Bearer token or API key)"
|
|
},
|
|
"maxFileSize": {
|
|
"index": 50,
|
|
"kind": "property",
|
|
"displayName": "Max File Size",
|
|
"group": "security",
|
|
"label": "security",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": 52428800,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Maximum file size in bytes for processing"
|
|
},
|
|
"oauthProfile": {
|
|
"index": 51,
|
|
"kind": "property",
|
|
"displayName": "Oauth Profile",
|
|
"group": "security",
|
|
"label": "security",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "OAuth profile name for obtaining an access token via the OAuth 2.0 Client Credentials grant. When set, the token is acquired from the configured identity provider and used as authenticationToken. Requires camel-oauth on the classpath."
|
|
}
|
|
},
|
|
"headers": {
|
|
"CamelDoclingOperation": {
|
|
"index": 0,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "DoclingOperations",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "The operation to perform",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#OPERATION"
|
|
},
|
|
"CamelDoclingOutputFormat": {
|
|
"index": 1,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "The output format for conversion",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#OUTPUT_FORMAT"
|
|
},
|
|
"CamelDoclingInputFilePath": {
|
|
"index": 2,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "The input file path or content",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#INPUT_FILE_PATH"
|
|
},
|
|
"CamelDoclingOutputFilePath": {
|
|
"index": 3,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "The output file path for saving result",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#OUTPUT_FILE_PATH"
|
|
},
|
|
"CamelDoclingProcessingOptions": {
|
|
"index": 4,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Map<String, Object>",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Additional processing options",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#PROCESSING_OPTIONS"
|
|
},
|
|
"CamelDoclingEnableOCR": {
|
|
"index": 5,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Whether to include OCR processing",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#ENABLE_OCR"
|
|
},
|
|
"CamelDoclingOCRLanguage": {
|
|
"index": 6,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Language for OCR processing",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#OCR_LANGUAGE"
|
|
},
|
|
"CamelDoclingCustomArguments": {
|
|
"index": 7,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "List<String>",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Custom command line arguments to pass to Docling",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#CUSTOM_ARGUMENTS"
|
|
},
|
|
"CamelDoclingUseAsyncMode": {
|
|
"index": 8,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Use asynchronous conversion mode (overrides endpoint configuration)",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#USE_ASYNC_MODE"
|
|
},
|
|
"CamelDoclingAsyncPollInterval": {
|
|
"index": 9,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Polling interval for async conversion status in milliseconds",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#ASYNC_POLL_INTERVAL"
|
|
},
|
|
"CamelDoclingAsyncTimeout": {
|
|
"index": 10,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Maximum time to wait for async conversion completion in milliseconds",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#ASYNC_TIMEOUT"
|
|
},
|
|
"CamelDoclingTaskId": {
|
|
"index": 11,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Task ID for checking async conversion status",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#TASK_ID"
|
|
},
|
|
"CamelDoclingBatchSize": {
|
|
"index": 12,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Integer",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Override batch size for this operation",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#BATCH_SIZE"
|
|
},
|
|
"CamelDoclingBatchParallelism": {
|
|
"index": 13,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Integer",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Override batch parallelism for this operation",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#BATCH_PARALLELISM"
|
|
},
|
|
"CamelDoclingBatchFailOnFirstError": {
|
|
"index": 14,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Override batch fail on first error setting for this operation",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#BATCH_FAIL_ON_FIRST_ERROR"
|
|
},
|
|
"CamelDoclingBatchTimeout": {
|
|
"index": 15,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Override batch timeout for this operation in milliseconds",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#BATCH_TIMEOUT"
|
|
},
|
|
"CamelDoclingBatchTotalDocuments": {
|
|
"index": 16,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Integer",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Total number of documents in the batch",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#BATCH_TOTAL_DOCUMENTS"
|
|
},
|
|
"CamelDoclingBatchSuccessCount": {
|
|
"index": 17,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Integer",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Number of successfully processed documents in the batch",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#BATCH_SUCCESS_COUNT"
|
|
},
|
|
"CamelDoclingBatchFailureCount": {
|
|
"index": 18,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Integer",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Number of failed documents in the batch",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#BATCH_FAILURE_COUNT"
|
|
},
|
|
"CamelDoclingBatchProcessingTime": {
|
|
"index": 19,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Total processing time for the batch in milliseconds",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#BATCH_PROCESSING_TIME"
|
|
},
|
|
"CamelDoclingBatchSplitResults": {
|
|
"index": 20,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Split batch results into individual exchanges instead of single BatchProcessingResults",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#BATCH_SPLIT_RESULTS"
|
|
},
|
|
"CamelDoclingMetadataPageCount": {
