Lay some of the groundwork for model/usermodel support for rich text

(note - RichTextRun is not fixed, and method signatures might well change)


git-svn-id: https://svn.apache.org/repos/asf/jakarta/poi/trunk@353788 13f79535-47bb-0310-9956-ffa450edef68

src/scratchpad/src/org/apache/poi/hslf/model/TextRun.java

@@ -19,7 +19,11 @@
 
 package org.apache.poi.hslf.model;
 
+import java.util.LinkedList;
+
 import org.apache.poi.hslf.record.*;
+import org.apache.poi.hslf.record.StyleTextPropAtom.TextPropCollection;
+import org.apache.poi.hslf.usermodel.RichTextRun;
 import org.apache.poi.util.StringUtil;
 
 /**
@@ -37,18 +41,17 @@ public class TextRun
 	private TextCharsAtom  _charAtom;
 	private StyleTextPropAtom _styleAtom;
 	private boolean _isUnicode;
+	private RichTextRun[] _rtRuns;
 
 	/**
 	* Constructs a Text Run from a Unicode text block
 	*
 	* @param tha the TextHeaderAtom that defines what's what
 	* @param tca the TextCharsAtom containing the text
+	* @param sta the StyleTextPropAtom which defines the character stylings
 	*/
 	public TextRun(TextHeaderAtom tha, TextCharsAtom tca, StyleTextPropAtom sta) {
-		_headerAtom = tha;
+		this(tha,null,tca,sta);
-		_charAtom = tca;
-		_styleAtom = sta;
-		_isUnicode = true;
 	}
 
