feat: implement limit without consider image

close: #14

src/github/renderer/make-mark.ts

@@ -0,0 +1,54 @@
+import { fromMarkdown } from 'mdast-util-from-markdown';
+import { Content, Parent } from 'mdast-util-from-markdown/lib/index';
+import { gfmFromMarkdown, gfmToMarkdown } from 'mdast-util-gfm';
+import { toMarkdown } from 'mdast-util-to-markdown';
+import { gfm } from 'micromark-extension-gfm';
+
+export function parseMarkdown(text: string) {
+  const tree = fromMarkdown(trimLeadingWS(text), {
+    extensions: [gfm()],
+    mdastExtensions: [gfmFromMarkdown()],
+  });
+  return tree;
+}
+
+function trimLeadingWS(str: string) {
+  /*
+    Get the initial indentation
+    But ignore new line characters
+  */
+  const matcher = /^[\r\n]?(\s+)/;
+  if (matcher.test(str)) {
+    /*
+      Replace the initial whitespace 
+      globally and over multiple lines
+    */
+    return str.replace(new RegExp('^' + str.match(matcher)![1], 'gm'), '');
+  } else {
+    // Regex doesn't match so return the original string
+    return str;
+  }
+}
+
+export function walk(root: Parent, cb: (token: Content) => boolean | void) {
+  root.children.forEach((node) => {
+    if (node) {
+      const skip = cb(node);
+      if (!skip) {
+        if ((node as Parent).children) {
+          walk(node as Parent, cb);
+        }
+      }
+    }
+  });
+}
+
+export function makeMarkdown(tree: Parent) {
+  return toMarkdown(tree, {
+    extensions: [gfmToMarkdown()],
+    listItemIndent: 'one',
+    rule: '-',
+    fence: '`',
+    fences: true,
+  });
+}

src/github/templates/utils.ts

@@ -1,6 +1,10 @@
 import capitalize from 'lodash/capitalize';
 
-import { StringBuilder, limitTextByPosition } from '@/utils/string-builder';
+import {
+  StringBuilder,
+  limitTextByPosition,
+  tryReplaceImageToNull,
+} from '@/utils/string-builder';
 
 import { render } from '../renderer';
 

src/github/utils.ts

@@ -1,20 +1,11 @@
 import { EmitterWebhookEventName } from '@octokit/webhooks';
-import { fromMarkdown } from 'mdast-util-from-markdown';
-import {
-  Root,
-  Content,
-  Parent,
-  Link,
-  Text,
-} from 'mdast-util-from-markdown/lib/index';
-import { gfmFromMarkdown, gfmToMarkdown } from 'mdast-util-gfm';
-import { toMarkdown } from 'mdast-util-to-markdown';
-import { gfm } from 'micromark-extension-gfm';
+import { Link, Text } from 'mdast-util-from-markdown/lib/index';
 
 import { markdown } from '@/im/message';
 import { ISetting } from '@/kv/types';
 import { send } from '@opensumi/dingtalk-bot/lib/utils';
 
+import { makeMarkdown, parseMarkdown, walk } from './renderer/make-mark';
 import { MarkdownContent } from './types';
 
 function securityInterception(text: string) {
@@ -47,45 +38,6 @@ function checkLinkIsLiteral(link: Link) {
   }
 }
 
-export function walk(root: Parent, cb: (token: Content) => boolean | void) {
-  root.children.forEach((node) => {
-    if (node) {
-      const skip = cb(node);
-      if (!skip) {
-        if ((node as Parent).children) {
-          walk(node as Parent, cb);
-        }
-      }
-    }
-  });
-}
-
-function trimLeadingWS(str: string) {
-  /*
-    Get the initial indentation
-    But ignore new line characters
-  */
-  const matcher = /^[\r\n]?(\s+)/;
-  if (matcher.test(str)) {
-    /*
-      Replace the initial whitespace 
-      globally and over multiple lines
-    */
-    return str.replace(new RegExp('^' + str.match(matcher)![1], 'gm'), '');
-  } else {
-    // Regex doesn't match so return the original string
-    return str;
-  }
-}
-
-function parseMarkdown(text: string) {
-  const tree = fromMarkdown(trimLeadingWS(text), {
-    extensions: [gfm()],
-    mdastExtensions: [gfmFromMarkdown()],
-  });
-  return tree;
-}
-
 export function replaceGitHubUrlToMarkdown(
   text: string,
   options: ReplaceOptions,
@@ -134,27 +86,13 @@ export function replaceGitHubUrlToMarkdown(
       return true;
     }
   });
-  text = toMarkdown(tree, {
-    extensions: [gfmToMarkdown()],
-    listItemIndent: 'one',
-    rule: '-',
-  });
 
