Example usage for org.dom4j.io SAXReader setXMLReaderClassName

Introduction

In this page you can find the example usage for org.dom4j.io SAXReader setXMLReaderClassName.

Prototype

public void setXMLReaderClassName(String xmlReaderClassName) throws SAXException 

Source Link

Document

Sets the class name of the XMLReader to be used to parse SAX events.

Usage

From source file:com.globalsight.everest.tm.util.ttx.TtxToTmx.java

License:Apache License

/**
 * Main method to call, returns the new filename of the result.
 *///from w w w.  j  a va  2  s.  com
public String convertTtxToTmx(String p_url) throws Exception {
    final String baseName = getBaseName(p_url);
    final String extension = getExtension(baseName);

    info("Converting TTX file to TMX: `" + p_url + "'");

    m_entryCount = 0;

    // Reading from a file, need to use Xerces.
    SAXReader reader = new SAXReader();
    reader.setXMLReaderClassName("org.apache.xerces.parsers.SAXParser");
    //reader.setEntityResolver(DtdResolver.getInstance());
    //reader.setValidation(true);

    // Fetch the version info early.
    reader.addHandler("/TRADOStag", new ElementHandler() {
        public void onStart(ElementPath path) {
            Element element = path.getCurrent();

            m_version = element.attributeValue(Ttx.VERSION);
        }

        public void onEnd(ElementPath path) {
        }
    });

    // Fetch the header info early.
    reader.addHandler("/TRADOStag/FrontMatter", new ElementHandler() {
        public void onStart(ElementPath path) {
        }

        public void onEnd(ElementPath path) {
            Element element = path.getCurrent();

            setTtxHeader(element);

            try {
                startOutputFile(baseName);
            } catch (Exception ex) {
                error(ex.toString());
                System.exit(1);
            }

            // prune the current element to reduce memory
            element.detach();
            element = null;
        }
    });

    ElementHandler tuHandler = new ElementHandler() {
        public void onStart(ElementPath path) {
            ++m_entryCount;
        }

        public void onEnd(ElementPath path) {
            Element element = path.getCurrent();

            element = cleanupTu(element);

            writeEntry(element.asXML());

            // prune the current element to reduce memory
            element.detach();
            element = null;

            if (m_entryCount % 50 == 0) {
                debug("Entry " + m_entryCount);
            }
        }
    };

    // Path handlers cannot use "//", sooo specify all known paths.
    reader.addHandler("/TRADOStag/Body/Raw/Tu", tuHandler);
    reader.addHandler("/TRADOStag/Body/Raw/df/Tu", tuHandler);
    reader.addHandler("/TRADOStag/Body/Raw/ut/Tu", tuHandler);
    reader.addHandler("/TRADOStag/Body/Raw/df/ut/Tu", tuHandler);

    // Read in the entire file (it's not too big normally).
    Document document = reader.read(p_url);

    closeOutputFile();

    info("Processed " + m_entryCount + " TUs into file `" + m_filename + "'");

    return m_filename;
}

---

From source file:com.globalsight.ling.docproc.DiplomatWordCounter.java

License:Apache License

private SAXReader createXmlParser() {
    SAXReader result = new SAXReader();

    try {/*w ww .  j a  v  a 2 s . co  m*/
        result.setXMLReaderClassName("org.dom4j.io.aelfred.SAXDriver");
        result.setValidation(false);
    } catch (Exception ex) {
        System.err.println("org.dom4j.io.aelfred.SAXDriver not found");

        // Use the system default parser, better than nothing.
        result = new SAXReader();
        result.setValidation(false);
    }

    return result;
}

---

From source file:com.globalsight.terminology.importer.GTXmlReader.java

License:Apache License

/**
 * Reads an XML file and checks for correctness. If there's any
 * error in the file, an exception is thrown.
 *//*  w  w  w  .j ava  2s  .  com*/
private void analyzeXml(String p_url) throws Exception {
    SAXReader reader = new SAXReader();
    reader.setXMLReaderClassName("org.apache.xerces.parsers.SAXParser");

    CATEGORY.debug("Analyzing document: " + p_url);

    // enable element complete notifications to conserve memory
    reader.addHandler("/entries/conceptGrp", new ElementHandler() {
        public void onStart(ElementPath path) {
            ++m_entryCount;
        }

        public void onEnd(ElementPath path) {
            Element element = path.getCurrent();

            // prune the current element to reduce memory
            element.detach();
        }
    });

    ImportUtil.filterateXmlIllegal(p_url, m_options.getEncoding());
    Document document = reader.read(p_url);

