Slice object information
Used by actions: ChunkDocument.
| Name | Type | Required | Description |
|---|---|---|---|
| Index | Integer | No | chunk index. Slice sequence id. |
| Content | String | No | chunk content. |
Document sharding configuration
Used by actions: ChunkDocument.
| Name | Type | Required | Description |
|---|---|---|---|
| MaxChunkSize | Integer | No | After splitting by the separator, the fragment length is validated. If it exceeds the maximum fragment length, it is split using the next-level separator. If there is no next-level separator, the original length is retained. Default value: 1000. |
| Delimiters | Array of String | No | Separator list, separators positioned towards the front take precedence; when the file type is TXT, default value: ["\n\n", "\n", ".", ".", "?", ",", ""]. |
| ChunkOverlap | Integer | No | The number of overlapping characters between adjacent fragments must be less than the fragment length. Fragments that form completely redundant slices are automatically filtered. Default value: 0.2*MaxChunkSize |
Document slicing async task
Used by actions: ChunkDocumentAsync.
| Name | Type | Required | Description |
|---|---|---|---|
| MaxChunkSize | Integer | No | Maximum fragment length |
Document slice information
Used by actions: ChunkDocument.
| Name | Type | Required | Description |
|---|---|---|---|
| FileType | String | No | File type, supports MD and TXT formats. |
| FileContent | String | No | Text original, use string format input. |
Document Information
Used by actions: ChunkDocumentAsync, ParseDocumentAsync.
| Name | Type | Required | Description |
|---|---|---|---|
| FileType | String | Yes | Supported file types: PDF, DOC, DOCX, PPT, PPTX, MD, TXT, XLS. XLSX,CSV,PNG,JPG,JPEG,BMP,GIF,WEBP,HEIC,EPS,ICNS, IM,PCX,PPM,TIFF,XBM,HEIF,JP2 Supported file size for document parsing: -PDF, DOC, DOCX, PPT, PPTX support 100M. -MD, TXT, XLS, XLSX, CSV support 10M. -Other supports 20M Supported file size for text slicing: -Maximum PDF size: 300 MB -D0CX, D0C, PPT, PPTX maximum 200M -Maximum 10M for TXT, MD -Another maximum of 20M |
| FileUrl | String | No | File storage on Tencent Cloud's URL ensures higher download speed and stability by using Tencent Cloud COS file address. |
| FileContent | String | No | base64 value of the file, carrying the MineType prefix information. The encoded file is no more than 10M. Supported file size: The downloaded file is no more than 8M after Base64 encoding. File download time is not more than 3 seconds. Supported image pixel: Unilateral between 20-10000px. |
| FileName | String | No | File name, used when uploading with base64. |
| FileStartPageNumber | Integer | No | Starting page number of the document |
| FileEndPageNumber | Integer | No | End page number of the document |
Document slice amount
Used by actions: GetDocumentChunkResult.
| Name | Type | Required | Description |
|---|---|---|---|
| PageNumber | Integer | No | Parse page count |
| TotalTokens | Integer | No | token consumption |
Document parsing configuration
Used by actions: ParseDocument.
| Name | Type | Required | Description |
|---|---|---|---|
| ImageResponseType | Integer | No | 0: Images are returned in link format 1:Return the text content extracted from the image |
Vector content
Used by actions: GetTextEmbedding.
| Name | Type | Description |
|---|---|---|
| Embedding | Array of Float | embedding content Note: This field may return null, indicating that no valid values can be obtained. |
| Index | Integer | Index serial number Note: This field may return null, indicating that no valid values can be obtained. |
Page number consumption
Used by actions: GetDocumentParseResult, ParseDocument.
| Name | Type | Required | Description |
|---|---|---|---|
| TotalPages | Integer | No | Total number of pages consumed |
Document Information
Used by actions: ParseDocument.
| Name | Type | Required | Description |
|---|---|---|---|
| FileType | String | Yes | Supported file types: PDF, DOC, DOCX, PPT, PPTX, MD, TXT, XLS. XLSX,CSV,PNG,JPG,JPEG,BMP,GIF,WEBP,HEIC,EPS,ICNS, IM,PCX,PPM,TIFF,XBM,HEIF,JP2 Supported file size for document parsing: -PDF, DOC, DOCX, PPT, PPTX support 100M. -MD, TXT, XLS, XLSX, CSV support 10M. -Another supports 20M Supported file size for text slicing: -Maximum PDF size: 300 MB -D0CX, D0C, PPT, PPTX maximum 200M -Maximum 10M for TXT, MD -Other maximum 20M |
| FileUrl | String | No | File storage on Tencent Cloud's URL ensures higher download speed and stability by using Tencent Cloud COS file address. |
| FileContent | String | No | base64 value of the file, carrying the MineType prefix information. The encoded file is no more than 10M. Supported file size: The downloaded file is no more than 8M after Base64 encoding. File download time is no more than 3 seconds. Supported image pixel: Unilateral between 20-10000px. Either FileUrl or FileContent of the file must be provided. If both are provided, only FileUrl is used. |
| DocumentParseConfig | DocumentParseConfig | No | Document parsing configuration |
| FileStartPageNumber | Integer | No | Starting page number of the document |
| FileEndPageNumber | Integer | No | End page number of the document |
Output result
Used by actions: RunRerank.
| Name | Type | Description |
|---|---|---|
| Index | Integer | Location index value of the corresponding doc in the array of candidate docs Note: This field may return null, indicating that no valid values can be obtained. |
| RelevanceScore | Float | Similarity score Note: This field may return null, indicating that no valid values can be obtained. |
| Document | String | Document content Note: This field may return null, indicating that no valid values can be obtained. |
token consumption total count
Used by actions: ChunkDocument, GetTextEmbedding, RunRerank.
| Name | Type | Description |
|---|---|---|
| TotalTokens | Integer | Total number of tokens Note: This field may return null, indicating that no valid values can be obtained. |
Was this page helpful?
You can also Contact sales or Submit a Ticket for help.
Help us improve! Rate your documentation experience in 5 mins.
Feedback