使用ANTLR4，用于代码镜像和基于Web的DSL的Primefaces扩展

DSL是很酷的东西，但是我不清楚它们有什么用。

然后我意识到它们对以下方面有好处：

摆脱复杂的UI

意思是

更快的做事方式

而已。当我阅读此博客时，我得出了这个结论。

如果您的用户是技术人员，并且不惧怕类似SQL的语法处理方式，则DSL特别适合

你有语法高亮
你有代码补全

否则，DSL有点烂。

因此，我不得不向客户提出一些概念证明。他有模糊的要求，要准确地提取团队需要的东西并不容易（他们需要很多东西，而且他很忙），因此DSL在此过程中可以提供很多帮助，因为人们在被迫时必须认真考虑他们的需求面对一种语法（甚至很小的一种）。

因此，我采用了以下技术：

用于代码镜像的JSF库Primefaces扩展
ANTLR4 （与ANTLR3相比有很大的改进，并且本书很棒）

不幸的是，我无法在两个工具中重用语法。实际上，我找不到能做到这一点的解决方案。至少对于基于Web的JSF解决方案。而且没有时间学习。因此，我不得不稍作修改。

首先，我们需要语法。 ANTLR4比ANTLR3更好，因为现在接线代码是通过访问者和侦听器完成的。语法内没有更多的Java代码。那很棒，而且更容易使用。

所以你可以有一个这样的语法

grammar Grammar;
options
{language = Java;
}
@lexer::header {package parsers;
}@parser::header {package parsers;
}
eval     :     expr EOF;
expr    :    'JOB' (jobName)? type 'TARGET' targetList ('START' startExpr)?
startExpr   :    'AT' cronTerm|    'AFTER' timeAmount timeUnits;
timeAmount: INT;   
jobName: STRING;   
targetList: STRING (',' STRING)*;   
type    :    deleteUser|    createUser;deleteUser: opDelete userName;
createUser: opCreate userName;
opDelete: 'DELETE';
opCreate: 'CREATE';
userName: STRING;cronTerm: '!'? (INT | '-' | '/' | '*' | '>' | '<')+;timeUnits:    'MINUTES'|    'HOURS'|    'DAYS'|    'WEEKS'|    'MONTHS';WS  :   [ \t\r\n]+ -> skip;STRING:  '"' ( ESC_SEQ | ~('\\'|'"') )* '"';fragment
HEX_DIGIT : ('0'..'9'|'a'..'f'|'A'..'F') ;fragment
ESC_SEQ:   '\\' ('b'|'t'|'n'|'f'|'r'|'\"'|'\''|'\\')|   UNICODE_ESC|   OCTAL_ESC;fragment
OCTAL_ESC:   '\\' ('0'..'3') ('0'..'7') ('0'..'7')|   '\\' ('0'..'7') ('0'..'7')|   '\\' ('0'..'7');fragment
UNICODE_ESC:   '\\' 'u' HEX_DIGIT HEX_DIGIT HEX_DIGIT HEX_DIGIT;ID  :    ('a'..'z'|'A'..'Z'|'_') ('a'..'z'|'A'..'Z'|'0'..'9'|'_')*;INT :    '0'..'9'+;

要编译语法，请尝试

public static void main(String[] args) {String[] arg0 = {"-visitor","/pathto/Grammar.g4"};org.antlr.v4.Tool.main(arg0);}