 	/**
@@ -56,12 +59,136 @@ public class TextRun
 	*
 	* @param tha the TextHeaderAtom that defines what's what
 	* @param tba the TextBytesAtom containing the text
+	* @param sta the StyleTextPropAtom which defines the character stylings
 	*/
 	public TextRun(TextHeaderAtom tha, TextBytesAtom tba, StyleTextPropAtom sta) {
+		this(tha,tba,null,sta);
+	}
+	
+	/**
+	 * Internal constructor and initializer
+	 */
+	private TextRun(TextHeaderAtom tha, TextBytesAtom tba, TextCharsAtom tca, StyleTextPropAtom sta) {
 		_headerAtom = tha;
-		_byteAtom = tba;
 		_styleAtom = sta;
-		_isUnicode = false;
+		if(tba != null) {
+			_byteAtom = tba;
+			_isUnicode = false;
+		} else {
+			_charAtom = tca;
+			_isUnicode = true;
+		}
+		
+		// Figure out the rich text runs
+		// TODO: Handle when paragraph style and character styles don't match up
+		LinkedList pStyles = new LinkedList();
+		LinkedList cStyles = new LinkedList();
+		if(_styleAtom != null) {
+			pStyles = _styleAtom.getParagraphStyles();
+			cStyles = _styleAtom.getCharacterStyles();
+		}
+		if(pStyles.size() != cStyles.size()) {
+			throw new RuntimeException("Don't currently handle case of overlapping styles");
+		}
+		_rtRuns = new RichTextRun[pStyles.size()];
+		//for(int i=0; i<)
+	}
+	
+	
+	// Update methods follow
+
+	/**
+	 * Saves the given string to the records. Doesn't touch the stylings. 
+	 */
+	private void storeText(String s) {
+		if(_isUnicode) {
+			// The atom can safely convert to unicode
+			_charAtom.setText(s);
+		} else {
+			// Will it fit in a 8 bit atom?
+			boolean hasMultibyte = StringUtil.hasMultibyte(s);
+			if(! hasMultibyte) {
+				// Fine to go into 8 bit atom
+				byte[] text = new byte[s.length()];
+				StringUtil.putCompressedUnicode(s,text,0);
+				_byteAtom.setText(text);
+			} else {
+				throw new RuntimeException("Setting of unicode text is currently only possible for Text Runs that are Unicode in the file, sorry. For now, please convert that text to us-ascii and re-try it");
+			}
+		}
+	}
+	
+	/**
+	 * Handles an update to the text stored in one of the Rich Text Runs
+	 * @param run
+	 * @param s
+	 */
+	public synchronized void changeTextInRichTextRun(RichTextRun run, String s) {
+		// Figure out which run it is
+		int runID = -1;
+		for(int i=0; i<_rtRuns.length; i++) {
+			if(run.equals(_rtRuns[i])) {
+				runID = i;
+			}
+		}
+		if(runID == -1) {
+			throw new IllegalArgumentException("Supplied RichTextRun wasn't from this TextRun");
+		}
+		
+		// Update the text length for its Paragraph and Character stylings
+		LinkedList pStyles = _styleAtom.getParagraphStyles();
+		LinkedList cStyles = _styleAtom.getCharacterStyles();
+		TextPropCollection pCol = (TextPropCollection)pStyles.get(runID);
+		TextPropCollection cCol = (TextPropCollection)cStyles.get(runID);
+		pCol.updateTextSize(s.length());
+		cCol.updateTextSize(s.length());
+		
+		// Build up the new text
+		// As we go through, update the start position for all subsequent runs
+		// The building relies on the old text still being present
+		StringBuffer newText = new StringBuffer();
+		for(int i=0; i<_rtRuns.length; i++) {
+			// Update start position
+			if(i > runID) {
+				_rtRuns[i].updateStartPosition(newText.length());
+			}
+			// Grab new text
+			if(i != runID) {
+				newText.append(_rtRuns[i].getRawText());
+			} else {
+				newText.append(s);
+			}
+		}
+		
+		// Save the new text
+		storeText(newText.toString());
+	}
+
+	/**
+	 * Changes the text, and sets it all to have the same styling
+	 *  as the the first character has. 
+	 * If you care about styling, do setText on a RichTextRun instead 
+	 */
+	public synchronized void setText(String s) {
+		// Save the new text to the atoms
+		storeText(s);
+
+		// Now handle record stylings:
+		//  everthing gets the same style that the first block has
+		LinkedList pStyles = _styleAtom.getParagraphStyles();
+		while(pStyles.size() > 1) { pStyles.removeLast(); }
+		
+		LinkedList cStyles = _styleAtom.getCharacterStyles();
+		while(cStyles.size() > 1) { cStyles.removeLast(); }
+		
+		TextPropCollection pCol = (TextPropCollection)pStyles.getFirst();
+		TextPropCollection cCol = (TextPropCollection)cStyles.getFirst();
+		pCol.updateTextSize(s.length());
+		cCol.updateTextSize(s.length());
+		
+		// Finally, zap and re-do the RichTextRuns
+		_rtRuns = new RichTextRun[1];
+		_rtRuns[0] = new RichTextRun(this,0,s.length());
 	}
 
 
@@ -93,36 +220,16 @@ public class TextRun
 			return _byteAtom.getText();
 		}
 	}
-
+	
 	/**
-	 * Changes the text. Chance are, this won't work just yet, because
+	 * Fetch the rich text runs (runs of text with the same styling) that
-	 *  we also need to update some other bits of the powerpoint file
+	 *  are contained within this block of text
-	 *  to match the change in the Text Atom, especially byte offsets
+	 * @return
 	 */
-	public void setText(String s) {
+	public RichTextRun[] getRichTextRuns() {
-		// If size changed, warn
+		return 	_rtRuns;
-		if(s.length() != getText().length()) {
-			System.err.println("Warning: Your powerpoint file may no longer readable by powerpoint, as the text run has changed size!");
-		}
-
-		if(_isUnicode) {
-			// The atom can safely convert to unicode
-			_charAtom.setText(s);
-		} else {
-			// Will it fit in a 8 bit atom?
-			boolean hasMultibyte = StringUtil.hasMultibyte(s);
-			if(! hasMultibyte) {
-				// Fine to go into 8 bit atom
-				byte[] text = new byte[s.length()];
-				StringUtil.putCompressedUnicode(s,text,0);
-				_byteAtom.setText(text);
-			} else {
-				throw new RuntimeException("Setting of unicode text is currently only possible for Text Runs that are Unicode in the file, sorry. For now, please convert that text to us-ascii and re-try it");
-			}
-		}
-		
 	}
-
+	
 	/**
 	* Returns the type of the text, from the TextHeaderAtom.
 	* Possible values can be seen from TextHeaderAtom

