目录
前言
本文是关于文本对比和通过diff算法的简单介绍,如果你想快速通过java+html或者纯JavaScript来实现两文件并排对比可看我另一篇文章:
一、文本差异对比介绍
举一个最常见的例子,我们使用git进行提交时,通常会使用git diff --cached
来查看这次提交做了哪些改动,这里我们先简单定义一下什么是diff:diff就是目标文本和源文本之间的区别,也就是将源文本变成目标文本所需要的操作。
在 Git 中,有四种diff算法,分别是Myers、Minimal、Patience和Histogram,它们用于获取位于两个不同提交中的两个相同文件的差异。
Myers算法由Eugene W.Myers在1986年发表的一篇论文中提出,是一个能在大部分情况产生”最短的直观的“diff的一个算法。
文本差异对比涉及到的算法介绍 :
How different are different diff algorithms in Git
Myers Diff 差分算法
好用的在线文本对比: 在线文本比较工具
好用的桌面文本对比软件:Meld、DiffMerge、xxdiff、Diffuse、Kompare
常见的文本差异对比库:
本文通过java-diff-utils库来实现文本差异对比
二、依赖下载
jar下载网址:io.github.java-diff-utils:java-diff-utils:4.11
官方git地址:java-diff-utils
官方demo示例: java-diff-utils Examples
如果你使用Maven:
<dependency>
<groupId>io.github.java-diff-utils</groupId>
<artifactId>java-diff-utils</artifactId>
<version>4.11</version>
</dependency>
如果你使用Gradle:
implementation 'io.github.java-diff-utils:java-diff-utils:4.11'
本文下面会用到的三个对比文本如下:
test1.txt 和 test3.txt :
_this.ispc = function(){
var userAgentInfo = navigator.userAgent;
var Agents = ["Android", "iPhone",
"SymbianOS", "Windows Phone",
"iPad", "iPod"];
var flag = true;
for (var v = 0; v < Agents.length; v++) {
if (userAgentInfo.indexOf(Agents[v]) > 0) {
flag = false;
break;
}
}
return flag;
}
window._assignInfo = {};
window._curnodepersons = [];
window.attachmode = '0';
window.isEsignature = false;
window.upList = [];
window.downList = [];
window.billId = '';
window.upProcessIdList = [];
window.downProcessIdList = [];
var urlsplit = window.location.href.split("#");
if(urlsplit.length>1){
//alert(window.location.href);
location.href = urlsplit[0];
}
test2.txt:
_this.ispc = function(){
var userAgentInfo = navigator.userAgent;
var Agents = ["Android", "iPhone",
"SymbianOS", "Windows Phone",
"iPad", "iPod"];
var flag = true;
for (var v = 0; v < Agents.length; v++) {
if (userAgentInfo.indexOf(Agents[v]) > 0) {
flag = false;
insert1;
insert2;
break;
}
}
return flag;
}
window._assignInfo = {};
window._curnodepersons = [];
window.attachmode = '0';
window.isEsignature = false;
add
window.billId = '';
window.upProcessIdList = [];
window.downProcessIdList = [];
var urlsplit = window.location.href.split("#");
if(urlsplit.length>1){
//alert(window.location.href);
location.href = qazwer[0];
}
三、获取两文件的不同点-patch
//原始文件
List<String> original = Files.readAllLines(new File("D:\\test1.txt").toPath());
//对比文件
List<String> revised = Files.readAllLines(new File("D:\\test2.txt").toPath());
//两文件的不同点
Patch<String> patch = DiffUtils.diff(original, revised);
for (AbstractDelta<String> delta : patch.getDeltas()) {
System.out.println(delta);
}
输出:
[InsertDelta, position: 9, lines: [insert1;, insert2;]]
[ChangeDelta, position: 18, lines: [window.upList = [];, window.downList = [];] to [, , add]]
[ChangeDelta, position: 26, lines: [ location.href = urlsplit[0];] to [ location.href = qazwer[0];]]
InsertDelta代表插入的,ChangeDelta代表删除的或修改的。position代表第几行,lines代表内容。