|
|
"index": 21,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Integer",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Number of pages in the document",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#METADATA_PAGE_COUNT"
|
|
},
|
|
"CamelDoclingMetadataLanguage": {
|
|
"index": 22,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Document language code",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#METADATA_LANGUAGE"
|
|
},
|
|
"CamelDoclingMetadataDocumentType": {
|
|
"index": 23,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Document type/format",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#METADATA_DOCUMENT_TYPE"
|
|
},
|
|
"CamelDoclingMetadataFormat": {
|
|
"index": 24,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Document format (MIME type)",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#METADATA_FORMAT"
|
|
},
|
|
"CamelDoclingMetadataFileSize": {
|
|
"index": 25,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "File size in bytes",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#METADATA_FILE_SIZE"
|
|
},
|
|
"CamelDoclingMetadataFileName": {
|
|
"index": 26,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "File name",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#METADATA_FILE_NAME"
|
|
},
|
|
"CamelDoclingMetadataRaw": {
|
|
"index": 27,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Map<String, Object>",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Raw metadata fields as a Map",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#METADATA_RAW"
|
|
},
|
|
"CamelDoclingChunkingTokenizer": {
|
|
"index": 28,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Tokenizer for hybrid chunking (e.g. sentence-transformers/all-MiniLM-L6-v2)",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#CHUNKING_TOKENIZER"
|
|
},
|
|
"CamelDoclingChunkingMaxTokens": {
|
|
"index": 29,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Integer",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Maximum tokens per chunk for hybrid chunking",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#CHUNKING_MAX_TOKENS"
|
|
},
|
|
"CamelDoclingChunkingMergePeers": {
|
|
"index": 30,
|
|
"kind": "header",
|
|
"displayName": "",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"javaType": "Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "Whether to merge peer chunks in hybrid chunking",
|
|
"constantName": "org.apache.camel.component.docling.DoclingHeaders#CHUNKING_MERGE_PEERS"
|
|
}
|
|
},
|
|
"properties": {
|
|
"operationId": {
|
|
"index": 0,
|
|
"kind": "path",
|
|
"displayName": "Operation Id",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": true,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"description": "The operation identifier"
|
|
},
|
|
"contentInBody": {
|
|
"index": 1,
|
|
"kind": "parameter",
|
|
"displayName": "Content In Body",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Include the content of the output file in the exchange body and delete the output file"
|
|
},
|
|
"doclingServeUrl": {
|
|
"index": 2,
|
|
"kind": "parameter",
|
|
"displayName": "Docling Serve Url",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": "http://localhost:5001",
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Docling-serve API URL (e.g., http://localhost:5001)"
|
|
},
|
|
"enableOCR": {
|
|
"index": 3,
|
|
"kind": "parameter",
|
|
"displayName": "Enable OCR",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": true,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable OCR processing for scanned documents"
|
|
},
|
|
"includeLayoutInfo": {
|
|
"index": 4,
|
|
"kind": "parameter",
|
|
"displayName": "Include Layout Info",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Show layout information with bounding boxes"
|
|
},
|
|
"ocrLanguage": {
|
|
"index": 5,
|
|
"kind": "parameter",
|
|
"displayName": "Ocr Language",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": "en",
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Language code for OCR processing"
|
|
},
|
|
"operation": {
|
|
"index": 6,
|
|
"kind": "parameter",
|
|
"displayName": "Operation",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": true,
|
|
"type": "enum",
|
|
"javaType": "org.apache.camel.component.docling.DoclingOperations",
|
|
"enum": [
|
|
"CONVERT_TO_MARKDOWN",
|
|
"CONVERT_TO_HTML",
|
|
"CONVERT_TO_JSON",
|
|
"EXTRACT_TEXT",
|
|
"EXTRACT_STRUCTURED_DATA",
|
|
"SUBMIT_ASYNC_CONVERSION",
|
|
"CHECK_CONVERSION_STATUS",
|
|
"BATCH_CONVERT_TO_MARKDOWN",
|
|
"BATCH_CONVERT_TO_HTML",
|
|
"BATCH_CONVERT_TO_JSON",
|
|
"BATCH_EXTRACT_TEXT",
|
|
"BATCH_EXTRACT_STRUCTURED_DATA",
|
|
"EXTRACT_METADATA",
|
|
"CHUNK_HYBRID",
|
|
"CHUNK_HIERARCHICAL"
|
|
],
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": "CONVERT_TO_MARKDOWN",
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "The operation to perform"
|
|
},
|
|
"outputFormat": {
|
|
"index": 7,
|
|
"kind": "parameter",
|
|
"displayName": "Output Format",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": "markdown",
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Output format for document conversion"
|
|
},
|
|
"useDoclingServe": {
|
|
"index": 8,
|
|
"kind": "parameter",
|
|
"displayName": "Use Docling Serve",
|
|
"group": "producer",
|
|
"label": "",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Use docling-serve API instead of CLI command"
|
|
},
|
|
"lazyStartProducer": {
|
|
"index": 9,
|
|
"kind": "parameter",
|
|
"displayName": "Lazy Start Producer",
|
|
"group": "producer (advanced)",
|
|
"label": "producer,advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"description": "Whether the producer should be started lazy (on the first message). By starting lazy you can use this to allow CamelContext and routes to startup in situations where a producer may otherwise fail during starting and cause the route to fail being started. By deferring this startup to be lazy then the startup failure can be handled during routing messages via Camel's routing error handlers. Beware that when the first message is processed then creating and starting the producer may take a little time and prolong the total processing time of the processing."
