Support TempFileCommentsTable (monitorjbl#65)

src/main/java/com/github/pjfanning/xlsx/StreamingReader.java

@@ -40,6 +40,8 @@ public static class Builder {
     private boolean avoidTempFiles = false;
     private boolean useSstTempFile = false;
     private boolean encryptSstTempFile = false;
+    private boolean useCommentsTempFile = false;
+    private boolean encryptCommentsTempFile = false;
     private boolean adjustLegacyComments = false;
     private boolean readComments = false;
     private boolean readCoreProperties = false;
@@ -101,6 +103,22 @@ public boolean readComments() {
       return readComments;
     }
 
+    /**
+     * @return Whether to use a temp file for the Comments data. If false, no
+     * temp file will be used and the entire table will be loaded into memory.
+     */
+    public boolean useCommentsTempFile() {
+      return useCommentsTempFile;
+    }
+
+    /**
+     * @return Whether to encrypt the temp file for the Comments data. Only applies if <code>useCommentsTempFile()</code>
+     * is true.
+     */
+    public boolean encryptCommentsTempFile() {
+      return encryptCommentsTempFile;
+    }
+
     /**
      * @return Whether to read the core document properties.
      */
@@ -159,7 +177,7 @@ public Builder bufferSize(int bufferSize) {
     }
 
     /**
-     * For password protected files specify password to open file.
+     * For password protected files, specify password to open file.
      * If the password is incorrect a {@code ReadException} is thrown on
      * {@code read}.
      * <p>NULL indicates that no password should be used, this is the
@@ -189,7 +207,7 @@ public Builder convertFromOoXmlStrict(boolean convertFromOoXmlStrict) {
 
     /**
      * Enables a mode where the code tries to avoid creating temp files. This is independent of
-     * {@code #setUseSstTempFile}.
+     * {@code #setUseSstTempFile} and {@code #setUseCommentsTempFile}.
      * <p>
      * By default, temp files are used to avoid holding onto too much data in memory.
      *
@@ -203,7 +221,7 @@ public Builder setAvoidTempFiles(boolean avoidTempFiles) {
 
     /**
      * Enables use of Shared Strings Table temp file. This option exists to accommodate
-     * extremely large workbooks with millions of unique strings. Normally the SST is entirely
+     * extremely large workbooks with millions of unique strings. Normally, the SST is entirely
      * loaded into memory, but with large workbooks with high cardinality (i.e., very few
      * duplicate values) the SST may not fit entirely into memory.
      * <p>
@@ -234,6 +252,42 @@ public Builder setEncryptSstTempFile(boolean encryptSstTempFile) {
       return this;
     }
 
+    /**
+     * Enables use of Comments temp file. This option exists to accommodate
+     * workbooks with lots of comments. Normally, the Comments are all
+     * loaded into memory.
+     * <p>
+     * By default, all the Comments data *will* be loaded into memory. <strong>However</strong>,
+     * enabling this option at all will have some noticeable performance degradation as you are
+     * trading memory for disk space.
+     *
+     * @param useCommentsTempFile whether to use a temp file to store the Comments data
+     * @return reference to current {@code Builder}
+     * @see #setReadComments(boolean)
+     * @see #setEncryptCommentsTempFile(boolean)
+     */
+    public Builder setUseCommentsTempFile(boolean useCommentsTempFile) {
+      this.useCommentsTempFile = useCommentsTempFile;
+      return this;
+    }
+
+    /**
+     * Enables use of encryption in the Comments temp file. This only applies if <code>setUseCommentsTempFile</code>
+     * is set to true.
+     * <p>
+     * By default, the temp file is not encrypted. <strong>However</strong>,
+     * enabling this option could slow down the processing of Comments data.
+     *
+     * @param encryptCommentsTempFile whether to encrypt the temp file used to store the Comments data
+     * @return reference to current {@code Builder}
+     * @see #setReadComments(boolean)
+     * @see #setUseCommentsTempFile(boolean)
+     */
+    public Builder setEncryptCommentsTempFile(boolean encryptCommentsTempFile) {
+      this.encryptCommentsTempFile = encryptCommentsTempFile;
+      return this;
+    }
+
     /**
      * Enables the reading of the comments.
      *