四、根据patch生成统一的差异格式unifiedDiff
//原始文件
List<String> original = Files.readAllLines(new File("D:\\test1.txt").toPath());
//对比文件
List<String> revised = Files.readAllLines(new File("D:\\test2.txt").toPath());
//两文件的不同点
Patch<String> patch = DiffUtils.diff(original, revised);
//生成统一的差异格式
List<String> unifiedDiff = UnifiedDiffUtils.generateUnifiedDiff("test1.txt", "test2.txt", original, patch, 0);
unifiedDiff.forEach(System.out::println);
输出diff:
--- test1.txt
+++ test2.txt
@@ -10,0 +10,2 @@
+insert1;
+insert2;
@@ -19,2 +21,3 @@
-window.upList = [];
-window.downList = [];
+
+
+add
@@ -27,1 +30,1 @@
- location.href = urlsplit[0];
+ location.href = qazwer[0];
说明:
关于输出的diff可参考 :读懂diff
减号代表原始文件test1.txt,加号代表对比文件test2.txt
@@ -10,0 +10,2 @@
+insert1;
+insert2
表示test2增加了2行,分别加了insert1;和insert2这两行
@@ -19,2 +21,3 @@
-window.upList = [];
-window.downList = [];
+
+
+add
表示test2删除了2行,删除的这两行是test1从19行开始的2行;然后test2又增加了3行,增加是3行是从test2的21行开始的。
@@ -27,1 +30,1 @@
- location.href = urlsplit[0];
+ location.href = qazwer[0];
表示test2删除了1行,删除的这行是test1从27行开始的1行;然后test2又增加了1行,增加是1行是从test2的30行开始的。(相当于这一行进行了修改)
前端页面美化输出:
如果想将上面通过Java代码得到的两个文件的差异美化输出可以 参考我另一篇博客:Java+html实现文本对比 实现效果如下:
如果想获取全部内容对比,要把diff加入到文件内容中,也就是上面代码的diffString 替换为如下(没改动的代码前要加一个空格才会显示):
const diffString = `
--- test1.txt
+++ test2.txt
@@ -0,0 +0,0 @@
this.ispc = function(){
var userAgentInfo = navigator.userAgent;
var Agents = ["Android", "iPhone",
"SymbianOS", "Windows Phone",
"iPad", "iPod"];
var flag = true;
for (var v = 0; v < Agents.length; v++) {
if (userAgentInfo.indexOf(Agents[v]) > 0) {
flag = false;
@@ -10,0 +10,2 @@
+insert1;
+insert2;
break;
}
}
return flag;
}
window._assignInfo = {};
window._curnodepersons = [];
window.attachmode = '0';
window.isEsignature = false;
@@ -19,2 +21,3 @@
-window.upList = [];
-window.downList = [];
+
+
+add
window.billId = ;
window.upProcessIdList = [];
window.downProcessIdList = [];
var urlsplit = window.location.href.split();
if(urlsplit.length>1){
//alert(window.location.href);
@@ -27,1 +30,1 @@
- location.href = urlsplit[0];
+ location.href = qazwer[0];
}
`;
如果你想通过js获取diff,然后通过diff2html渲染到界面可以参考:
实现Diff页面的工程实践
vue-diff-demo
五、根据unifiedDiff打补丁
假设你有test1.txt、test2.txt、test3.txt 三个内容一样的文件,有一天你改了test2.txt的文件内容,你想把test2.txt修改的地方也同步的修改运用到 test3.txt上就可以通过打补丁的方式来实现。简单的来说就是test1.txt、test2.txt进行对比得到 unifiedDiff ,把 unifiedDiff运用到 test3.txt 。(相当于git的不同分支的代码同步)
//原始文件
List<String> original = Files.readAllLines(new File("D:\\test1.txt").toPath());
//对比文件
List<String> revised = Files.readAllLines(new File("D:\\test2.txt").toPath());
//两文件的不同点
Patch<String> patch = DiffUtils.diff(original, revised);
//生成统一的差异格式
List<String> unifiedDiff = UnifiedDiffUtils.generateUnifiedDiff("test1.txt", "test2.txt", original, patch, 0);
//从文件或此处从内存导入统一差异格式到补丁
Patch<String> importedPatch = UnifiedDiffUtils.parseUnifiedDiff(unifiedDiff);
List<String> test3 = Files.readAllLines(new File("D:\\test3.txt").toPath());