|
|
},
|
|
"abortOnError": {
|
|
"index": 10,
|
|
"kind": "parameter",
|
|
"displayName": "Abort On Error",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Abort processing on error"
|
|
},
|
|
"asyncPollInterval": {
|
|
"index": 11,
|
|
"kind": "parameter",
|
|
"displayName": "Async Poll Interval",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": 2000,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Polling interval for async conversion status in milliseconds"
|
|
},
|
|
"asyncTimeout": {
|
|
"index": 12,
|
|
"kind": "parameter",
|
|
"displayName": "Async Timeout",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": 300000,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Maximum time to wait for async conversion completion in milliseconds"
|
|
},
|
|
"doclingCommand": {
|
|
"index": 13,
|
|
"kind": "parameter",
|
|
"displayName": "Docling Command",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Path to Docling Python executable or command"
|
|
},
|
|
"doCodeEnrichment": {
|
|
"index": 14,
|
|
"kind": "parameter",
|
|
"displayName": "Do Code Enrichment",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable code enrichment in document processing"
|
|
},
|
|
"documentTimeout": {
|
|
"index": 15,
|
|
"kind": "parameter",
|
|
"displayName": "Document Timeout",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "java.lang.Long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Document processing timeout in seconds"
|
|
},
|
|
"doFormulaEnrichment": {
|
|
"index": 16,
|
|
"kind": "parameter",
|
|
"displayName": "Do Formula Enrichment",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable formula enrichment in document processing"
|
|
},
|
|
"doOcr": {
|
|
"index": 17,
|
|
"kind": "parameter",
|
|
"displayName": "Do Ocr",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable OCR processing in docling-serve API mode. When not set, the server uses its own defaults. Set enableOCR to false to explicitly disable OCR."
|
|
},
|
|
"doPictureClassification": {
|
|
"index": 18,
|
|
"kind": "parameter",
|
|
"displayName": "Do Picture Classification",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable picture classification in document processing"
|
|
},
|
|
"doPictureDescription": {
|
|
"index": 19,
|
|
"kind": "parameter",
|
|
"displayName": "Do Picture Description",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable picture description generation in document processing"
|
|
},
|
|
"doTableStructure": {
|
|
"index": 20,
|
|
"kind": "parameter",
|
|
"displayName": "Do Table Structure",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable table structure recognition"
|
|
},
|
|
"forceOcr": {
|
|
"index": 21,
|
|
"kind": "parameter",
|
|
"displayName": "Force Ocr",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Force OCR processing even for digital documents"
|
|
},
|
|
"imageExportMode": {
|
|
"index": 22,
|
|
"kind": "parameter",
|
|
"displayName": "Image Export Mode",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Image export mode for referenced images"
|
|
},
|
|
"imagesScale": {
|
|
"index": 23,
|
|
"kind": "parameter",
|
|
"displayName": "Images Scale",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "number",
|
|
"javaType": "java.lang.Double",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Scale factor for exported images"
|
|
},
|
|
"includeImages": {
|
|
"index": 24,
|
|
"kind": "parameter",
|
|
"displayName": "Include Images",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Include images in the conversion output"
|
|
},
|
|
"mdPageBreakPlaceholder": {
|
|
"index": 25,
|
|
"kind": "parameter",
|
|
"displayName": "Md Page Break Placeholder",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Placeholder string for page breaks in markdown output"
|
|
},
|
|
"ocrEngine": {
|
|
"index": 26,
|
|
"kind": "parameter",
|
|
"displayName": "Ocr Engine",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "OCR engine to use"
|
|
},
|
|
"pdfBackend": {
|
|
"index": 27,
|
|
"kind": "parameter",
|
|
