Merge pull request #6 from chazeon/feat-sanitize

Sanitize filenames to prevent invalid characters

pyproject.toml

@@ -4,7 +4,7 @@ authors = [{ name = "Chenxing Luo", email = "chenxing.luo@gmail.com" }]
 readme = "README.md"
 license = { file = "LICENSE" }
 dynamic = ["version"]
-dependencies = ["requests", "mutagen", "click", "tabulate", "wcwidth"]
+dependencies = ["requests", "mutagen", "click", "tabulate", "wcwidth", "pathvalidate"]
 requires-python = ">=3.6"
 
 [project.scripts]
@@ -17,4 +17,4 @@ Repository = "https://github.com/chazeon/python-vistopia.git"
 include = ["vistopia"]
 
 [tool.setuptools.dynamic]
-version = {attr = "vistopia.__version__"}
+version = {attr = "vistopia.__version__"}

requirements.txt

@@ -3,3 +3,4 @@ mutagen
 click
 tabulate
 wcwidth
+pathvalidate

vistopia/visitor.py

@@ -4,6 +4,7 @@
 from logging import getLogger
 from functools import lru_cache
 from typing import Optional
+from pathvalidate import sanitize_filename
 
 
 logger = getLogger(__name__)
@@ -73,7 +74,9 @@ def save_show(self, id: int,
                         int(article["sort_number"]) not in episodes:
                     continue
 
-                fname = show_dir / "{}.mp3".format(article["title"])
+                fname = show_dir / "{}.mp3".format(
+                    sanitize_filename(article["title"])
+                )
                 if not fname.exists():
                     urlretrieve(article["media_key_full_url"], fname)
 
@@ -99,7 +102,9 @@ def save_transcript(self, id: int, episodes: Optional[set] = None):
                         int(article["sort_number"]) not in episodes:
                     continue
 
-                fname = show_dir / "{}.html".format(article["title"])
+                fname = show_dir / "{}.html".format(
+                    sanitize_filename(article["title"])
+                )
                 if not fname.exists():
                     urlretrieve(article["content_url"], fname)
 
@@ -131,7 +136,9 @@ def save_transcript_with_single_file(self, id: int,
                 if episodes and int(article["sort_number"]) not in episodes:
                     continue
 
-                fname = show_dir / "{}.html".format(article["title"])
+                fname = show_dir / "{}.html".format(
+                    sanitize_filename(article["title"])
+                )
                 if not fname.exists():
                     command = [
                         single_file_exec_path,