src/main/java/com/github/pjfanning/xlsx/impl/StreamingWorkbookReader.java

@@ -217,7 +217,7 @@ void loadSheets(OoxmlReader reader, SharedStringsTable sst, StylesTable stylesTa
       PackagePart part = iter.getSheetPart();
       sheetStreams.put(part, is);
       if (builder.readComments()) {
-        sheetComments.put(part, iter.getSheetComments());
+        sheetComments.put(part, iter.getSheetComments(builder));
       }
     }
 

src/main/java/com/github/pjfanning/xlsx/impl/ooxml/OoxmlReader.java

@@ -31,9 +31,10 @@ Licensed to the Apache Software Foundation (ASF) under one or more
 
 import javax.xml.parsers.ParserConfigurationException;
 
+import com.github.pjfanning.poi.xssf.streaming.TempFileCommentsTable;
+import com.github.pjfanning.xlsx.StreamingReader;
 import org.apache.poi.ooxml.POIXMLException;
 import org.apache.poi.openxml4j.exceptions.InvalidFormatException;
-import org.apache.poi.openxml4j.exceptions.OpenXML4JException;
 import org.apache.poi.openxml4j.opc.OPCPackage;
 import org.apache.poi.openxml4j.opc.PackagePart;
 import org.apache.poi.openxml4j.opc.PackagePartName;
@@ -84,7 +85,7 @@ public class OoxmlReader {
   /**
    * Creates a new XSSFReader, for the given package
    */
-  public OoxmlReader(OPCPackage pkg) throws IOException, OpenXML4JException {
+  public OoxmlReader(OPCPackage pkg) {
     this.pkg = pkg;
 
     PackageRelationship coreDocRelationship = this.pkg.getRelationshipsByType(
@@ -133,7 +134,7 @@ public StylesTable getStylesTable() throws IOException, InvalidFormatException {
     }
     return styles;
   }
-
+  
   /**
    * Returns an InputStream to read the contents of the
    * main Workbook, which contains key overall data for
@@ -286,7 +287,7 @@ public String getSheetName() {
      * Returns the comments associated with this sheet,
      * or null if there aren't any
      */
-    public Comments getSheetComments() {
+    public Comments getSheetComments(StreamingReader.Builder builder) {
       PackagePart sheetPkg = getSheetPart();
 
       // Do we have a comments relationship? (Only ever one if so)
@@ -297,7 +298,7 @@ public Comments getSheetComments() {
           PackageRelationship comments = commentsList.getRelationship(0);
           PackagePartName commentsName = PackagingURIHelper.createPartName(comments.getTargetURI());
           PackagePart commentsPart = sheetPkg.getPackage().getPart(commentsName);
-          return parseComments(commentsPart);
+          return parseComments(builder, commentsPart);
         }
       } catch (InvalidFormatException|IOException e) {
         LOGGER.log(POILogger.WARN, e);
@@ -306,9 +307,16 @@ public Comments getSheetComments() {
       return null;
     }
 
-    //to allow subclassing
-    protected Comments parseComments(PackagePart commentsPart) throws IOException {
-      return new CommentsTable(commentsPart);
+    private Comments parseComments(StreamingReader.Builder builder, PackagePart commentsPart) throws IOException {
+      if (builder.useCommentsTempFile()) {
+        try (InputStream is = commentsPart.getInputStream()) {
+          TempFileCommentsTable ct = new TempFileCommentsTable(builder.encryptCommentsTempFile());
+          ct.readFrom(is);
+          return ct;
+        }
+      } else {
+        return new CommentsTable(commentsPart);
+      }
     }
 
     /**