"displayName": "Pdf Backend",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "PDF parsing backend"
|
|
},
|
|
"pipeline": {
|
|
"index": 28,
|
|
"kind": "parameter",
|
|
"displayName": "Pipeline",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Processing pipeline to use"
|
|
},
|
|
"processTimeout": {
|
|
"index": 29,
|
|
"kind": "parameter",
|
|
"displayName": "Process Timeout",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": 30000,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Timeout for Docling process execution in milliseconds"
|
|
},
|
|
"tableCellMatching": {
|
|
"index": 30,
|
|
"kind": "parameter",
|
|
"displayName": "Table Cell Matching",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Enable table cell matching post-processing"
|
|
},
|
|
"tableMode": {
|
|
"index": 31,
|
|
"kind": "parameter",
|
|
"displayName": "Table Mode",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Table structure recognition mode"
|
|
},
|
|
"useAsyncMode": {
|
|
"index": 32,
|
|
"kind": "parameter",
|
|
"displayName": "Use Async Mode",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Use asynchronous conversion mode (docling-serve API only)"
|
|
},
|
|
"workingDirectory": {
|
|
"index": 33,
|
|
"kind": "parameter",
|
|
"displayName": "Working Directory",
|
|
"group": "advanced",
|
|
"label": "advanced",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Working directory for Docling execution"
|
|
},
|
|
"batchFailOnFirstError": {
|
|
"index": 34,
|
|
"kind": "parameter",
|
|
"displayName": "Batch Fail On First Error",
|
|
"group": "batch",
|
|
"label": "batch",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": true,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Fail entire batch on first error (true) or continue processing remaining documents (false)"
|
|
},
|
|
"batchParallelism": {
|
|
"index": 35,
|
|
"kind": "parameter",
|
|
"displayName": "Batch Parallelism",
|
|
"group": "batch",
|
|
"label": "batch",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "int",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": 4,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Number of parallel threads for batch processing"
|
|
},
|
|
"batchSize": {
|
|
"index": 36,
|
|
"kind": "parameter",
|
|
"displayName": "Batch Size",
|
|
"group": "batch",
|
|
"label": "batch",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "int",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": 10,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Number of documents to submit per sub-batch. Documents are partitioned into sub-batches of this size and each sub-batch is processed before starting the next one. Within each sub-batch, up to batchParallelism threads are used concurrently. This controls memory usage and back-pressure when processing large document sets."
|
|
},
|
|
"batchTimeout": {
|
|
"index": 37,
|
|
"kind": "parameter",
|
|
"displayName": "Batch Timeout",
|
|
"group": "batch",
|
|
"label": "batch",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": 300000,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Maximum time to wait for batch completion in milliseconds"
|
|
},
|
|
"splitBatchResults": {
|
|
"index": 38,
|
|
"kind": "parameter",
|
|
"displayName": "Split Batch Results",
|
|
"group": "batch",
|
|
"label": "batch",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Split batch results into individual exchanges (one per document) instead of single BatchProcessingResults"
|
|
},
|
|
"chunkingIncludeRawText": {
|
|
"index": 39,
|
|
"kind": "parameter",
|
|
"displayName": "Chunking Include Raw Text",
|
|
"group": "chunking",
|
|
"label": "chunking",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Include raw text in chunk output"
|
|
},
|
|
"chunkingMaxTokens": {
|
|
"index": 40,
|
|
"kind": "parameter",
|
|
"displayName": "Chunking Max Tokens",
|
|
"group": "chunking",
|
|
"label": "chunking",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "java.lang.Integer",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Maximum number of tokens per chunk for hybrid chunking"