src/scratchpad/src/org/apache/poi/hslf/record/ParentAwareRecord.java

@@ -0,0 +1,12 @@
+package org.apache.poi.hslf.record;
+
+/**
+ * Interface to define how a record can indicate it cares about what its
+ *  parent is, and how it wants to be told which record is its parent.
+ * 
+ * @author Nick Burch (nick at torchbox dot com)
+ */
+public interface ParentAwareRecord {
+	public RecordContainer getParentRecord();
+	public void setParentRecord(RecordContainer parentRecord);
+}

src/scratchpad/src/org/apache/poi/hslf/record/StyleTextPropAtom.java

@@ -64,6 +64,11 @@ public class StyleTextPropAtom extends RecordAtom
 	 */
 	private LinkedList paragraphStyles;
 	public LinkedList getParagraphStyles() { return paragraphStyles; }
+	/**
+	 * Updates the link list of TextPropCollections which make up the
+	 *  paragraph stylings
+	 */
+	public void setParagraphStyles(LinkedList ps) { paragraphStyles = ps; }
 	/** 
 	 * The list of all the different character stylings we code for.
 	 * Each entry is a TextPropCollection, which tells you how many
@@ -72,6 +77,11 @@ public class StyleTextPropAtom extends RecordAtom
 	 */
 	private LinkedList charStyles;
 	public LinkedList getCharacterStyles() { return charStyles; }
+	/**
+	 * Updates the link list of TextPropCollections which make up the
+	 *  character stylings
+	 */
+	public void setCharacterStyles(LinkedList cs) { charStyles = cs; }
 
 	/** All the different kinds of paragraph properties we might handle */
 	public TextProp[] paragraphTextPropTypes = new TextProp[] {
@@ -370,6 +380,14 @@ public class StyleTextPropAtom extends RecordAtom
 			reservedField = -1;
 			textPropList = new LinkedList();
 		}
+		
+		/**
+		 * Update the size of the text that this set of properties
+		 *  applies to 
+		 */
+		public void updateTextSize(int textSize) {
+			charactersCovered = textSize;
+		}
 
 		/**
 		 * Writes out to disk the header, and then all the properties

src/scratchpad/src/org/apache/poi/hslf/record/TextHeaderAtom.java

@@ -30,10 +30,11 @@ import java.io.OutputStream;
  * @author Nick Burch
  */
 
-public class TextHeaderAtom extends RecordAtom
+public class TextHeaderAtom extends RecordAtom implements ParentAwareRecord
 {
 	private byte[] _header;
 	private static long _type = 3999l;
+	private RecordContainer parentRecord;
 
 	public static final int TITLE_TYPE = 0;
 	public static final int BODY_TYPE = 1;
@@ -49,6 +50,9 @@ public class TextHeaderAtom extends RecordAtom
 
 	public int getTextType() { return textType; }
 	public void setTextType(int type) { textType = type; }
+	
+	public RecordContainer getParentRecord() { return parentRecord; }
+	public void setParentRecord(RecordContainer record) { this.parentRecord = record; } 
 