    // all done
}

---

From source file:com.globalsight.terminology.importer.GTXmlReaderThread.java

License:Apache License

public void run() {
    try {//from  w  w  w.j a  v  a2s . c o  m
        SAXReader reader = new SAXReader();
        reader.setXMLReaderClassName("org.apache.xerces.parsers.SAXParser");

        // enable pruning to call me back as each Element is complete
        reader.addHandler("/entries/conceptGrp", new ElementHandler() {
            public void onStart(ElementPath path) {
            }

            public void onEnd(ElementPath path) {
                Element element = path.getCurrent();

                // prune the current element to reduce memory
                element.detach();

                Document doc = m_factory.createDocument(element);
                Entry entry = new Entry(doc);

                m_result = m_results.hireResult();
                m_result.setResultObject(entry);

                boolean done = m_results.put(m_result);
                m_result = null;

                // Stop reading the TMX file.
                if (done) {
                    throw new ThreadDeath();
                }
            }
        });

        String url = m_options.getFileName();

        reader.read(url);
    } catch (ThreadDeath ignore) {
        CATEGORY.info("ReaderThread: interrupted.");
    } catch (Throwable ignore) {
        // Should never happen, and I don't know how to handle
        // this case other than passing the exception in
        // m_results, which I won't do for now.
    } finally {
        if (m_result != null) {
            m_results.fireResult(m_result);
            m_result = null;
        }

        m_results.producerDone();
        m_results = null;

        CATEGORY.debug("ReaderThread: done.");
    }
}

---

From source file:com.globalsight.terminology.importer.MtfReader.java

License:Apache License

/**
 * Reads an XML file and checks for correctness. If there's any
 * error in the file, an exception is thrown.
 *//* w  w  w.j a v  a2  s .  com*/
private void analyzeXml(String p_url) throws Exception {
    SAXReader reader = new SAXReader();
    reader.setXMLReaderClassName("org.apache.xerces.parsers.SAXParser");

    CATEGORY.debug("Analyzing document: " + p_url);

    // enable element complete notifications to conserve memory
    reader.addHandler("/mtf/conceptGrp", new ElementHandler() {
        public void onStart(ElementPath path) {
            ++m_entryCount;
        }

        public void onEnd(ElementPath path) {
            Element element = path.getCurrent();

            // prune the current element to reduce memory
            element.detach();

            // TODO: validate entry and report errors.
        }
    });

    Document document = reader.read(p_url);

    // all done
}

---

From source file:com.globalsight.terminology.importer.MtfReaderThread.java

License:Apache License

public void run() {
    try {// www . ja v a  2 s .c  o m
        SAXReader reader = new SAXReader();
        reader.setXMLReaderClassName("org.apache.xerces.parsers.SAXParser");

        // enable pruning to call me back as each Element is complete
        reader.addHandler("/mtf/conceptGrp", new ElementHandler() {
            public void onStart(ElementPath path) {
                m_count++;
            }

            public void onEnd(ElementPath path) {
                Element element = path.getCurrent();

                // prune the current element to reduce memory
                element.detach();

                m_result = m_results.hireResult();

                try {
                    // Convert MultiTerm to GlobalSight.
                    element = convertMtf(element);

                    Document doc = m_factory.createDocument(element);
                    Entry entry = new Entry(doc);

                    if (CATEGORY.isDebugEnabled()) {
                        CATEGORY.debug(entry.getXml());
                    }

                    m_result.setResultObject(entry);
                } catch (Throwable ex) {
                    String msg = "Entry " + m_count + ": " + ex.getMessage();

                    m_result.setError(msg);

                    if (CATEGORY.isDebugEnabled()) {
                        CATEGORY.debug(msg, ex);
                    } else {
                        CATEGORY.warn(msg, ex);
                    }
                }

                boolean done = m_results.put(m_result);
                m_result = null;

                // Stop reading the XML file.
                if (done) {
                    throw new ThreadDeath();
                }
            }
        });

        String url = m_options.getFileName();

        Document document = reader.read(url);
    } catch (ThreadDeath ignore) {
        CATEGORY.info("ReaderThread: interrupted.");
    } catch (Throwable ignore) {
        // Should never happen, and I don't know how to handle
        // this case other than passing the exception in
        // m_results, which I won't do for now.
        CATEGORY.error("unexpected error", ignore);
    } finally {
        if (m_result != null) {
            m_results.fireResult(m_result);
            m_result = null;
        }

        m_results.producerDone();
        m_results = null;

