claunia/SabreTools
1
0
Fork 0
mirror of https://github.com/claunia/SabreTools.git synced 2025-12-16 19:14:27 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add archive.org deserialization test, fix issues

Browse Source
This commit is contained in:
Matt Nadareski
2023-07-12 17:21:19 -04:00
parent 225be86248
commit a351a02ee1
6 changed files with 336 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

60
SabreTools.Models/ArchiveDotOrg/File.cs
View File

@@ -49,18 +49,56 @@ namespace SabreTools.Models.ArchiveDotOrg
[XmlElement("rotation")]
public long? Rotation { get; set; }
[XmlElement("hocr_char_to_word_module_version")]
public string? hOCRCharToWordModuleVersion { get; set; }
#region OCR-Related
[XmlElement("cloth_cover_detection_module_version")]
public string? ClothCoverDetectionModuleVersion { get; set; }
[XmlElement("hocr_char_to_word_hocr_version")]
public string? hOCRCharToWordhOCRVersion { get; set; }
[XmlElement("ocr_module_version")]
public string? TesseractOCRModuleVersion { get; set; }
[XmlElement("hocr_char_to_word_module_version")]
public string? hOCRCharToWordModuleVersion { get; set; }
[XmlElement("hocr_fts_text_hocr_version")]
public string? hOCRFtsTexthOCRVersion { get; set; }
[XmlElement("hocr_fts_text_module_version")]
public string? hOCRFtsTextModuleVersion { get; set; }
[XmlElement("hocr_pageindex_hocr_version")]
public string? hOCRPageIndexhOCRVersion { get; set; }
[XmlElement("hocr_pageindex_module_version")]
public string? hOCRPageIndexModuleVersion { get; set; }
[XmlElement("ocr")]
public string? TesseractOCR { get; set; }
[XmlElement("ocr_converted")]
public string? TesseractOCRConverted { get; set; }
[XmlElement("ocr_detected_lang")]
public string? TesseractOCRDetectedLang { get; set; }
[XmlElement("ocr_detected_lang_conf")]
public string? TesseractOCRDetectedLangConf { get; set; }
[XmlElement("ocr_detected_script")]
public string? TesseractOCRDetectedScript { get; set; }
[XmlElement("ocr_detected_script_conf")]
public string? TesseractOCRDetectedScriptConf { get; set; }
[XmlElement("ocr_parameters")]
public string? TesseractOCRParameters { get; set; }
[XmlElement("ocr_module_version")]
public string? TesseractOCRModuleVersion { get; set; }
[XmlElement("pdf_module_version")]
public string? PDFModuleVersion { get; set; }
[XmlElement("word_conf_0_10")]
public long? WordConfidenceInterval0To10 { get; set; }
@@ -90,5 +128,19 @@ namespace SabreTools.Models.ArchiveDotOrg
[XmlElement("word_conf_91_100")]
public long? WordConfidenceInterval91To100 { get; set; }
#endregion
#region DO NOT USE IN PRODUCTION
/// <remarks>Should be empty</remarks>
[XmlAnyAttribute]
public XmlAttribute[]? ADDITIONAL_ATTRIBUTES { get; set; }
/// <remarks>Should be empty</remarks>
[XmlAnyElement]
public object[]? ADDITIONAL_ELEMENTS { get; set; }
#endregion
}
}