From 206587dfe2d1067d9e85c8de655331826500d6c7 Mon Sep 17 00:00:00 2001 From: pwalkow <pawel.walkowiak@hotmail.com> Date: Tue, 30 May 2023 11:52:37 +0200 Subject: [PATCH] Mv code --- .../java/pl/clarin/any2txt/Converter.java | 19 +++++++++++-------- 1 file changed, 11 insertions(+), 8 deletions(-) diff --git a/module/src/main/java/pl/clarin/any2txt/Converter.java b/module/src/main/java/pl/clarin/any2txt/Converter.java index 5b0e3f3..ae91d30 100644 --- a/module/src/main/java/pl/clarin/any2txt/Converter.java +++ b/module/src/main/java/pl/clarin/any2txt/Converter.java @@ -76,19 +76,22 @@ public class Converter extends Worker { File file = new File(fileIn); TikaInputStream inS; - try { - DocumentBuilder builder = DocumentBuilderFactory.newInstance().newDocumentBuilder(); - Document doc = builder.parse(file); - } - catch (IOException | SAXException | ParserConfigurationException e) { - Logger.getLogger(Converter.class.getName()).log(Level.WARNING, "Problems with XML parsing " + e.getMessage() , e.getMessage()); - } +// try { +// DocumentBuilder builder = DocumentBuilderFactory.newInstance().newDocumentBuilder(); +// Document doc = builder.parse(file); +// } +// catch (IOException | SAXException | ParserConfigurationException e) { +// Logger.getLogger(Converter.class.getName()).log(Level.WARNING, "Problems with XML parsing " + e.getMessage() , e.getMessage()); +// } try { Metadata metadata = new Metadata(); metadata.set("Content-Encoding", "CP1250"); inS = TikaInputStream.get(file, metadata); + DocumentBuilder builder = DocumentBuilderFactory.newInstance().newDocumentBuilder(); + Document doc = builder.parse(file); + Tika tika = new Tika(); Logger.getLogger(Converter.class.getName()).log(Level.INFO, "Detected file type: " + tika.detect(inS)); // || "text/plain".equals(tika.detect(inS)) @@ -119,7 +122,7 @@ public class Converter extends Worker { } catch (FileNotFoundException ex) { Logger.getLogger(Converter.class.getName()).log(Level.WARNING, "File not found", ex.getMessage()); saveEmptyFile(fileOut); - } catch (IOException | TikaException ex) { + } catch (IOException | TikaException | SAXException | ParserConfigurationException ex) { Logger.getLogger(Converter.class.getName()).log(Level.WARNING, "Problems in Tika processing " + ex.getMessage() , ex.getMessage()); if (ex.getMessage().contains("XML parse error")) { // try { -- GitLab