CsvParser.java

   1 /*
   2  * Copyright (C) 2007-2012 Mathieu Baudier
   3  *
   4  * Licensed under the Apache License, Version 2.0 (the "License");
   5  * you may not use this file except in compliance with the License.
   6  * You may obtain a copy of the License at
   7  *
   8  *         http://www.apache.org/licenses/LICENSE-2.0
   9  *
  10  * Unless required by applicable law or agreed to in writing, software
  11  * distributed under the License is distributed on an "AS IS" BASIS,
  12  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  13  * See the License for the specific language governing permissions and
  14  * limitations under the License.
  15  */
  16 package org.argeo.util;
  17
  18 import java.io.BufferedReader;
  19 import java.io.IOException;
  20 import java.io.InputStream;
  21 import java.io.InputStreamReader;
  22 import java.util.ArrayList;
  23 import java.util.Collections;
  24 import java.util.List;
  25
  26 import org.argeo.ArgeoException;
  27
  28 /**
  29  * Parses a CSV file interpreting the first line as a header. The
  30  * {@link #parse(InputStream)} method and the setters are synchronized so that
  31  * the object cannot be modified when parsing.
  32  */
  33 public abstract class CsvParser {
  34         private char separator = ',';
  35         private char quote = '\"';
  36
  37         private Boolean noHeader = false;
  38         private Boolean strictLineAsLongAsHeader = true;
  39
  40         /**
  41          * Actually process a parsed line. If
  42          * {@link #setStrictLineAsLongAsHeader(Boolean)} is true (default) the
  43          * header and the tokens are guaranteed to have the same size.
  44          *
  45          * @param lineNumber
  46          *            the current line number, starts at 1 (the header, if header
  47          *            processing is enabled, the first line otherwise)
  48          * @param header
  49          *            the read-only header or null if {@link #setNoHeader(Boolean)}
  50          *            is true (default is false)
  51          * @param tokens
  52          *            the parsed tokens
  53          */
  54         protected abstract void processLine(Integer lineNumber,
  55                         List<String> header, List<String> tokens);
  56
  57         public synchronized void parse(InputStream in) {
  58                 parse(in, null);
  59         }
  60
  61         public synchronized void parse(InputStream in, String encoding) {
  62                 BufferedReader reader = null;
  63                 Integer lineCount = 0;
  64                 try {
  65                         if (encoding == null)
  66                                 reader = new BufferedReader(new InputStreamReader(in));
  67                         else
  68                                 reader = new BufferedReader(new InputStreamReader(in, encoding));
  69                         List<String> header = null;
  70                         if (!noHeader) {
  71                                 String headerStr = reader.readLine();
  72                                 if (headerStr == null)// empty file
  73                                         return;
  74                                 lineCount++;
  75                                 header = new ArrayList<String>();
  76                                 StringBuffer currStr = new StringBuffer("");
  77                                 Boolean wasInquote = false;
  78                                 while (parseLine(headerStr, header, currStr, wasInquote)) {
  79                                         headerStr = reader.readLine();
  80                                         if (headerStr == null)
  81                                                 break;
  82                                         wasInquote = true;
  83                                 }
  84                                 header = Collections.unmodifiableList(header);
  85                         }
  86
  87                         String line = null;
  88                         lines: while ((line = reader.readLine()) != null) {
  89                                 line = preProcessLine(line);
  90                                 if (line == null) {
  91                                         // skip line
  92                                         continue lines;
  93                                 }
  94                                 lineCount++;
  95                                 List<String> tokens = new ArrayList<String>();
  96                                 StringBuffer currStr = new StringBuffer("");
  97                                 Boolean wasInquote = false;
  98                                 sublines: while (parseLine(line, tokens, currStr, wasInquote)) {
  99                                         line = reader.readLine();
 100                                         if (line == null)
 101                                                 break sublines;
 102                                         wasInquote = true;
 103                                 }
 104                                 if (!noHeader && strictLineAsLongAsHeader) {
 105                                         int headerSize = header.size();
 106                                         int tokenSize = tokens.size();
 107                                         if (tokenSize == 1 && line.trim().equals(""))
 108                                                 continue lines;// empty line
 109                                         if (headerSize != tokenSize) {
 110                                                 throw new ArgeoException("Token size " + tokenSize
 111                                                                 + " is different from header size "
 112                                                                 + headerSize + " at line " + lineCount
 113                                                                 + ", line: " + line + ", header: " + header
 114                                                                 + ", tokens: " + tokens);
 115                                         }
 116                                 }
 117                                 processLine(lineCount, header, tokens);
 118                         }
 119                 } catch (ArgeoException e) {
 120                         throw e;
 121                 } catch (IOException e) {
 122                         throw new ArgeoException("Cannot parse CSV file (line: "
 123                                         + lineCount + ")", e);
 124                 } finally {
 125                         if (reader != null)
 126                                 try {
 127                                         reader.close();
 128                                 } catch (Exception e2) {
 129                                         // silent
 130                                 }
 131                 }
 132         }
 133
 134         /**
 135          * Called before each (logical) line is processed, giving a change to modify
 136          * it (typically for cleaning dirty files). To be overridden, return the
 137          * line unchanged by default. Skip the line if 'null' is returned.