        CATEGORY.debug("ReaderThread: done.");
    }
}

---

From source file:com.globalsight.terminology.importer.TbxReader.java

License:Apache License

private void analyzeTbx(String p_url) throws Exception {
    SAXReader reader = new SAXReader();
    reader.setXMLReaderClassName("org.apache.xerces.parsers.SAXParser");

    CATEGORY.debug("Analyzing document: " + p_url);
    // enable element complete notifications to conserve memory

    reader.addHandler("/martif/text/body/termEntry", new ElementHandler() {
        public void onStart(ElementPath path) {
            ++m_entryCount;//from w  w w. j ava2s.  com
        }

        public void onEnd(ElementPath path) {
            Element element = path.getCurrent();

            // prune the current element to reduce memory
            element.detach();
        }
    });

    Document document = reader.read(p_url);
}

---

From source file:com.globalsight.terminology.importer.TbxReaderThread.java

License:Apache License

public void run() {
    try {/*from   w ww  . j  a v a2 s  .  co  m*/
        SAXReader reader = new SAXReader();
        reader.setXMLReaderClassName("org.apache.xerces.parsers.SAXParser");

        // enable pruning to call me back as each Element is complete
        reader.addHandler("/martif/text/body/termEntry", new ElementHandler() {
            public void onStart(ElementPath path) {
            }

            public void onEnd(ElementPath path) {
                Element element = path.getCurrent();

                // prune the current element to reduce memory
                element.detach();

                Document doc = m_factory.createDocument(element);
                Entry entry = new Entry(doc);

                m_result = m_results.hireResult();
                m_result.setResultObject(entry);

                boolean done = m_results.put(m_result);
                m_result = null;

                // Stop reading the TMX file.
                if (done) {
                    throw new ThreadDeath();
                }
            }
        });

        String url = m_options.getFileName();

        Document document = reader.read(url);
    } catch (ThreadDeath ignore) {
        CATEGORY.info("ReaderThread: interrupted.");
    } catch (Throwable ignore) {
        // Should never happen, and I don't know how to handle
        // this case other than passing the exception in
        // m_results, which I won't do for now.
    } finally {
        if (m_result != null) {
            m_results.fireResult(m_result);
            m_result = null;
        }

        m_results.producerDone();
        m_results = null;

        CATEGORY.debug("ReaderThread: done.");
    }
}

---

From source file:com.globalsight.terminology.util.MtfAnalyzer.java

License:Apache License

public void analyze(String p_url) throws Exception {
    m_entryCount = 0;//from  ww  w. j  a va  2  s .c o m

    SAXReader reader = new SAXReader();
    reader.setXMLReaderClassName("org.apache.xerces.parsers.SAXParser");

    System.err.println("Analyzing document: " + p_url);

    // enable element complete notifications to conserve memory
    reader.addHandler("/mtf/conceptGrp", new ElementHandler() {
        public void onStart(ElementPath path) {
            ++m_entryCount;

            if (m_entryCount % 200 == 0) {
                log("Entry " + m_entryCount);
            }
        }

        public void onEnd(ElementPath path) {
            Element element = path.getCurrent();

            // prune the current element to reduce memory
            element.detach();

            element = null;
        }
    });

    Document document = reader.read(p_url);

    log("Total entries: " + m_entryCount);

    // all done
}

---

From source file:com.globalsight.terminology.util.MtfSplitter.java

License:Apache License

public void split(String p_url, String p_numEntries) throws Exception {
    final int maxEntries = Integer.parseInt(p_numEntries);
    final String baseName = getBaseName(p_url);
    final String extension = getExtension(p_url);

    m_entryCount = 0;/*from  w w  w  .  j  a v a2 s . c o m*/

    SAXReader reader = new SAXReader();
    reader.setXMLReaderClassName("org.apache.xerces.parsers.SAXParser");

    log("Splitting document `" + p_url + "'");

    startFile(baseName, extension);

    // enable element complete notifications to conserve memory
    reader.addHandler("/mtf/conceptGrp", new ElementHandler() {
        public void onStart(ElementPath path) {
            ++m_entryCount;

            if (m_entryCount % maxEntries == 0) {
                try {
                    closeFile();
                    startFile(baseName, extension);
                } catch (Exception ex) {
                    log(ex.toString());
                    System.exit(1);
                }
            }
        }

        public void onEnd(ElementPath path) {
            Element element = path.getCurrent();

            writeEntry(element.asXML());

            // prune the current element to reduce memory
            element.detach();

            element = null;
        }
    });

    Document document = reader.read(p_url);

    closeFile();

    // all done
}

---

• «
• 1
• 2
• 3
• »