然后，ANTLR将为您生成类。

在我们的例子中，我们想访问解析树并检索我们想要的值。我们这样做扩展了生成的抽象类。

import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.StringTokenizer;import org.antlr.v4.runtime.tree.ErrorNode;import bsh.EvalError;
import bsh.Interpreter;public class MyLoader2 extends GrammarBaseVisitor<Void> {private String jobName,cronTerm,timeUnits,userName,jobType;private List<String> targetList;private boolean now,errorFound;private int timeAmount;   private Interpreter bsh = new Interpreter();private String eval(String s) throws EvaluationException{try {if (!s.startsWith("\"")){return s;}bsh.eval("String s="+s);return (String)bsh.eval("s");} catch (EvalError e) {throw new EvaluationException(s);}}@Overridepublic Void visitTimeAmount(TimeAmountContext ctx) {try{this.timeAmount = Integer.parseInt(ctx.getText());}catch(java.lang.NumberFormatException nfe){throw new InvalidTimeAmountException(ctx.getText());}return super.visitTimeAmount(ctx);}@Overridepublic Void visitUserName(UserNameContext ctx) {this.userName = eval(ctx.getText());return super.visitUserName(ctx);}@Overridepublic Void visitCronTerm(CronTermContext ctx) {this.cronTerm = eval(ctx.getText());return super.visitCronTerm(ctx);}@Overridepublic Void visitTimeUnits(TimeUnitsContext ctx) {this.timeUnits = ctx.getText();return super.visitTimeUnits(ctx);}@Overridepublic Void visitTargetList(TargetListContext ctx) {this.targetList = toStringList(ctx.getText());return super.visitTargetList(ctx);}@Overridepublic Void visitJobName(JobNameContext ctx) {this.jobName = eval(ctx.getText());return super.visitJobName(ctx);}@Overridepublic Void visitOpCreate(OpCreateContext ctx) {this.jobType = ctx.getText();return super.visitOpCreate(ctx);}@Overridepublic Void visitOpDelete(OpDeleteContext ctx) {this.jobType = ctx.getText();return super.visitOpDelete(ctx);}private List<String> toStringList(String text) {List<String> l = new ArrayList<String>();StringTokenizer st = new StringTokenizer(text," ,");while(st.hasMoreElements()){l.add(eval(st.nextToken()));}return l;}private Map<String, String> toMapList(String text) throws InvalidItemsException, InvalidKeyvalException {Map<String, String> m = new HashMap<String, String>();if (text == null || text.trim().length() == 0){return m;}String[] items = text.split(",");if (items.length == 0){throw new InvalidItemsException();}for(String item:items){String[] keyval = item.split("=");if (keyval.length == 2){m.put(keyval[0], keyval[1]);}else{throw new InvalidKeyvalException(keyval.length);}}return m;}public String getJobName() {return jobName;}public String getCronTerm() {return cronTerm;}public String getTimeUnits() {return timeUnits;}public String getUserName() {return userName;}public String getJobType() {return jobType;}public List<String> getTargetList() {return targetList;}public boolean isNow() {return now;}public int getTimeAmount() {return timeAmount;}@Overridepublic Void visitOpNow(OpNowContext ctx) {this.now = ctx.getText().equals("NOW");return super.visitOpNow(ctx);}public boolean isErrorFound() {return errorFound;}@Overridepublic Void visitErrorNode(ErrorNode node) {this.errorFound = true;return super.visitErrorNode(node);}
}

请注意， beanshell解释器用于将“ xyz”之类的字符串评估为xyz。这对于其中包含转义引号和字符的字符串特别有用。

因此，您有了语法和visiter / loader bean，然后我们可以对其进行测试：

private static MyLoader getLoader(String str){ANTLRInputStream input = new ANTLRInputStream(str);GrammarLexer lexer = new GrammarLexer(input);CommonTokenStream tokens = new CommonTokenStream(lexer);GrammarParser parser = new GrammarParser(tokens);ParseTree tree = parser.eval();MyLoader loader = new MyLoader();loader.visit(tree);return loader;
}public static void main(String[] args){MyLoader loader = getLoader("JOB \"jobName\" CREATE \"myuser\" TARGET \"site1\",\"site2\" START AFTER 1 DAY");System.out.println(loader.getJobName());System.out.println(loader.getJobType());
}

大。现在是黑客。 Code Mirror支持自定义语法，但JSF Primefaces扩展中不存在。因此，我打开了resources-codemirror-1.2.0.jar，打开了/META-INF/resources/primefaces-extensions/codemirror/mode/modes.js文件，对其进行了格式化（以便我可以阅读），然后我刚刚选择了最简单的语言作为我的新自定义sintax荧光笔！

我改名了

(...)
}, "xml"), CodeMirror.defineMIME("text/x-markdown", "markdown"), CodeMirror.defineMode("mylanguage", function (e) {
(...)var t = e.indentUnit,n, i = r(["site", "type", "targetList"]),s = r(["AT","AFTER","CREATE","MINUTES","HOURS","TARGET","MONTHS","JOB","DAYS","DELETE","START","WEEKS" ]),
(...)
}), CodeMirror.defineMIME("text/x-mylanguage", "mylanguage"), CodeMirror.defineMode("ntriples", function () {(...)

那些在“ s = r”中用大写字母表示的标记是将被突出显示的标记，而在“ i = r”中的那些标记则是将被突出显示的标记。为什么我们都想要？因为第二种类型是“占位符”，我的意思是，我们将它们用于自动填充内容。

好，那么您的JSF xhtml页面将如下所示

<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml"xmlns:f="http://java.sun.com/jsf/core"xmlns:h="http://java.sun.com/jsf/html"xmlns:ui="http://java.sun.com/jsf/facelets"xmlns:p="http://primefaces.org/ui"xmlns:pe="http://primefaces.org/ui/extensions"><h:body>
<h:form id="form">
<pe:codeMirrorid="codeMirror"style="width:600px;" mode="myLanguage" widgetVar="myCodeMirror" theme="eclipse"value="#{myMB.script}"lineNumbers="true"             completeMethod="#{myMB.complete}" extraKeys="{ 'Ctrl-Space': function(cm) { PF('myCodeMirror').complete(); }}"/> 
<p:commandButton value="Verify" action="#{myMB.verify}" />
(...)

