✨ Improve code block function
This commit is contained in:
@@ -4,6 +4,7 @@
|
||||
|
||||
import { EditorState } from '@codemirror/state';
|
||||
import { syntaxTree, syntaxTreeAvailable } from '@codemirror/language';
|
||||
import { Block as BlockNode, BlockDelimiter, BlockContent, BlockLanguage, Document } from './lang-parser/parser.terms.js';
|
||||
import {
|
||||
CodeBlock,
|
||||
SupportedLanguage,
|
||||
@@ -17,121 +18,81 @@ import {
|
||||
Block
|
||||
} from './types';
|
||||
|
||||
/**
|
||||
* 语言检测工具
|
||||
*/
|
||||
class LanguageDetector {
|
||||
// 语言关键字映射
|
||||
private static readonly LANGUAGE_PATTERNS: Record<string, RegExp[]> = {
|
||||
javascript: [
|
||||
/\b(function|const|let|var|class|extends|import|export|async|await)\b/,
|
||||
/\b(console\.log|document\.|window\.)\b/,
|
||||
/=>\s*[{(]/
|
||||
],
|
||||
typescript: [
|
||||
/\b(interface|type|enum|namespace|implements|declare)\b/,
|
||||
/:\s*(string|number|boolean|object|any)\b/,
|
||||
/<[A-Z][a-zA-Z0-9<>,\s]*>/
|
||||
],
|
||||
python: [
|
||||
/\b(def|class|import|from|if __name__|print|len|range)\b/,
|
||||
/^\s*#.*$/m,
|
||||
/\b(True|False|None)\b/
|
||||
],
|
||||
java: [
|
||||
/\b(public|private|protected|static|final|class|interface)\b/,
|
||||
/\b(System\.out\.println|String|int|void)\b/,
|
||||
/import\s+[a-zA-Z0-9_.]+;/
|
||||
],
|
||||
html: [
|
||||
/<\/?[a-zA-Z][^>]*>/,
|
||||
/<!DOCTYPE\s+html>/i,
|
||||
/<(div|span|p|h[1-6]|body|head|html)\b/
|
||||
],
|
||||
css: [
|
||||
/[.#][a-zA-Z][\w-]*\s*{/,
|
||||
/\b(color|background|margin|padding|font-size):\s*[^;]+;/,
|
||||
/@(media|keyframes|import)\b/
|
||||
],
|
||||
json: [
|
||||
/^\s*[{\[][\s\S]*[}\]]\s*$/,
|
||||
/"[^"]*":\s*(".*"|[\d.]+|true|false|null)/,
|
||||
/,\s*$/m
|
||||
],
|
||||
sql: [
|
||||
/\b(SELECT|FROM|WHERE|INSERT|UPDATE|DELETE|CREATE|ALTER|DROP)\b/i,
|
||||
/\b(JOIN|LEFT|RIGHT|INNER|OUTER|ON|GROUP BY|ORDER BY)\b/i,
|
||||
/;\s*$/m
|
||||
],
|
||||
shell: [
|
||||
/^#!/,
|
||||
/\b(echo|cd|ls|grep|awk|sed|cat|chmod)\b/,
|
||||
/\$\{?\w+\}?/
|
||||
],
|
||||
markdown: [
|
||||
/^#+\s+/m,
|
||||
/\*\*.*?\*\*/,
|
||||
/\[.*?\]\(.*?\)/,
|
||||
/^```/m
|
||||
]
|
||||
};
|
||||
|
||||
/**
|
||||
* 检测文本的编程语言
|
||||
*/
|
||||
static detectLanguage(text: string): LanguageDetectionResult {
|
||||
if (!text.trim()) {
|
||||
return { language: 'text', confidence: 1.0 };
|
||||
}
|
||||
|
||||
const scores: Record<string, number> = {};
|
||||
|
||||
// 对每种语言计算匹配分数
|
||||
for (const [language, patterns] of Object.entries(this.LANGUAGE_PATTERNS)) {
|
||||
let score = 0;
|
||||
const textLower = text.toLowerCase();
|
||||
|
||||
for (const pattern of patterns) {
|
||||
const matches = text.match(pattern);
|
||||
if (matches) {
|
||||
score += matches.length;
|
||||
}
|
||||
}
|
||||
|
||||
// 根据文本长度标准化分数
|
||||
scores[language] = score / Math.max(text.length / 100, 1);
|
||||
}
|
||||
|
||||
// 找到最高分的语言
|
||||
const bestMatch = Object.entries(scores)
|
||||
.sort(([, a], [, b]) => b - a)[0];
|
||||
|
||||
if (bestMatch && bestMatch[1] > 0) {
|
||||
return {
|
||||
language: bestMatch[0] as SupportedLanguage,
