From 346486c563fe7201714c12034f9504034e8e00c6 Mon Sep 17 00:00:00 2001
From: Ryszard Knop <dragoon@dragonic.eu>
Date: Mon, 27 Jan 2025 12:13:47 +0100
Subject: [PATCH] Allow filtering downloaded files by glob/regex

---
 itch_dl/cli.py        | 11 +++++++++--
 itch_dl/config.py     |  3 +++
 itch_dl/downloader.py | 18 ++++++++++++++++++
 3 files changed, 30 insertions(+), 2 deletions(-)

diff --git a/itch_dl/cli.py b/itch_dl/cli.py
index a91010a..bc18b39 100644
--- a/itch_dl/cli.py
+++ b/itch_dl/cli.py
@@ -29,6 +29,10 @@ def parse_args() -> argparse.Namespace:
                         help="how many threads to use for downloading games (default: 1)")
     parser.add_argument("--mirror-web", action="store_true",
                         help="try to fetch assets on game sites")
+    parser.add_argument("--filter-files-glob", metavar="glob", default=None,
+                        help="filter downloaded files with a shell-style glob/fnmatch (unmatched files are skipped)")
+    parser.add_argument("--filter-files-regex", metavar="regex", default=None,
+                        help="filter downloaded files with a Python regex (unmatched files are skipped)")
     parser.add_argument("--verbose", action="store_true",
                         help="print verbose logs")
     return parser.parse_args()
@@ -36,8 +40,11 @@ def parse_args() -> argparse.Namespace:
 
 
 def apply_args_on_settings(args: argparse.Namespace, settings: Settings):
-    if args.api_key:
-        settings.api_key = args.api_key
+    """Apply settings overrides from provided command line arguments, if set."""
+    for key in ("api_key", "filter_files_glob", "filter_files_regex"):
+        value = getattr(args, key)
+        if value:
+            setattr(settings, key, value)
 
 
 def run() -> int:
diff --git a/itch_dl/config.py b/itch_dl/config.py
index ec35102..0f52aa2 100644
--- a/itch_dl/config.py
+++ b/itch_dl/config.py
@@ -17,6 +17,9 @@ class Settings(BaseModel):
     api_key: Optional[str] = None
     user_agent: str = f"python-requests/{requests.__version__} itch-dl/{__version__}"
 
+    filter_files_glob: Optional[str] = None
+    filter_files_regex: Optional[str] = None
+
 
 def create_and_get_config_path() -> str:
     """Returns the configuration directory in the appropriate
diff --git a/itch_dl/downloader.py b/itch_dl/downloader.py
index e95dea2..087e3f2 100644
--- a/itch_dl/downloader.py
+++ b/itch_dl/downloader.py
@@ -1,6 +1,7 @@
 import os
 import json
 import re
+import fnmatch
 import logging
 import urllib.parse
 import zipfile
@@ -65,6 +66,7 @@ class GameDownloader:
         self.download_to = download_to
         self.mirror_web = mirror_web
 
+        self.settings = settings
         self.download_keys = keys
         self.client = ItchApiClient(settings.api_key, settings.user_agent)
 
@@ -307,6 +309,22 @@ class GameDownloader:
                 expected_size = upload.get("size")
                 upload_is_external = upload["storage"] == "external"
 
+                if self.settings.filter_files_glob and not fnmatch.fnmatch(file_name, self.settings.filter_files_glob):
+                    logging.info(
+                        "File '%s' does not match the glob filter '%s', skipping",
+                        file_name,
+                        self.settings.filter_files_glob
+                    )
+                    continue
+
+                if self.settings.filter_files_regex and not re.fullmatch(self.settings.filter_files_regex, file_name):
+                    logging.info(
+                        "File '%s' does not match the regex filter '%s', skipping",
+                        file_name,
+                        self.settings.filter_files_regex
+                    )
+                    continue
+
                 logging.debug(
                     "Downloading '%s' (%d), %s",
                     file_name,