   return text;
 }
 
-export function toDingtalkMarkdown(tree: Root) {
-  return toMarkdown(tree, {
-    extensions: [gfmToMarkdown()],
-    listItemIndent: 'one',
-    fence: '`',
-    fences: true,
-  });
-}
-
 export function standardizeMarkdown(text: string) {
   const tree = parseMarkdown(text);
-  return toDingtalkMarkdown(tree);
+  return makeMarkdown(tree);
 }
 
 export async function sendContentToDing(
@@ -211,6 +149,10 @@ export async function sendToDing(
 }
 
 export function replaceGitHubText(text: string) {
+  if (!text.includes('<img')) {
+    return text;
+  }
+
   let tmp = text;
   let regexResult: RegExpExecArray | null = null;
   do {

src/utils/string-builder.ts

@@ -1,4 +1,5 @@
 import { render } from '@/github/renderer';
+import { makeMarkdown, parseMarkdown, walk } from '@/github/renderer/make-mark';
 import { replaceGitHubText } from '@/github/utils';
 
 export class StringBuilder {
@@ -50,7 +51,37 @@ export class StringBuilder {
 
 const LIMIT_MIN_LINE = 5;
 
+export function tryReplaceImageToNull(text: string) {
+  text = replaceGitHubText(text);
+
+  const ast = parseMarkdown(text);
+
+  let imageCount = 0;
+
+  walk(ast, (node) => {
+    if (node.type === 'image') {
+      imageCount++;
+      node.url = '';
+      node.alt = '';
+      node.title = '';
+      return true;
+    }
+  });
+
+  const result = makeMarkdown(ast);
+
+  return { result, imageCount };
+}
+
 export function limitTextByPosition(text: string, position: number) {
+  // replace all image url to null
+  const { result, imageCount } = tryReplaceImageToNull(text);
+
+  // if images count less than 8, we only process the non-image text
+  if (imageCount < 8) {
+    text = result;
+  }
+
   const arrayOfLines = text.replace(/\r\n|\n\r|\n|\r/g, '\n').split('\n');
 
   let count = 0;

test/github/__snapshots__/utils.test.ts.snap

@@ -1,5 +1,14 @@
 // Jest Snapshot v1, https://goo.gl/fbAQLP
 
+exports[`github utils can replace image to null 1`] = `
+{
+  "imageCount": 1,
+  "result": "![]()
+看起来这个分支有个报错关于 TerminalClient 的,有可能是 init 时机有点问题
+",
+}
+`;
+
 exports[`github utils can transform image 1`] = `
 "![](https://user-images.githubusercontent.com/2226423/153811718-2babbfa7-e63f-4ec7-9fd3-9f450beaad9b.png)
 看起来这个分支有个报错关于 TerminalClient 的,有可能是 init 时机有点问题"

test/github/utils.test.ts

@@ -6,9 +6,16 @@ import {
   parseGitHubUrl,
   standardizeMarkdown,
 } from '@/github/utils';
-import { StringBuilder, limitTextByPosition } from '@/utils/string-builder';
+import {
+  StringBuilder,
+  limitTextByPosition,
+  tryReplaceImageToNull,
+} from '@/utils/string-builder';
 import * as DingUtils from '@opensumi/dingtalk-bot/lib/utils';
 
+const commentWithImg = `<img width="954" alt="image" src="https://user-images.githubusercontent.com/2226423/153811718-2babbfa7-e63f-4ec7-9fd3-9f450beaad9b.png">
+看起来这个分支有个报错关于 TerminalClient 的,有可能是 init 时机有点问题`;
+
 describe('github utils', () => {
   it('can limit lines', () => {
     const text = `line1\nline2\nline3\n`;
@@ -17,9 +24,7 @@ describe('github utils', () => {
   });
 
   it('can transform image', () => {
-    const comment = `<img width="954" alt="image" src="https://user-images.githubusercontent.com/2226423/153811718-2babbfa7-e63f-4ec7-9fd3-9f450beaad9b.png">
-看起来这个分支有个报错关于 TerminalClient 的,有可能是 init 时机有点问题`;
-    const data = new StringBuilder(comment);
+    const data = new StringBuilder(commentWithImg);
     expect(data.build()).toMatchSnapshot();
   });
 
@@ -293,4 +298,9 @@ https://github.com/opensumi/core/blob/3455b10620badfe7b03a02d66136d3226b7891b8/p
     expect(result).toContain(`hi\\~`);
     console.log(`it ~ result:`, result);
   });
+
+  it('can replace image to null', () => {
+    const d = tryReplaceImageToNull(commentWithImg);
+    expect(d).toMatchSnapshot();
+  });
 });