From e525fd634b07600110cf0ac940a4b5e9c48f1ab1 Mon Sep 17 00:00:00 2001
From: Luka Hietala <luka.hietala08@gmail.com>
Date: Thu, 20 Nov 2025 17:16:26 +0200
Subject: [PATCH] add pygments syntax highlighting

---
 git/blob.py         | 10 ++++++++--
 highlight.py        | 25 +++++++++++++++++++++++++
 requirements.txt    |  1 +
 templates/blob.html |  3 +--
 4 files changed, 35 insertions(+), 4 deletions(-)
 create mode 100644 highlight.py

diff --git a/git/blob.py b/git/blob.py
index 96496a6..a950dae 100644
--- a/git/blob.py
+++ b/git/blob.py
@@ -1,4 +1,5 @@
 import pygit2 as git
+from highlight import highlight_code
 
 # retrieves a blob content for given ref and path
 def get_blob(repo_path, ref="HEAD", blob_path=""):
@@ -22,14 +23,19 @@ def get_blob(repo_path, ref="HEAD", blob_path=""):
                 if entry.name == part:
                     if entry.type == git.GIT_OBJECT_BLOB:
                         blob = repo.get(entry.id)
-                        return {
+                        content = blob.data.decode('utf-8', errors='replace')
+                        result = {
                             'name': entry.name,
                             'id': str(entry.id),
                             'path': blob_path,
                             'size': blob.size,
                             'is_binary': blob.is_binary,
-                            'content': blob.data.decode('utf-8', errors='replace')
+                            'content': content
                         }
+                        # add highlighted content
+                        if not blob.is_binary:
+                            result['highlighted'] = highlight_code(content, entry.name)
+                        return result
                     elif entry.type == git.GIT_OBJECT_TREE:
                         tree = repo.get(entry.id)
                         found = True
diff --git a/highlight.py b/highlight.py
new file mode 100644
index 0000000..ca5b64e
--- /dev/null
+++ b/highlight.py
@@ -0,0 +1,25 @@
+from pygments import highlight
+from pygments.util import ClassNotFound
+from pygments.lexers import TextLexer
+from pygments.lexers import guess_lexer
+from pygments.lexers import guess_lexer_for_filename
+from pygments.formatters import HtmlFormatter
+
+# server-side syntax highlighting
+
+# https://git.kernel.org/pub/scm/infra/cgit.git/tree/filters/syntax-highlighting.py?id=dbaee2672be14374acb17266477c19294c6155f3
+
+def highlight_code(data, filename):
+    formatter = HtmlFormatter(style='sas', nobackground=True, linenos=True)
+    try:
+        lexer = guess_lexer_for_filename(filename, data)
+    except ClassNotFound:
+        if data.startswith('#!'):
+            lexer = guess_lexer(data)
+        else:
+            lexer = TextLexer()
+    except TypeError:
+        lexer = TextLexer()
+    css = formatter.get_style_defs('.highlight')
+    highlighted = highlight(data, lexer, formatter)
+    return f'<style>{css}</style>{highlighted}'
\ No newline at end of file
diff --git a/requirements.txt b/requirements.txt
index 1da7394..39f0eec 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -9,5 +9,6 @@ MarkupSafe==3.0.3
 packaging==25.0
 pycparser==2.23
 pygit2==1.19.0
+Pygments==2.19.2
 python-dotenv==1.2.1
 Werkzeug==3.1.3
diff --git a/templates/blob.html b/templates/blob.html
index f75c70d..e49ee3d 100644
--- a/templates/blob.html
+++ b/templates/blob.html
@@ -5,10 +5,9 @@
 <p>Blob id: {{ blob.id }}</p>
 <p>Size: {{ blob.size }} bytes</p>
 <p><a href="{{ url_for('repo_blame_path', repo_name=repo_name, path=path) }}?ref={{ ref }}">Blame</a></p>
-<h2>Content:</h2>
 {% if blob.is_binary %}
     <pre>Binary...</pre>
 {% else %}
-<pre>{{ blob.content }}</pre>
+{{ blob.highlighted|safe }}
 {% endif %}
 {% endblock %}
\ No newline at end of file
-- 
2.47.3