现在，我们需要自动完成功能。这是无聊的部分。您必须手动完成大多数完成工作，因为没有上下文信息（请记住，我没有时间学习……），所以快速而肮脏的方式是这样的

in myMBpublic List<String> complete(final CompleteEvent event) {try {return this.myEJB.complete(event.getToken());} catch (Exception e) {jsfUtilEJB.addErrorMessage(e,"Could not complete");return null;}} in myEJBprivate static final String    SITE    = "site_";
public List<String> complete(String token) throws Exception {if (token == null || token.trim().length() == 0){return null;}else{List<String> suggestions = new ArrayList<String>();switch(token){//first search variablescase "targetlist":for(String v:TARGETS){suggestions.add(v);   }break;case "site":List<Site> allSites = this.baseService.getSiteDAO().getAll();for(Site s:allSites){suggestions.add("\""+SITE+s.getName()+"\"");}break;case "type":suggestions.add("DELETE \"userName\"");suggestions.add("CREATE \"userName\"");break;case "AT":suggestions.add("AT \"cronExpression\"");suggestions.add("AT \"0 * * * * * * *\"");break;case "AFTER":for(int a:AMOUNTS){for(String u:UNITS){if (a == 1){suggestions.add("AFTER"+" "+a+" "+u);}else{suggestions.add("AFTER"+" "+a+" "+u+"S");}}}break;case "TARGET":for(String v:TARGETS){suggestions.add("TARGET "+v+"");   }break;case "JOB":suggestions.add("JOB \"jobName\" \ntype \nTARGET targetlist \nSTART");break;case "START":suggestions.add("START AT \"cronExpression\"");suggestions.add("START AT \"0 * * * * * * *\"");for(int a:AMOUNTS){for(String u:UNITS){if (a == 1){suggestions.add("START AFTER"+" "+a+" "+u);}else{suggestions.add("START AFTER"+" "+a+" "+u+"S");}}}suggestions.add("START NOW");break;case "DELETE":suggestions.add("DELETE \"userName\"");break;case "CREATE":suggestions.add("CREATE \"userName\"");break;default:if (token.startsWith(SITE)){List<Site> matchedSites = this.baseService.getSiteDAO().getByPattern(token.substring(SITE.length())+"*");for(Site s:matchedSites){suggestions.add("\""+SITE+s.getName()+"\"");}}else{//then search substringsfor(String kw:KEYWORDS){if (kw.toLowerCase().startsWith(token.toLowerCase())){suggestions.add(kw);}}}}//end switch//remove dups and sortSet<String> ts = new TreeSet<String>(suggestions);           return new ArrayList<String>(ts);}
}private static final int[] AMOUNTS = {1,5,10};
private static final String[] UNITS = {"MINUTE","HOUR","DAY","WEEK","MONTH"};
private static final String[] TARGETS = {"site"};/** KEYWORDS are basic suggestions*/
private static final String[] KEYWORDS = {"AT","AFTER","CREATE","MINUTES","HOURS","TARGET","MONTHS","JOB","DAYS","DELETE","START","WEEKS"};

因此，关键字的自动填充内容将仅向您显示字段和更多关键字，而“占位符”（还记得jar中的Codemirror javascript中的小写关键字吗？）用从数据库中获取的实际值来完成。另外，您可以使用部分字符串来检索那些以子字符串开头的字符串，如下所示：

屏幕截图

当然，在JPA中类似模式的搜索可以这样执行：

public abstract class GenericDAO<E> {protected EntityManager entityManager;private Class<E> clazz;private EntityType<E> pClass;@SuppressWarnings("unchecked")public GenericDAO(EntityManager entityManager) {this.entityManager = entityManager;ParameterizedType genericSuperclass = (ParameterizedType) getClass().getGenericSuperclass();this.clazz = (Class<E>) genericSuperclass.getActualTypeArguments()[0];EntityManagerFactory emf = this.entityManager.getEntityManagerFactory();Metamodel metamodel = emf.getMetamodel();this.pClass = metamodel.entity(clazz);}public List<E> getByPattern(String pattern) {pattern = pattern.replace("?", "_").replace("*", "%");CriteriaBuilder cb = entityManager.getCriteriaBuilder();CriteriaQuery<E> q = cb.createQuery(clazz);Root<E> entity = q.from(clazz);SingularAttribute<E, String> singularAttribute = (SingularAttribute<E, String>) pClass.getDeclaredSingularAttribute(getNameableField(clazz));Path<String> path = entity.get(singularAttribute);q.where(cb.like(path, pattern));q.select(entity);TypedQuery<E> tq = entityManager.createQuery(q);List<E> all = tq.getResultList();return all;}private String getNameableField(Class<E> clazz) {for(Field f : clazz.getDeclaredFields()) {for(Annotation a : f.getAnnotations()) {if(a.annotationType() == Nameable.class) {return f.getName();}}}return null;}
(...)

其中Nameable是您的实体类的注释：

@Retention(RetentionPolicy.RUNTIME)
@Target({ElementType.FIELD})
public @interface Nameable {
}

使用它来注释实体类中的单个列，即String。像这样：

@Entity
@Table(uniqueConstraints=@UniqueConstraint(columnNames={"name"}))
public class Site implements Serializable {/****/private static final long serialVersionUID = 8008732613898597654L;@Id@GeneratedValue(strategy = GenerationType.AUTO)private Long id;@Nameable@Column(nullable=false)private String name;
(...)

当然，“验证”按钮只是获取您的脚本并将其推送到加载程序中。

参考：来自Leo's Notepad博客的JCG合作伙伴 Leonardo Shikida 使用ANTLR4，用于代码镜像的Primefaces扩展和基于Web的DSL玩。