From 3c20d726cf94550a6cc2b4bc9059815b216aa1e9 Mon Sep 17 00:00:00 2001
From: Denis-Cosmin NUTIU <dnutiu@bitdefender.com>
Date: Thu, 1 Jun 2023 22:19:57 +0300
Subject: [PATCH] add tests for test_fix_html_tags_stripe_tag

---
 app/converter/regex_heuristics.py             |  2 +-
 app/converter/wordpress_markdown.py           |  2 +-
 .../converter/wordpress_markdown_test.py      | 36 +++++++++++++++++++
 config.yaml                                   |  6 ++--
 4 files changed, 42 insertions(+), 4 deletions(-)

diff --git a/app/converter/regex_heuristics.py b/app/converter/regex_heuristics.py
index 2d03e21..f47a352 100644
--- a/app/converter/regex_heuristics.py
+++ b/app/converter/regex_heuristics.py
@@ -1,5 +1,5 @@
 def handle_regex_heuristics(line: str) -> str:
     """
-    Manipulates a line without tags by using regex heuristics.
+    Manipulates a line by using regex heuristics.
     """
     return line
diff --git a/app/converter/wordpress_markdown.py b/app/converter/wordpress_markdown.py
index 9eba560..c9c363c 100644
--- a/app/converter/wordpress_markdown.py
+++ b/app/converter/wordpress_markdown.py
@@ -82,7 +82,7 @@ class WordpressMarkdownConverter:
             if tags:
                 fixed_tags = self.fix_html_tags(tags)
                 if fixed_tags:
-                    fixed_lines.extend(fixed_tags)
+                    fixed_lines.append("".join(fixed_tags))
 
     def convert_post_content(self, post_content: str) -> str:
         """
diff --git a/app/tests/converter/wordpress_markdown_test.py b/app/tests/converter/wordpress_markdown_test.py
index 57c46ca..1a34a5e 100644
--- a/app/tests/converter/wordpress_markdown_test.py
+++ b/app/tests/converter/wordpress_markdown_test.py
@@ -15,6 +15,9 @@ from app.tests.utils import make_fake_configurator
     ],
 )
 def test_header_rewrite_author(author_rewrite, input_header, expected_header):
+    """
+    Test the header rewrite author functionality.
+    """
     configurator = make_fake_configurator(
         "wordpress_markdown_converter",
         ConverterOptions(
@@ -35,6 +38,9 @@ def test_header_rewrite_author(author_rewrite, input_header, expected_header):
     ],
 )
 def test_header_fields_drop(header_fields_drop, input_header, expected_header):
+    """
+    Test the header fields drop functionality.
+    """
     configurator = make_fake_configurator(
         "wordpress_markdown_converter",
         ConverterOptions(
@@ -63,6 +69,36 @@ def test_header_fields_drop(header_fields_drop, input_header, expected_header):
     ],
 )
 def test_fix_html_tags_detect_youtube_links(input_lines, expected_lines):
+    """
+    Ensure fix_html_tags detects youtube links.
+    """
+    configurator = make_fake_configurator(
+        "wordpress_markdown_converter",
+        ConverterOptions(),
+    )
+    converter = WordpressMarkdownConverter(configurator)
+    assert converter.fix_html_tags(input_lines) == expected_lines
+
+
+@pytest.mark.parametrize(
+    "input_lines, expected_lines",
+    [
+        ([], []),
+        ([""], ["\n"]),
+        (["<p>Test</p>"], ["Test"]),
+        (["<p>Test</p>", "<p>Test</p>"], ["Test", "Test"]),
+        (["<p>Te<span>st</span></p>"], ["Test"]),
+        (["<p>Te<span><span>s</span><span class='hi'>t</span></span></p>"], ["Test"]),
+        (
+            ["<pre>preTe<ImportantC#Tag>st</ImportantC#Tag></pre></pre>"],
+            ["preTest"],
+        ),
+    ],
+)
+def test_fix_html_tags_stripe_tag(input_lines, expected_lines):
+    """
+    Ensure fix_html_tags removes the tag and tested tags.
+    """
     configurator = make_fake_configurator(
         "wordpress_markdown_converter",
         ConverterOptions(),
diff --git a/config.yaml b/config.yaml
index dd8be8b..0d468b5 100644
--- a/config.yaml
+++ b/config.yaml
@@ -1,6 +1,6 @@
 logging_level: "INFO"
-source_path: "/home/denis/PycharmProjects/jekyll-to-hugo/my_test_data/_posts"
-output_path: "/home/denis/PycharmProjects/jekyll-to-hugo/my_test_data/quickstart/content/posts"
+source_path: "/Users/dnutiu/PycharmProjects/jekyll-to-hugo/my_test_data/_posts"
+output_path: "/Users/dnutiu/NucuLabsProjects/NucuLabsDevBlog/content/posts"
 converter: "wordpress_markdown_converter"
 converter_options:
   author_rewrite: "NucuLabs.dev"
@@ -9,6 +9,8 @@ converter_options:
       target: "/"
     - source: "https://nuculabs.wordpress.com/"
       target: "https://nuculabs.dev/posts/"
+    - source: "https://twitter.com/metonymyqt"
+      target: "https://twitter.com/nuculabs"
   header_fields_drop:
     - restapi_import_id
     - original_post_id