|
|
},
|
|
"chunkingMergePeers": {
|
|
"index": 41,
|
|
"kind": "parameter",
|
|
"displayName": "Chunking Merge Peers",
|
|
"group": "chunking",
|
|
"label": "chunking",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": true,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Whether to merge peer chunks in hybrid chunking"
|
|
},
|
|
"chunkingTokenizer": {
|
|
"index": 42,
|
|
"kind": "parameter",
|
|
"displayName": "Chunking Tokenizer",
|
|
"group": "chunking",
|
|
"label": "chunking",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Tokenizer model for hybrid chunking (e.g. sentence-transformers/all-MiniLM-L6-v2)"
|
|
},
|
|
"chunkingUseMarkdownTables": {
|
|
"index": 43,
|
|
"kind": "parameter",
|
|
"displayName": "Chunking Use Markdown Tables",
|
|
"group": "chunking",
|
|
"label": "chunking",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "java.lang.Boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Use markdown format for tables in chunk output"
|
|
},
|
|
"includeMetadataInHeaders": {
|
|
"index": 44,
|
|
"kind": "parameter",
|
|
"displayName": "Include Metadata In Headers",
|
|
"group": "metadata",
|
|
"label": "metadata",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": true,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Include metadata in message headers when extracting metadata"
|
|
},
|
|
"includeRawMetadata": {
|
|
"index": 45,
|
|
"kind": "parameter",
|
|
"displayName": "Include Raw Metadata",
|
|
"group": "metadata",
|
|
"label": "metadata",
|
|
"required": false,
|
|
"type": "boolean",
|
|
"javaType": "boolean",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Include raw metadata as returned by the parser"
|
|
},
|
|
"apiKeyHeader": {
|
|
"index": 46,
|
|
"kind": "parameter",
|
|
"displayName": "Api Key Header",
|
|
"group": "security",
|
|
"label": "security",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": "X-API-Key",
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Header name for API key authentication"
|
|
},
|
|
"authenticationScheme": {
|
|
"index": 47,
|
|
"kind": "parameter",
|
|
"displayName": "Authentication Scheme",
|
|
"group": "security",
|
|
"label": "security",
|
|
"required": false,
|
|
"type": "enum",
|
|
"javaType": "org.apache.camel.component.docling.AuthenticationScheme",
|
|
"enum": [
|
|
"NONE",
|
|
"BEARER",
|
|
"API_KEY"
|
|
],
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": "NONE",
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Authentication scheme (BEARER, API_KEY, NONE)"
|
|
},
|
|
"authenticationToken": {
|
|
"index": 48,
|
|
"kind": "parameter",
|
|
"displayName": "Authentication Token",
|
|
"group": "security",
|
|
"label": "security",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": true,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Authentication token for docling-serve API (Bearer token or API key)"
|
|
},
|
|
"maxFileSize": {
|
|
"index": 49,
|
|
"kind": "parameter",
|
|
"displayName": "Max File Size",
|
|
"group": "security",
|
|
"label": "security",
|
|
"required": false,
|
|
"type": "integer",
|
|
"javaType": "long",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"defaultValue": 52428800,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "Maximum file size in bytes for processing"
|
|
},
|
|
"oauthProfile": {
|
|
"index": 50,
|
|
"kind": "parameter",
|
|
"displayName": "Oauth Profile",
|
|
"group": "security",
|
|
"label": "security",
|
|
"required": false,
|
|
"type": "string",
|
|
"javaType": "java.lang.String",
|
|
"deprecated": false,
|
|
"deprecationNote": "",
|
|
"autowired": false,
|
|
"secret": false,
|
|
"configurationClass": "org.apache.camel.component.docling.DoclingConfiguration",
|
|
"configurationField": "configuration",
|
|
"description": "OAuth profile name for obtaining an access token via the OAuth 2.0 Client Credentials grant. When set, the token is acquired from the configured identity provider and used as authenticationToken. Requires camel-oauth on the classpath."
