fix: support only report the url in pr change && fix _en.rst's url

Author: ooooo-create

ci_scripts/report_preview_url.sh

@@ -11,12 +11,26 @@ generate_preview_url() {
     local file_path="$1"
     local pr_id="$2"
     local path_no_ext="${file_path%.*}"
-    local base_url="http://preview-pr-${pr_id}.paddle-docs-preview.paddlepaddle.org.cn/documentation/docs/zh/"
+
+    # Check if file ends with _en (English version)
+    # _en.rst will only reach English preview, _en.md can be reached in both English and Chinese previews
+    # To Simplify, we treat all _en files as English version
+    if [[ "$path_no_ext" == *_en ]]; then
+        local base_url="http://preview-pr-${pr_id}.paddle-docs-preview.paddlepaddle.org.cn/documentation/docs/en/"
+    else
+        # Use /zh/ path for Chinese version
+        local base_url="http://preview-pr-${pr_id}.paddle-docs-preview.paddlepaddle.org.cn/documentation/docs/zh/"
+    fi
+
     local final_url="${base_url}${path_no_ext}.html"
     echo "$final_url"
 }
 
-mapfile -t all_git_files < <(git diff --name-only --diff-filter=ACMR develop | sed 's#^docs/##')
+# Use merge-base to find the common ancestor between PR branch and develop
+# This ensures we only get changes from this PR, excluding commits merged to develop after PR creation
+BASE_COMMIT=$(git merge-base HEAD develop 2>/dev/null || echo "develop")
+
+mapfile -t all_git_files < <(git diff --name-only --diff-filter=ACMR "$BASE_COMMIT" | sed 's#^docs/##')
 
 output_lines=()