From 7921839f089e0185d62df5beb822d07cf1ba2fb6 Mon Sep 17 00:00:00 2001
From: Kirill Markin <markinkirill@gmail.com>
Date: Sun, 9 Jun 2024 09:38:27 +0200
Subject: [PATCH] ignore-content setting

---
 .repo-to-text-settings.yaml | 12 ++++++++++++
 repo_to_text/main.py        | 31 +++++++++++++++++++++++++------
 requirements.txt            |  1 +
 3 files changed, 38 insertions(+), 6 deletions(-)
 create mode 100644 .repo-to-text-settings.yaml

diff --git a/.repo-to-text-settings.yaml b/.repo-to-text-settings.yaml
new file mode 100644
index 0000000..0b1d0e1
--- /dev/null
+++ b/.repo-to-text-settings.yaml
@@ -0,0 +1,12 @@
+# Details: https://github.com/kirill-markin/repo-to-text
+
+# Ignore files and directories for "Contents of ..." section
+# Syntax: gitignore
+ignore-content:
+  - ".repo-to-text-settings.yaml"
+  - "README.md"
+  - "LICENSE"
+  - "examples/"
+  - "tests/"
+  - "MANIFEST.in"
+  - "setup.py"
diff --git a/repo_to_text/main.py b/repo_to_text/main.py
index a275276..82295ef 100644
--- a/repo_to_text/main.py
+++ b/repo_to_text/main.py
@@ -3,6 +3,7 @@ import subprocess
 import pathspec
 import logging
 import argparse
+import yaml
 from datetime import datetime
 import pyperclip
 
@@ -33,14 +34,32 @@ def get_tree_structure(path='.', gitignore_spec=None) -> str:
     logging.debug('Tree structure filtering complete')
     return '\n'.join(filtered_lines)
 
-def load_gitignore(path='.'):
+def load_ignore_specs(path='.'):
+    gitignore_spec = None
+    content_ignore_spec = None
+
     gitignore_path = os.path.join(path, '.gitignore')
     if os.path.exists(gitignore_path):
         logging.debug(f'Loading .gitignore from path: {gitignore_path}')
         with open(gitignore_path, 'r') as f:
-            return pathspec.PathSpec.from_lines('gitwildmatch', f)
-    logging.debug('.gitignore not found')
-    return None
+            gitignore_spec = pathspec.PathSpec.from_lines('gitwildmatch', f)
+
+    repo_settings_path = os.path.join(path, '.repo-to-text-settings.yaml')
+    if os.path.exists(repo_settings_path):
+        logging.debug(f'Loading .repo-to-text-settings.yaml from path: {repo_settings_path}')
+        with open(repo_settings_path, 'r') as f:
+            ignore_data = yaml.safe_load(f)
+            if 'ignore-content' in ignore_data:
+                content_ignore_spec = pathspec.PathSpec.from_lines('gitwildmatch', ignore_data['ignore-content'])
+
+    return gitignore_spec, content_ignore_spec
+
+def should_ignore_file(file_path, relative_path, gitignore_spec, content_ignore_spec):
+    return (
+        is_ignored_path(file_path) or
+        (gitignore_spec and gitignore_spec.match_file(relative_path)) or
+        (content_ignore_spec and content_ignore_spec.match_file(relative_path))
+    )
 
 def is_ignored_path(file_path: str) -> bool:
     ignored_dirs = ['.git']
@@ -83,7 +102,7 @@ def remove_empty_dirs(tree_output: str, path='.') -> str:
 
 def save_repo_to_text(path='.', output_dir=None) -> str:
     logging.debug(f'Starting to save repo structure to text for path: {path}')
-    gitignore_spec = load_gitignore(path)
+    gitignore_spec, content_ignore_spec = load_ignore_specs(path)
     tree_structure = get_tree_structure(path, gitignore_spec)
     tree_structure = remove_empty_dirs(tree_structure, path)
     
@@ -115,7 +134,7 @@ def save_repo_to_text(path='.', output_dir=None) -> str:
                 file_path = os.path.join(root, filename)
                 relative_path = os.path.relpath(file_path, path)
                 
-                if is_ignored_path(file_path) or (gitignore_spec and gitignore_spec.match_file(relative_path)):
+                if should_ignore_file(file_path, relative_path, gitignore_spec, content_ignore_spec):
                     continue
 
                 relative_path = relative_path.replace('./', '', 1)
diff --git a/requirements.txt b/requirements.txt
index fcd830c..1d9e36e 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -3,3 +3,4 @@ pathspec==0.12.1
 pytest==8.2.2
 argparse==1.4.0
 pyperclip==1.8.2
+PyYAML==6.0.1