|
||||
confidence: Math.min(bestMatch[1], 1.0)
|
||||
};
|
||||
}
|
||||
|
||||
return { language: 'text', confidence: 1.0 };
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* 从语法树解析代码块
|
||||
*/
|
||||
export function getBlocksFromSyntaxTree(state: EditorState): CodeBlock[] | null {
|
||||
export function getBlocksFromSyntaxTree(state: EditorState): Block[] | null {
|
||||
if (!syntaxTreeAvailable(state)) {
|
||||
return null;
|
||||
}
|
||||
|
||||
const tree = syntaxTree(state);
|
||||
const blocks: CodeBlock[] = [];
|
||||
const blocks: Block[] = [];
|
||||
const doc = state.doc;
|
||||
|
||||
// TODO: 如果使用自定义 Lezer 语法，在这里实现语法树解析
|
||||
// 目前先返回 null，使用字符串解析作为后备
|
||||
return null;
|
||||
// 遍历语法树中的所有块
|
||||
tree.iterate({
|
||||
enter(node) {
|
||||
if (node.type.id === BlockNode) {
|
||||
// 查找块的分隔符和内容
|
||||
let delimiter: { from: number; to: number } | null = null;
|
||||
let content: { from: number; to: number } | null = null;
|
||||
let language = 'text';
|
||||
let auto = false;
|
||||
|
||||
// 遍历块的子节点
|
||||
const blockNode = node.node;
|
||||
blockNode.firstChild?.cursor().iterate(child => {
|
||||
if (child.type.id === BlockDelimiter) {
|
||||
delimiter = { from: child.from, to: child.to };
|
||||
|
||||
// 解析整个分隔符文本来获取语言和自动检测标记
|
||||
const delimiterText = doc.sliceString(child.from, child.to);
|
||||
console.log('🔍 [解析器] 分隔符文本:', delimiterText);
|
||||
|
||||
// 使用正则表达式解析分隔符
|
||||
const match = delimiterText.match(/∞∞∞([a-zA-Z0-9_-]+)(-a)?\n/);
|
||||
if (match) {
|
||||
language = match[1] || 'text';
|
||||
auto = match[2] === '-a';
|
||||
console.log(`🔍 [解析器] 解析结果: 语言=${language}, 自动=${auto}`);
|
||||
} else {
|
||||
// 回退到逐个解析子节点
|
||||
child.node.firstChild?.cursor().iterate(langChild => {
|
||||
if (langChild.type.id === BlockLanguage) {
|
||||
const langText = doc.sliceString(langChild.from, langChild.to);
|
||||
language = langText || 'text';
|
||||
}
|
||||
// 检查是否有自动检测标记
|
||||
if (doc.sliceString(langChild.from, langChild.to) === '-a') {
|
||||
auto = true;
|
||||
}
|
||||
});
|
||||
}
|
||||
} else if (child.type.id === BlockContent) {
|
||||
content = { from: child.from, to: child.to };
|
||||
}
|
||||
});
|
||||
|
||||
if (delimiter && content) {
|
||||
blocks.push({
|
||||
language: {
|
||||
name: language as SupportedLanguage,
|
||||
auto: auto,
|
||||
},
|
||||
content: content,
|
||||
delimiter: delimiter,
|
||||
range: {
|
||||
from: node.from,
|
||||
to: node.to,
|
||||
},
|
||||
});
|
||||
}
|
||||
}
|
||||
}
|
||||
});
|
||||
|
||||
return blocks.length > 0 ? blocks : null;
|
||||
}
|
||||
|
||||
// 跟踪第一个分隔符的大小
|
||||
@@ -308,6 +269,13 @@ export function getBlocksFromString(state: EditorState): Block[] {
|
||||
* 获取文档中的所有块
|
||||
*/
|
||||
export function getBlocks(state: EditorState): Block[] {
|
||||
// 优先使用语法树解析
|
||||
const syntaxTreeBlocks = getBlocksFromSyntaxTree(state);
|
||||
if (syntaxTreeBlocks) {
|
||||
return syntaxTreeBlocks;
|
||||
}
|
||||
|
||||
// 如果语法树不可用，回退到字符串解析
|
||||
return getBlocksFromString(state);
|
||||
}
|
||||
|
||||
|
||||
Reference in New Issue
Block a user