 138          */
 139         protected String preProcessLine(String line) {
 140                 return line;
 141         }
 142
 143         /**
 144          * Parses a line character by character for performance purpose
 145          *
 146          * @return whether to continue parsing this line
 147          */
 148         protected Boolean parseLine(String str, List<String> tokens,
 149                         StringBuffer currStr, Boolean wasInquote) {
 150                 // List<String> tokens = new ArrayList<String>();
 151
 152                 // System.out.println("#LINE: " + str);
 153
 154                 if (wasInquote)
 155                         currStr.append('\n');
 156
 157                 char[] arr = str.toCharArray();
 158                 boolean inQuote = wasInquote;
 159                 // StringBuffer currStr = new StringBuffer("");
 160                 for (int i = 0; i < arr.length; i++) {
 161                         char c = arr[i];
 162                         if (c == separator) {
 163                                 if (!inQuote) {
 164                                         tokens.add(currStr.toString());
 165                                         // System.out.println("# TOKEN: " + currStr);
 166                                         currStr.delete(0, currStr.length());
 167                                 } else {
 168                                         // we don't remove separator that are in a quoted substring
 169                                         // System.out
 170                                         // .println("IN QUOTE, got a separator: [" + c + "]");
 171                                         currStr.append(c);
 172                                 }
 173                         } else if (c == quote) {
 174                                 if (inQuote && (i + 1) < arr.length && arr[i + 1] == quote) {
 175                                         // case of double quote
 176                                         currStr.append(quote);
 177                                         i++;
 178                                 } else {// standard
 179                                         inQuote = inQuote ? false : true;
 180                                 }
 181                         } else {
 182                                 currStr.append(c);
 183                         }
 184                 }
 185
 186                 if (!inQuote) {
 187                         tokens.add(currStr.toString());
 188                         // System.out.println("# TOKEN: " + currStr);
 189                 }
 190                 // if (inQuote)
 191                 // throw new ArgeoException("Missing quote at the end of the line "
 192                 // + str + " (parsed: " + tokens + ")");
 193                 if (inQuote)
 194                         return true;
 195                 else
 196                         return false;
 197                 // return tokens;
 198         }
 199
 200         public char getSeparator() {
 201                 return separator;
 202         }
 203
 204         public synchronized void setSeparator(char separator) {
 205                 this.separator = separator;
 206         }
 207
 208         public char getQuote() {
 209                 return quote;
 210         }
 211
 212         public synchronized void setQuote(char quote) {
 213                 this.quote = quote;
 214         }
 215
 216         public Boolean getNoHeader() {
 217                 return noHeader;
 218         }
 219
 220         public synchronized void setNoHeader(Boolean noHeader) {
 221                 this.noHeader = noHeader;
 222         }
 223
 224         public Boolean getStrictLineAsLongAsHeader() {
 225                 return strictLineAsLongAsHeader;
 226         }
 227
 228         public synchronized void setStrictLineAsLongAsHeader(
 229                         Boolean strictLineAsLongAsHeader) {
 230                 this.strictLineAsLongAsHeader = strictLineAsLongAsHeader;
 231         }
 232
 233 }