 	/* *************** record code follows ********************** */
 

src/scratchpad/src/org/apache/poi/hslf/usermodel/RichTextRun.java

@@ -0,0 +1,85 @@
+
+/* ====================================================================
+   Copyright 2002-2004   Apache Software Foundation
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.
+==================================================================== */
+        
+
+
+package org.apache.poi.hslf.usermodel;
+
+import org.apache.poi.hslf.model.TextRun;
+
+/**
+ * Represents a run of text, all with the same style
+ * 
+ * TODO: get access to the font/character properties
+ *
+ * @author Nick Burch
+ */
+
+public class RichTextRun
+{
+	/** The TextRun we belong to */
+	private TextRun parentRun;
+	
+	/** Where in the parent TextRun we start from */
+	private int startPos;
+	
+	/** How long a string (in the parent TextRun) we represent */
+	private int length;
+	
+	/** Our paragraph and character style */
+	
+	/**
+	 * Create a new wrapper around a rich text string
+	 * @param parent
+	 * @param startAt
+	 * @param len
+	 */
+	public RichTextRun(TextRun parent, int startAt, int len) {
+		parentRun = parent;
+		startPos = startAt;
+		length = len;
+	}
+	
+	/**
+	 * Fetch the text, in output suitable form
+	 */
+	public String getText() {
+		return parentRun.getText().substring(startPos, startPos+length);
+	}
+	/**
+	 * Fetch the text, in raw storage form
+	 */
+	public String getRawText() {
+		return parentRun.getRawText().substring(startPos, startPos+length);
+	}
+	
+	/**
+	 * Change the text
+	 */
+	public void setText(String text) {
+		length = text.length();
+		parentRun.changeTextInRichTextRun(this,text);
+	}
+	
+	/**
+	 * Tells the RichTextRun its new position in the parent TextRun
+	 * @param startAt
+	 */
+	public void updateStartPosition(int startAt) {
+		startPos = startAt;
+	}
+}

src/scratchpad/src/org/apache/poi/hslf/usermodel/SlideShow.java

@@ -27,7 +27,9 @@ import org.apache.poi.util.LittleEndian;
 import org.apache.poi.hslf.*;
 import org.apache.poi.hslf.model.*;
 import org.apache.poi.hslf.record.FontCollection;
+import org.apache.poi.hslf.record.ParentAwareRecord;
 import org.apache.poi.hslf.record.Record;
+import org.apache.poi.hslf.record.RecordContainer;
 import org.apache.poi.hslf.record.RecordTypes;
 import org.apache.poi.hslf.record.SlideAtom;
 import org.apache.poi.hslf.record.SlideListWithText;
@@ -78,6 +80,11 @@ public class SlideShow
     _hslfSlideShow = hslfSlideShow;
 	_records = _hslfSlideShow.getRecords();
 	byte[] _docstream = _hslfSlideShow.getUnderlyingBytes();
+	
+	// Handle Parent-aware Reocrds
+	for(int i=0; i<_records.length; i++) {
+		handleParentAwareRecords(_records[i]);
+	}
 
 	// Find the versions of the core records we'll want to use
 	findMostRecentCoreRecords();
@@ -85,6 +92,32 @@ public class SlideShow
 	// Build up the model level Slides and Notes
 	buildSlidesAndNotes();
   }
+  
+  
+  /**
+   * Find the records that are parent-aware, and tell them
+   *  who their parent is
+   */
+  private void handleParentAwareRecords(Record baseRecord) {
+	  // Only need to do something if this is a container record
+	  if(baseRecord instanceof RecordContainer) {
+		RecordContainer br = (RecordContainer)baseRecord;
+		Record[] childRecords = br.getChildRecords();
+		
+		// Loop over child records, looking for interesting ones
+		for(int i=0; i<childRecords.length; i++) {
+			Record record = childRecords[i];
+			// Tell parent aware records of their parent
+			if(record instanceof ParentAwareRecord) {
+				((ParentAwareRecord)record).setParentRecord(br);
+			}
+			// Walk on down for the case of container records
+			if(record instanceof RecordContainer) {
+				handleParentAwareRecords(record);
+			}
+		}
+	  }
+  }
 
 
   /**