|
|
}
|
|
},
|
|
"propertiesSchema": {
|
|
"properties": {
|
|
"operationId": {
|
|
"title": "Operation Id",
|
|
"description": "The operation identifier",
|
|
"type": "string",
|
|
"$comment": "group:producer"
|
|
},
|
|
"contentInBody": {
|
|
"title": "Content In Body",
|
|
"description": "Include the content of the output file in the exchange body and delete the output file",
|
|
"type": "boolean",
|
|
"$comment": "group:producer",
|
|
"default": false
|
|
},
|
|
"doclingServeUrl": {
|
|
"title": "Docling Serve Url",
|
|
"description": "Docling-serve API URL (e.g., http://localhost:5001)",
|
|
"type": "string",
|
|
"$comment": "group:producer",
|
|
"default": "http://localhost:5001"
|
|
},
|
|
"enableOCR": {
|
|
"title": "Enable OCR",
|
|
"description": "Enable OCR processing for scanned documents",
|
|
"type": "boolean",
|
|
"$comment": "group:producer",
|
|
"default": true
|
|
},
|
|
"includeLayoutInfo": {
|
|
"title": "Include Layout Info",
|
|
"description": "Show layout information with bounding boxes",
|
|
"type": "boolean",
|
|
"$comment": "group:producer",
|
|
"default": false
|
|
},
|
|
"ocrLanguage": {
|
|
"title": "Ocr Language",
|
|
"description": "Language code for OCR processing",
|
|
"type": "string",
|
|
"$comment": "group:producer",
|
|
"default": "en"
|
|
},
|
|
"operation": {
|
|
"title": "Operation",
|
|
"description": "The operation to perform",
|
|
"type": "string",
|
|
"enum": [
|
|
"CONVERT_TO_MARKDOWN",
|
|
"CONVERT_TO_HTML",
|
|
"CONVERT_TO_JSON",
|
|
"EXTRACT_TEXT",
|
|
"EXTRACT_STRUCTURED_DATA",
|
|
"SUBMIT_ASYNC_CONVERSION",
|
|
"CHECK_CONVERSION_STATUS",
|
|
"BATCH_CONVERT_TO_MARKDOWN",
|
|
"BATCH_CONVERT_TO_HTML",
|
|
"BATCH_CONVERT_TO_JSON",
|
|
"BATCH_EXTRACT_TEXT",
|
|
"BATCH_EXTRACT_STRUCTURED_DATA",
|
|
"EXTRACT_METADATA",
|
|
"CHUNK_HYBRID",
|
|
"CHUNK_HIERARCHICAL"
|
|
],
|
|
"$comment": "group:producer",
|
|
"default": "CONVERT_TO_MARKDOWN"
|
|
},
|
|
"outputFormat": {
|
|
"title": "Output Format",
|
|
"description": "Output format for document conversion",
|
|
"type": "string",
|
|
"$comment": "group:producer",
|
|
"default": "markdown"
|
|
},
|
|
"useDoclingServe": {
|
|
"title": "Use Docling Serve",
|
|
"description": "Use docling-serve API instead of CLI command",
|
|
"type": "boolean",
|
|
"$comment": "group:producer",
|
|
"default": false
|
|
},
|
|
"lazyStartProducer": {
|
|
"title": "Lazy Start Producer",
|
|
"description": "Whether the producer should be started lazy (on the first message). By starting lazy you can use this to allow CamelContext and routes to startup in situations where a producer may otherwise fail during starting and cause the route to fail being started. By deferring this startup to be lazy then the startup failure can be handled during routing messages via Camel's routing error handlers. Beware that when the first message is processed then creating and starting the producer may take a little time and prolong the total processing time of the processing.",
|
|
"type": "boolean",
|
|
"$comment": "group:producer (advanced)",
|
|
"default": false
|
|
},
|
|
"abortOnError": {
|
|
"title": "Abort On Error",
|
|
"description": "Abort processing on error",
|
|
"type": "boolean",
|
|
"$comment": "group:advanced",
|
|
"default": false
|
|
},
|
|
"asyncPollInterval": {
|
|
"title": "Async Poll Interval",
|
|
"description": "Polling interval for async conversion status in milliseconds",
|
|
"type": "integer",
|
|
"$comment": "group:advanced",
|
|
"default": 2000
|
|
},
|
|
"asyncTimeout": {
|
|
"title": "Async Timeout",
|
|
"description": "Maximum time to wait for async conversion completion in milliseconds",
|
|
"type": "integer",
|
|
"$comment": "group:advanced",
|
|
"default": 300000
|
|
},
|
|
"doclingCommand": {
|
|
"title": "Docling Command",
|
|
"description": "Path to Docling Python executable or command",
|
|
"type": "string",
|
|
"$comment": "group:advanced"
|
|
},
|
|
"doCodeEnrichment": {
|
|
"title": "Do Code Enrichment",
|
|
"description": "Enable code enrichment in document processing",
|
|
"type": "boolean",
|
|
"$comment": "group:advanced",
|
|
"default": false
|
|
},