//将差异运用到其他文件打补丁,即将不同点运用到其他文件(相当于git的冲突合并)
List<String> patchedText = DiffUtils.patch(test3, importedPatch);
for (String patchedTextPow : patchedText) {
System.out.println(patchedTextPow);
}
输出:
_this.ispc = function(){
var userAgentInfo = navigator.userAgent;
var Agents = ["Android", "iPhone",
"SymbianOS", "Windows Phone",
"iPad", "iPod"];
var flag = true;
for (var v = 0; v < Agents.length; v++) {
if (userAgentInfo.indexOf(Agents[v]) > 0) {
flag = false;
insert1;
insert2;
break;
}
}
return flag;
}
window._assignInfo = {};
window._curnodepersons = [];
window.attachmode = '0';
window.isEsignature = false;
add
window.billId = '';
window.upProcessIdList = [];
window.downProcessIdList = [];
var urlsplit = window.location.href.split("#");
if(urlsplit.length>1){
//alert(window.location.href);
location.href = qazwer[0];
}
六、对比两文件的不同点并按行显示不同
//原始文件
List<String> text1= Files.readAllLines(new File("D:\\test2.txt").toPath());
//对比文件
List<String> text2=Files.readAllLines(new File("D:\\test1.txt").toPath());
//行比较器,原文件删除的内容用"~"包裹,对比文件新增的内容用"**"包裹
DiffRowGenerator generator = DiffRowGenerator.create()
.showInlineDiffs(true)
.inlineDiffByWord(true)
.oldTag(f -> "~")
.newTag(f -> "**")
.build();
//通过行比较器对比得到每一行的不同
List<DiffRow> rows = generator.generateDiffRows(text1, text2);
//输出每一行的原始文件和对比文件,每一行的原始文件和对比文件通过 "|"分割
for (DiffRow row : rows) {
System.out.println(row.getOldLine() + "|" + row.getNewLine());
}
输出:
_this.ispc = function(){|_this.ispc = function(){
var userAgentInfo = navigator.userAgent;| var userAgentInfo = navigator.userAgent;
var Agents = ["Android", "iPhone",| var Agents = ["Android", "iPhone",
"SymbianOS", "Windows Phone",| "SymbianOS", "Windows Phone",
"iPad", "iPod"];| "iPad", "iPod"];
var flag = true;| var flag = true;
for (var v = 0; v < Agents.length; v++) {| for (var v = 0; v < Agents.length; v++) {
if (userAgentInfo.indexOf(Agents[v]) > 0) {| if (userAgentInfo.indexOf(Agents[v]) > 0) {
flag = false;| flag = false;
~insert1;~|
~insert2;~|
break;| break;
}| }
}| }
return flag;| return flag;
}|}
window._assignInfo = {};|window._assignInfo = {};
window._curnodepersons = [];|window._curnodepersons = [];
window.attachmode = '0';|window.attachmode = '0';
window.isEsignature = false;|window.isEsignature = false;
|**window.upList = [];**
|**window.downList = [];**
~add~|
window.billId = '';|window.billId = '';
window.upProcessIdList = [];|window.upProcessIdList = [];
window.downProcessIdList = [];|window.downProcessIdList = [];
var urlsplit = window.location.href.split("#");|var urlsplit = window.location.href.split("#");
if(urlsplit.length>1){|if(urlsplit.length>1){
//alert(window.location.href);| //alert(window.location.href);
location.href = ~qazwer~[0];| location.href = **urlsplit**[0];
}|}
参考:
Diffing more quickly
Compare files side by side and hightlight diff using Java | Apache Commons Text diff | Myers algorithm
Myers diff algorithm vs Hunt–McIlroy algorithm