Merge branch 'master' of git.oschina.net:flashsword20/webmagic

Conflicts:
	webmagic-core/src/main/java/us/codecraft/webmagic/selector/RegexSelector.java

webmagic-core/src/main/java/us/codecraft/webmagic/selector/RegexSelector.java

@@ -1,93 +1,93 @@
-package us.codecraft.webmagic.selector;
+package us.codecraft.webmagic.selector;
-
+
-import org.apache.commons.lang3.StringUtils;
+import org.apache.commons.lang3.StringUtils;
-
+
-import java.util.ArrayList;
+import java.util.ArrayList;
-import java.util.List;
+import java.util.List;
-import java.util.regex.Matcher;
+import java.util.regex.Matcher;
-import java.util.regex.Pattern;
+import java.util.regex.Pattern;
-import java.util.regex.PatternSyntaxException;
+import java.util.regex.PatternSyntaxException;
-
+
-/**
+/**
- * Selector in regex.<br>
+ * Selector in regex.<br>
- *
+ *
- * @author code4crafter@gmail.com <br>
+ * @author code4crafter@gmail.com <br>
- * @since 0.1.0
+ * @since 0.1.0
- */
+ */
-public class RegexSelector implements Selector {
+public class RegexSelector implements Selector {
-
+
-    private String regexStr;
+    private String regexStr;
-
+
-    private Pattern regex;
+    private Pattern regex;
-
+
-    private int group = 1;
+    private int group = 1;
-
+
-    public RegexSelector(String regexStr, int group) {
+    public RegexSelector(String regexStr, int group) {
-        if (StringUtils.isBlank(regexStr)) {
+        if (StringUtils.isBlank(regexStr)) {
-            throw new IllegalArgumentException("regex must not be empty");
+            throw new IllegalArgumentException("regex must not be empty");
-        }
+        }
-        // Check bracket for regex group. Add default group 1 if there is no group.
+        // Check bracket for regex group. Add default group 1 if there is no group.
-        // Only check if there exists the valid left parenthesis, leave regexp validation for Pattern.
+        // Only check if there exists the valid left parenthesis, leave regexp validation for Pattern.
-        if (StringUtils.countMatches(regexStr, "(") - StringUtils.countMatches(regexStr, "\\(") ==
+        if (StringUtils.countMatches(regexStr, "(") - StringUtils.countMatches(regexStr, "\\(") ==
-                StringUtils.countMatches(regexStr, "(?:") - StringUtils.countMatches(regexStr, "\\(?:")) {
+                StringUtils.countMatches(regexStr, "(?:") - StringUtils.countMatches(regexStr, "\\(?:")) {
-            regexStr = "(" + regexStr + ")";
+            regexStr = "(" + regexStr + ")";
-        }
+        }
-        this.regexStr = regexStr;
+        this.regexStr = regexStr;
-        try {
+        try {
-            regex = Pattern.compile(regexStr, Pattern.DOTALL | Pattern.CASE_INSENSITIVE);
+            regex = Pattern.compile(regexStr, Pattern.DOTALL | Pattern.CASE_INSENSITIVE);
-        } catch (PatternSyntaxException e) {
+        } catch (PatternSyntaxException e) {
-            throw new IllegalArgumentException("invalid regex", e);
+            throw new IllegalArgumentException("invalid regex", e);
-        }
+        }
-        this.group = group;
+        this.group = group;
-    }
+    }
-
+
-    public RegexSelector(String regexStr) {
+    public RegexSelector(String regexStr) {
-        this(regexStr, 1);
+        this(regexStr, 1);
-    }
+    }
-
+
-    @Override
+    @Override
-    public String select(String text) {
+    public String select(String text) {
-        return selectGroup(text).get(group);
+        return selectGroup(text).get(group);
-    }
+    }
-
+
-    @Override
+    @Override
-    public List<String> selectList(String text) {
+    public List<String> selectList(String text) {
-        List<String> strings = new ArrayList<String>();
+        List<String> strings = new ArrayList<String>();
-        List<RegexResult> results = selectGroupList(text);
+        List<RegexResult> results = selectGroupList(text);
-        for (RegexResult result : results) {
+        for (RegexResult result : results) {
-            strings.add(result.get(group));
+            strings.add(result.get(group));
-        }
+        }
-        return strings;
+        return strings;
-    }
+    }
-
+
-    public RegexResult selectGroup(String text) {
+    public RegexResult selectGroup(String text) {
-        Matcher matcher = regex.matcher(text);
+        Matcher matcher = regex.matcher(text);
-        if (matcher.find()) {
+        if (matcher.find()) {
-            String[] groups = new String[matcher.groupCount() + 1];
+            String[] groups = new String[matcher.groupCount() + 1];
-            for (int i = 0; i < groups.length; i++) {
+            for (int i = 0; i < groups.length; i++) {
-                groups[i] = matcher.group(i);
+                groups[i] = matcher.group(i);
-            }
+            }
-            return new RegexResult(groups);
+            return new RegexResult(groups);
-        }
+        }
-        return RegexResult.EMPTY_RESULT;
+        return RegexResult.EMPTY_RESULT;
-    }
+    }
-
+
-    public List<RegexResult> selectGroupList(String text) {
+    public List<RegexResult> selectGroupList(String text) {
-        Matcher matcher = regex.matcher(text);
+        Matcher matcher = regex.matcher(text);
-        List<RegexResult> resultList = new ArrayList<RegexResult>();
+        List<RegexResult> resultList = new ArrayList<RegexResult>();
-        while (matcher.find()) {
+        while (matcher.find()) {
-            String[] groups = new String[matcher.groupCount() + 1];
+            String[] groups = new String[matcher.groupCount() + 1];
-            for (int i = 0; i < groups.length; i++) {
+            for (int i = 0; i < groups.length; i++) {
-                groups[i] = matcher.group(i);
+                groups[i] = matcher.group(i);
-            }
+            }
-            resultList.add(new RegexResult(groups));
+            resultList.add(new RegexResult(groups));
-        }
+        }
-        return resultList;
+        return resultList;
-    }
+    }
-
+
-    @Override
+    @Override
-    public String toString() {
+    public String toString() {
-        return regexStr;
+        return regexStr;
-    }
+    }
-
+
-}
+}