|
|
"documentTimeout": {
|
|
"title": "Document Timeout",
|
|
"description": "Document processing timeout in seconds",
|
|
"type": "integer",
|
|
"$comment": "group:advanced"
|
|
},
|
|
"doFormulaEnrichment": {
|
|
"title": "Do Formula Enrichment",
|
|
"description": "Enable formula enrichment in document processing",
|
|
"type": "boolean",
|
|
"$comment": "group:advanced",
|
|
"default": false
|
|
},
|
|
"doOcr": {
|
|
"title": "Do Ocr",
|
|
"description": "Enable OCR processing in docling-serve API mode. When not set, the server uses its own defaults. Set enableOCR to false to explicitly disable OCR.",
|
|
"type": "boolean",
|
|
"$comment": "group:advanced",
|
|
"default": false
|
|
},
|
|
"doPictureClassification": {
|
|
"title": "Do Picture Classification",
|
|
"description": "Enable picture classification in document processing",
|
|
"type": "boolean",
|
|
"$comment": "group:advanced",
|
|
"default": false
|
|
},
|
|
"doPictureDescription": {
|
|
"title": "Do Picture Description",
|
|
"description": "Enable picture description generation in document processing",
|
|
"type": "boolean",
|
|
"$comment": "group:advanced",
|
|
"default": false
|
|
},
|
|
"doTableStructure": {
|
|
"title": "Do Table Structure",
|
|
"description": "Enable table structure recognition",
|
|
"type": "boolean",
|
|
"$comment": "group:advanced",
|
|
"default": false
|
|
},
|
|
"forceOcr": {
|
|
"title": "Force Ocr",
|
|
"description": "Force OCR processing even for digital documents",
|
|
"type": "boolean",
|
|
"$comment": "group:advanced",
|
|
"default": false
|
|
},
|
|
"imageExportMode": {
|
|
"title": "Image Export Mode",
|
|
"description": "Image export mode for referenced images",
|
|
"type": "string",
|
|
"$comment": "group:advanced"
|
|
},
|
|
"imagesScale": {
|
|
"title": "Images Scale",
|
|
"description": "Scale factor for exported images",
|
|
"type": "number",
|
|
"$comment": "group:advanced"
|
|
},
|
|
"includeImages": {
|
|
"title": "Include Images",
|
|
"description": "Include images in the conversion output",
|
|
"type": "boolean",
|
|
"$comment": "group:advanced",
|
|
"default": false
|
|
},
|
|
"mdPageBreakPlaceholder": {
|
|
"title": "Md Page Break Placeholder",
|
|
"description": "Placeholder string for page breaks in markdown output",
|
|
"type": "string",
|
|
"$comment": "group:advanced"
|
|
},
|
|
"ocrEngine": {
|
|
"title": "Ocr Engine",
|
|
"description": "OCR engine to use",
|
|
"type": "string",
|
|
"$comment": "group:advanced"
|
|
},
|
|
"pdfBackend": {
|
|
"title": "Pdf Backend",
|
|
"description": "PDF parsing backend",
|
|
"type": "string",
|
|
"$comment": "group:advanced"
|
|
},
|
|
"pipeline": {
|
|
"title": "Pipeline",
|
|
"description": "Processing pipeline to use",
|
|
"type": "string",
|
|
"$comment": "group:advanced"
|
|
},
|
|
"processTimeout": {
|
|
"title": "Process Timeout",
|
|
"description": "Timeout for Docling process execution in milliseconds",
|
|
"type": "integer",
|
|
"$comment": "group:advanced",
|
|
"default": 30000
|
|
},
|
|
"tableCellMatching": {
|
|
"title": "Table Cell Matching",
|
|
"description": "Enable table cell matching post-processing",
|
|
"type": "boolean",
|
|
"$comment": "group:advanced",
|
|
"default": false
|
|
},
|
|
"tableMode": {
|
|
"title": "Table Mode",
|
|
"description": "Table structure recognition mode",
|
|
"type": "string",
|
|
"$comment": "group:advanced"
|
|
},
|
|
"useAsyncMode": {
|
|
"title": "Use Async Mode",
|
|
"description": "Use asynchronous conversion mode (docling-serve API only)",
|
|
"type": "boolean",
|
|
"$comment": "group:advanced",
|
|
"default": false
|
|
},
|
|
"workingDirectory": {
|
|
"title": "Working Directory",
|
|
"description": "Working directory for Docling execution",
|
|
"type": "string",
|
|
"$comment": "group:advanced"
|
|
},
|
|
"batchFailOnFirstError": {
|
|
"title": "Batch Fail On First Error",
|
|
"description": "Fail entire batch on first error (true) or continue processing remaining documents (false)",
|
|
"type": "boolean",
|
|
"$comment": "group:batch",
|
|
"default": true
|
|
},
|
|
"batchParallelism": {
|
|
"title": "Batch Parallelism",
|
|
