From d04a6f41b30aa4eaf5251ba6d1350297a42f3065 Mon Sep 17 00:00:00 2001
From: kjimlau <neochannel@163.com>
Date: Mon, 20 Sep 2021 15:53:32 +0800
Subject: [PATCH] feat: parse dom queryPath

---
 src/content-scripts/parser/selection-meta.ts | 41 ++++++++---
 src/utils/dom.ts                             | 72 ++++++++++++++++----
 2 files changed, 90 insertions(+), 23 deletions(-)

diff --git a/src/content-scripts/parser/selection-meta.ts b/src/content-scripts/parser/selection-meta.ts
index 09cd75c..cab346d 100644
--- a/src/content-scripts/parser/selection-meta.ts
+++ b/src/content-scripts/parser/selection-meta.ts
@@ -1,5 +1,5 @@
 import { isObject } from "@/utils/utils";
-import { getNodeText } from "@/utils/dom";
+import { filterAncestorNodes, getDomQueryPath, getNodeText } from "@/utils/dom";
 import { Rect } from "@/types/common";
 
 /**
@@ -98,14 +98,37 @@ function filterInvalidCoorRects(rects: Rect[]) {
   return rects.filter((ele) => !excludeRects.includes(ele));
 }
 
+function parseSelectionRects(range: Range) {
+  let rects = Array.from(range.getClientRects()).map((r) => ({
+    x: r.x,
+    y: r.y + window.scrollY,
+    width: r.width,
+    height: r.height,
+  }));
+  rects = filterDuplicateRects(rects);
+  rects = filterInvalidCoorRects(rects);
+  return rects;
+}
+
+function parseSelectedQueryPaths(selection: Selection, range: Range) {
+  const allWithinRangeParentNodes = (range.commonAncestorContainer as HTMLElement)?.getElementsByTagName(
+    "*"
+  );
+  const allSelectedNodes = Array.from(allWithinRangeParentNodes).filter(n => selection.containsNode(n, true));
+  const textNodes = filterAncestorNodes(allSelectedNodes);
+  return textNodes.map(n => getDomQueryPath(n as HTMLElement));
+}
+
 /**
  * Get the `SelectionMeta` from current mouse selection object.
  */
 export interface SelectionMeta {
+  queryPaths: string[];
   rects: Rect[];
   texts: string[];
 }
 export function parseRectsAndTextFromSelection(): SelectionMeta {
+  let queryPaths: string[] = [];
   let rects: Rect[] = [];
   let texts: string[] = [];
   try {
@@ -114,22 +137,22 @@ export function parseRectsAndTextFromSelection(): SelectionMeta {
     if (selection) {
       const range = selection.getRangeAt(0);
       if (range) {
+        // 1. texts
         const cloneFragment = range.cloneContents();
         texts = getNodeTextList(cloneFragment);
-        rects = Array.from(range.getClientRects()).map((r) => ({
-          x: r.x,
-          y: r.y + window.scrollY,
-          width: r.width,
-          height: r.height,
-        }));
-        rects = filterDuplicateRects(rects);
-        rects = filterInvalidCoorRects(rects);
+
+        // 2. rects
+        rects = parseSelectionRects(range);
+        
+        // 3. queryPaths
+        queryPaths = parseSelectedQueryPaths(selection, range);
       }
     }
   } catch (err) {
     console.log(err);
   }
   return {
+    queryPaths,
     rects,
     texts,
   };
diff --git a/src/utils/dom.ts b/src/utils/dom.ts
index 2e9a86e..7e244fa 100644
--- a/src/utils/dom.ts
+++ b/src/utils/dom.ts
@@ -1,25 +1,69 @@
-import { getObjectType } from '../utils/utils'
+import { getObjectType } from "../utils/utils";
 
 export const getNodeText = (node: any): string => {
-  let text = ''
-  const objType = getObjectType(node)
+  let text = "";
+  const objType = getObjectType(node);
   switch (objType) {
-    case '[object HTMLTimeElement]': {
-      text = node.dateTime || ''
-      break
+    case "[object HTMLTimeElement]": {
+      text = node.dateTime || "";
+      break;
     }
-    case '[object Text]': {
-      text = node.textContent || ''
-      break
+    case "[object Text]": {
+      text = node.textContent || "";
+      break;
     }
-    case '[object HTMLSpanElement]': {
-      text = node.innerText || ''
-      break
+    case "[object HTMLSpanElement]": {
+      text = node.innerText || "";
+      break;
     }
     // TODO: more options
     default: {
-      break
+      break;
     }
   }
-  return text
+  return text;
+};
+
+export function getDomQueryPath(el: HTMLElement) {
+  if (!el) return '';
+
+  const stack = [];
+  while (el.parentNode != null) {
+    let sibCount = 0;
+    let sibIndex = 0;
+    for (let i = 0; i < el.parentNode.childNodes.length; i++) {
+      const sib = el.parentNode.childNodes[i];
+      if (sib.nodeName == el.nodeName) {
+        if (sib === el) {
+          sibIndex = sibCount;
+        }
+        sibCount++;
+      }
+    }
+    const name = el.nodeName.toLowerCase();
+    if (el.hasAttribute("id") && el.id != "") {
+      stack.unshift(`${name}#${el.id}`);
+    } else if (sibCount > 1) {
+      stack.unshift(`${name}:n-th-of-type(${sibIndex + 1})`);
+    } else {
+      stack.unshift(name);
+    }
+    el = el.parentNode as HTMLElement;
+  }
+
+  return stack.slice(1).join(' '); // removes the html element
+}
+
+export function filterAncestorNodes(nodes: Node[]) {
+  const ancestors = new Set();
+  for (const node of nodes) {
+    let parent = node.parentNode;
+    while(parent) {
+      if (nodes.includes(parent) && node.textContent) {
+        ancestors.add(parent);
+      }
+      parent = parent.parentNode;
+    }
+  }
+  return nodes.filter(n => !ancestors.has(n));
 }
\ No newline at end of file