From 126c685634a1bd18eea2ca0d979a78a4f43ef552 Mon Sep 17 00:00:00 2001
From: Raphael MANSUY <raphael.mansuy.partner@decathlon.com>
Date: Sun, 24 Mar 2024 12:24:26 +0100
Subject: [PATCH] improve markdown

---
 code2prompt/language_inference.py | 83 ++++++++++++++-----------------
 code2prompt/main.py               |  6 ++-
 2 files changed, 41 insertions(+), 48 deletions(-)

diff --git a/code2prompt/language_inference.py b/code2prompt/language_inference.py
index 8c8c53b..105dcd9 100644
--- a/code2prompt/language_inference.py
+++ b/code2prompt/language_inference.py
@@ -1,4 +1,6 @@
-""" This module contains the function to infer the programming language based on the file extension. """
+"""
+This module contains the function to infer the programming language based on the file extension.
+"""
 
 import os
 
@@ -13,49 +15,36 @@ def infer_language(filename: str) -> str:
     _, extension = os.path.splitext(filename)
     extension = extension.lower()
 
-    if extension in [".c", ".h"]:
-        return "c"
-    elif extension in [".cpp", ".hpp", ".cc", ".cxx"]:
-        return "cpp"
-    elif extension in [".java"]:
-        return "java"
-    elif extension in [".js", ".jsx"]:
-        return "javascript"
-    elif extension in [".cs"]:
-        return "csharp"
-    elif extension in [".php"]:
-        return "php"
-    elif extension in [".go"]:
-        return "go"
-    elif extension in [".rs"]:
-        return "rust"
-    elif extension in [".kt"]:
-        return "kotlin"
-    elif extension in [".swift"]:
-        return "swift"
-    elif extension in [".scala"]:
-        return "scala"
-    elif extension in [".dart"]:
-        return "dart"
-    elif extension in [".py"]:
-        return "python"
-    elif extension in [".rb"]:
-        return "ruby"
-    elif extension in [".pl", ".pm"]:
-        return "perl"
-    elif extension in [".sh"]:
-        return "bash"
-    elif extension in [".ps1"]:
-        return "powershell"
-    elif extension in [".html", ".htm"]:
-        return "html"
-    elif extension in [".xml"]:
-        return "xml"
-    elif extension in [".sql"]:
-        return "sql"
-    elif extension in [".m"]:
-        return "matlab"
-    elif extension in [".r"]:
-        return "r"
-    else:
-        return "unknown"
+    language_map = {
+        ".c": "c",
+        ".h": "c",
+        ".cpp": "cpp",
+        ".hpp": "cpp",
+        ".cc": "cpp",
+        ".cxx": "cpp",
+        ".java": "java",
+        ".js": "javascript",
+        ".jsx": "javascript",
+        ".cs": "csharp",
+        ".php": "php",
+        ".go": "go",
+        ".rs": "rust",
+        ".kt": "kotlin",
+        ".swift": "swift",
+        ".scala": "scala",
+        ".dart": "dart",
+        ".py": "python",
+        ".rb": "ruby",
+        ".pl": "perl",
+        ".pm": "perl",
+        ".sh": "bash",
+        ".ps1": "powershell",
+        ".html": "html",
+        ".htm": "html",
+        ".xml": "xml",
+        ".sql": "sql",
+        ".m": "matlab",
+        ".r": "r"
+    }
+
+    return language_map.get(extension, "unknown")
\ No newline at end of file
diff --git a/code2prompt/main.py b/code2prompt/main.py
index 6388047..f39557f 100644
--- a/code2prompt/main.py
+++ b/code2prompt/main.py
@@ -152,7 +152,11 @@ def create_markdown_file(path, output, gitignore, filter, suppress_comments):
             file_info += f"- Created: {file_creation_time}\n"
             file_info += f"- Modified: {file_modification_time}\n\n"
 
-            file_code = f"### Code\n```{file_extension}\n{file_content}\n```\n\n"
+            language = infer_language(file_path.name)
+            if language == "unknown":
+                language = format(file_extension[1:])
+
+            file_code = f"### Code\n```{language}\n{file_content}\n```\n\n"
 
             content.append(file_info + file_code)
             table_of_contents.append(