"description": "Number of parallel threads for batch processing",
|
|
"type": "integer",
|
|
"$comment": "group:batch",
|
|
"default": 4
|
|
},
|
|
"batchSize": {
|
|
"title": "Batch Size",
|
|
"description": "Number of documents to submit per sub-batch. Documents are partitioned into sub-batches of this size and each sub-batch is processed before starting the next one. Within each sub-batch, up to batchParallelism threads are used concurrently. This controls memory usage and back-pressure when processing large document sets.",
|
|
"type": "integer",
|
|
"$comment": "group:batch",
|
|
"default": 10
|
|
},
|
|
"batchTimeout": {
|
|
"title": "Batch Timeout",
|
|
"description": "Maximum time to wait for batch completion in milliseconds",
|
|
"type": "integer",
|
|
"$comment": "group:batch",
|
|
"default": 300000
|
|
},
|
|
"splitBatchResults": {
|
|
"title": "Split Batch Results",
|
|
"description": "Split batch results into individual exchanges (one per document) instead of single BatchProcessingResults",
|
|
"type": "boolean",
|
|
"$comment": "group:batch",
|
|
"default": false
|
|
},
|
|
"chunkingIncludeRawText": {
|
|
"title": "Chunking Include Raw Text",
|
|
"description": "Include raw text in chunk output",
|
|
"type": "boolean",
|
|
"$comment": "group:chunking",
|
|
"default": false
|
|
},
|
|
"chunkingMaxTokens": {
|
|
"title": "Chunking Max Tokens",
|
|
"description": "Maximum number of tokens per chunk for hybrid chunking",
|
|
"type": "integer",
|
|
"$comment": "group:chunking"
|
|
},
|
|
"chunkingMergePeers": {
|
|
"title": "Chunking Merge Peers",
|
|
"description": "Whether to merge peer chunks in hybrid chunking",
|
|
"type": "boolean",
|
|
"$comment": "group:chunking",
|
|
"default": true
|
|
},
|
|
"chunkingTokenizer": {
|
|
"title": "Chunking Tokenizer",
|
|
"description": "Tokenizer model for hybrid chunking (e.g. sentence-transformers/all-MiniLM-L6-v2)",
|
|
"type": "string",
|
|
"$comment": "group:chunking"
|
|
},
|
|
"chunkingUseMarkdownTables": {
|
|
"title": "Chunking Use Markdown Tables",
|
|
"description": "Use markdown format for tables in chunk output",
|
|
"type": "boolean",
|
|
"$comment": "group:chunking",
|
|
"default": false
|
|
},
|
|
"includeMetadataInHeaders": {
|
|
"title": "Include Metadata In Headers",
|
|
"description": "Include metadata in message headers when extracting metadata",
|
|
"type": "boolean",
|
|
"$comment": "group:metadata",
|
|
"default": true
|
|
},
|
|
"includeRawMetadata": {
|
|
"title": "Include Raw Metadata",
|
|
"description": "Include raw metadata as returned by the parser",
|
|
"type": "boolean",
|
|
"$comment": "group:metadata",
|
|
"default": false
|
|
},
|
|
"apiKeyHeader": {
|
|
"title": "Api Key Header",
|
|
"description": "Header name for API key authentication",
|
|
"type": "string",
|
|
"$comment": "group:security",
|
|
"default": "X-API-Key"
|
|
},
|
|
"authenticationScheme": {
|
|
"title": "Authentication Scheme",
|
|
"description": "Authentication scheme (BEARER, API_KEY, NONE)",
|
|
"type": "string",
|
|
"enum": [
|
|
"NONE",
|
|
"BEARER",
|
|
"API_KEY"
|
|
],
|
|
"$comment": "group:security",
|
|
"default": "NONE"
|
|
},
|
|
"authenticationToken": {
|
|
"title": "Authentication Token",
|
|
"description": "Authentication token for docling-serve API (Bearer token or API key)",
|
|
"type": "string",
|
|
"$comment": "group:security",
|
|
"format": "password"
|
|
},
|
|
"maxFileSize": {
|
|
"title": "Max File Size",
|
|
"description": "Maximum file size in bytes for processing",
|
|
"type": "integer",
|
|
"$comment": "group:security",
|
|
"default": 52428800
|
|
},
|
|
"oauthProfile": {
|
|
"title": "Oauth Profile",
|
|
"description": "OAuth profile name for obtaining an access token via the OAuth 2.0 Client Credentials grant. When set, the token is acquired from the configured identity provider and used as authenticationToken. Requires camel-oauth on the classpath.",
|
|
"type": "string",
|
|
"$comment": "group:security"
|
|
}
|
|
},
|
|
"$schema": "http://json-schema.org/draft-07/schema#",
|
|
"type": "object",
|
|
"required": [
|
|
"operationId",
|
|
"operation"
|
|
]
|
|
}
|
|
} |