Apache Tika 1.14 发布了，该版本包含了一些改进和 Bug 修复。Tika 是一个内容抽取的工具集合(a toolkit for text extracting)。它集成了 POI, Pdfbox 并且为文本抽取工作提供了一个统一的界面。其次，Tika 也提供了便利的扩展 API，用来丰富其对第三方文件格式的支持。

更新如下：

• Extract all headers from MSG/RFC822 (TIKA-2122).

• Upgrade metadata-extractor to 2.9.1 (TIKA-2113).

• Extract PDF DocInfo metadata into separate keys to prevent overwriting by XMP metadata (TIKA-2057).

• Re-enable fileUrl for tika-server (TIKA-2081).  If you choose,to use this feature, beware of the security vulnerabilities!See: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2015-3271

• Add Tesseract's hOCR output format as an option, via Eric Pugh(TIKA-2093)

• Extract macros from MSOffice files (TIKA-2069).

• Maintain passed-in mime in TXTParser (TIKA-2047).

• Upgrade to POI.3-15 (TIKA-2013).

• Upgrade to PDFBox 2.0.3 (TIKA-2051).

• Fix hyperlinks with formatting in DOC and DOCX (TIKA-1255 and TIKA-2078)

• Tika now is integrated with the Tensorflow library from Google and it can use its Inception v3 image classification model to identify objects in images (TIKA-1993).

• Parser configuration is now type-safe and parameters for parsers can have assigned types (TIKA-1508, TIKA-1986).

• Prevent OOM/permanent hang on some corrupt CHM files (TIKA-2040).

• Upgrade ICU4J charset detection components to fix multithreading bug (TIKA-2041).

• Upgrade to Jackcess 2.1.4 (TIKA-2039).

• Maintain more significant digits in cells of "General" format in XLS and XLSX (TIKA-2025).

• Avoid mark/reset issues when extracting or detecting embedded resources in RFC822 emails (TIKA-2037).

• Improving accuracy of Tesseract for better extraction of numeric and alphanumeric text from images (TIKA-2021, TIKA-2031).

• Improve extraction of embedded documents from PPT, PPTX and XLSX(TIKA-2026).

• Add parser for applefile (AppleSingle) (TIKA-2022).

• Add mime types, mime magic and/or globs for:

  • Endnote Import File (TIKA-2011)

  • DJVU files (TIKA-2009)

  • MS Owner File (TIKA-2008)

  • Windows Media Metafile (TIKA-2004)

  • iCal and vCalendar (TIKA-2006)

  • MBOX (TIKA-2042)

  • Stata DTA (TIKA-2064)

• Add configurable maximum threshold for number of events extracted from the XMP Media Management Schema in JempboxExtractor (TIKA-1999).

• Integrate TesseractOCR with full page image rendering for PDFs (TIKA-1994).

• Add mime detection via Nick C and parser for DBF files (TIKA-1513).

• Add mime detection and parsers for MSOffice 2003 XML Word and Excel formats (TIKA-1958).

• Extract hyperlinks from PPT, PPTX, XSLX (TIKA-1454).

• Upgrade to Commons Compress 1.12 (supports progress on TIKA-1358)

发布说明和完整更新内容

下载地址：

• http://www.apache.org/dyn/closer.cgi/tika/apache-tika-1.14-src.zip