Merge pull request #265 from cnblogs/extract-wikilink-images

feat: extract wikilink images

Author: cnblogs-dudu

src/service/extract-img/apply-replace-list.ts

@@ -1,18 +1,30 @@
 import { ImgInfo } from '@/service/extract-img/get-replace-list'
 import { RsText } from '@/wasm'
+import { FILTER_BYTE_OFFSET } from './find-img-link'
+import { escapeRegExp } from 'lodash'
 
 export function applyReplaceList(
     text: string,
     replaceList: [src: ImgInfo, newLink: string][],
     beforeEach: (newLink: string) => void
 ) {
+    const rsCandidate = replaceList.filter(x => x[0].byteOffset !== FILTER_BYTE_OFFSET)
+
     // replace from end
-    const sorted = replaceList.sort((a, b) => b[0].byteOffset - a[0].byteOffset)
+    const sorted = rsCandidate.sort((a, b) => b[0].byteOffset - a[0].byteOffset)
     for (const [src, newLink] of sorted) {
         beforeEach(newLink)
         const start = src.byteOffset
         const end = src.byteOffset + Buffer.from(src.data).length
         text = RsText.replaceWithByteOffset(text, start, end, newLink)
     }
+
+    const tsCandidate = replaceList.filter(x => x[0].byteOffset === FILTER_BYTE_OFFSET)
+    for (const [src, newLink] of tsCandidate) {
+        const prefix = src.prefix ?? ''
+        const regex = new RegExp(escapeRegExp(String.raw`${prefix}${src.data}`), 'g')
+        text = text.replace(regex, prefix + newLink)
+    }
+
     return text
 }

src/service/extract-img/extract-img.ts

@@ -68,6 +68,7 @@ export async function extractImg(text: string, fileDir: string, inputImgSrc?: Im
                 })
             )
     )
+
     const replaceListLen = replaceList.length
 
     return window.withProgress(

src/service/extract-img/find-img-link.ts

@@ -13,7 +13,23 @@ const imgTagDataUrlImgPat = r`(<img.*?src\s*=\s*")(${dataUrlPat})"[^/]*?\/?>`
 const mkdUrlImgPat = r`(!\[[^]]*\]\()([^)]+)\)`
 const imgTagUrlImgPat = r`(<img\s*.*?src\s*=\s*["'])(.*?)["'][^>]*?>`
 const mkdDataUrlImgPat = r`(!\[.*?]\()(${dataUrlPat})\)`
-const cnbDomain = r`\.cnblogs\.com\/`
+const wikilinkImages = /!\[(\[.+?\])\][\s\S]+?(?<prefix>\1:\s*)(?<uri>.*?)\s+/g
+const exludeDomains = /\.cnblogs\.com/i
+const webUrlPrefix = /^https?:\/\//i
+
+export const FILTER_BYTE_OFFSET = -9999
+
+function getImagesWithTs(text: string) {
+    return [...text.matchAll(wikilinkImages)].map(m => {
+        const uri = m.groups?.uri ?? ''
+        return <ImgInfo>{
+            byteOffset: FILTER_BYTE_OFFSET,
+            data: uri,
+            src: webUrlPrefix.test(uri) ? ImgSrc.web : ImgSrc.fs,
+            prefix: m.groups?.prefix,
+        }
+    })
+}
 
 export function findImgLink(text: string): ImgInfo[] {
     const imgTagUrlImgMgs = RsRegex.matches(imgTagUrlImgPat, text) as RsMatch[]
@@ -24,7 +40,7 @@ export function findImgLink(text: string): ImgInfo[] {
         const byteOffset = mg.byte_offset + Buffer.from(prefix).length
 
         let src
-        if (/https?:\/\//.test(data)) src = ImgSrc.web
+        if (webUrlPrefix.test(data)) src = ImgSrc.web
         else src = ImgSrc.fs
 
         return <ImgInfo>{
@@ -48,8 +64,7 @@ export function findImgLink(text: string): ImgInfo[] {
         }
     })
 
-    const acc = urlImgInfo.concat(dataUrlImgInfo)
-
-    // keep links while not cnb
-    return acc.filter(x => !RsRegex.isMatch(cnbDomain, x.data.toLowerCase()))
+    let images = urlImgInfo.concat(dataUrlImgInfo)
+    images = images.concat(getImagesWithTs(text))
+    return images.filter(x => !exludeDomains.test(x.data))
 }

src/service/extract-img/get-replace-list.ts

@@ -17,6 +17,7 @@ export type ImgInfo = {
     byteOffset: number
     data: string
     src: ImgSrc
+    prefix: string | undefined
 }
 
 export const enum ImgSrc {