[refactor] Extract duplicated arrange to make DRY

tests/test_show_external_link_icon.py

@@ -1,16 +1,12 @@
 import shutil
 from pathlib import Path
 
+import pytest
 from bs4 import BeautifulSoup
 
 
-def assert_is_external(reference, expected_url: str) -> None:
-    assert reference["href"] == expected_url
-    assert reference["target"] == "_blank"
-    assert reference["rel"] == ["noopener", "noreferrer"]
-
-
-def test_see_external_link_icon(
+@pytest.fixture
+def parsed_built_html(
     make_app,
     sphinx_test_tempdir: Path,
     rootdir: Path,
@@ -24,30 +20,27 @@ def test_see_external_link_icon(
     app.build()
 
     html = (app.outdir / "index.html").read_text()
-    soup = BeautifulSoup(html, "html.parser")
-    references = soup.find_all("a", {"class": "reference"})
+    return BeautifulSoup(html, "html.parser")
+
+
+def assert_is_external(reference, expected_url: str) -> None:
+    assert reference["href"] == expected_url
+    assert reference["target"] == "_blank"
+    assert reference["rel"] == ["noopener", "noreferrer"]
+
+
+def test_see_external_link_icon(parsed_built_html):
+    references = parsed_built_html.find_all("a", {"class": "reference"})
 
     ref = references[0]
     assert ref.text == "https://pypi.org/project/sphinx-new-tab-link/ "
     assert ref.svg
     assert_is_external(ref, "https://pypi.org/project/sphinx-new-tab-link/")
 
 
-def test_external_link_icon_as_image_target(
-    make_app, sphinx_test_tempdir: Path, rootdir: Path
-):
+def test_external_link_icon_as_image_target(parsed_built_html):
     # https://github.com/ftnext/sphinx-new-tab-link/issues/16
-    srcdir = sphinx_test_tempdir / "external-link-icon"
-    if not srcdir.exists():
-        testroot_path = rootdir / "test-external-link-icon"
-        shutil.copytree(testroot_path, srcdir)
-
-    app = make_app("html", srcdir=srcdir)
-    app.build()
-
-    html = (app.outdir / "index.html").read_text()
-    soup = BeautifulSoup(html, "html.parser")
-    references = soup.find_all("a", {"class": "reference"})
+    references = parsed_built_html.find_all("a", {"class": "reference"})
 
     ref = references[1]
     assert_is_external(