feat: clean and clear methods for the cache

.idea/vcs.xml

@@ -2,7 +2,5 @@
 <project version="4">
   <component name="VcsDirectoryMappings">
     <mapping directory="" vcs="Git" />
-    <mapping directory="$PROJECT_DIR$/../forks/sponsorblock.py" vcs="Git" />
-    <mapping directory="$PROJECT_DIR$/../rythmbox-id3-lyrics-support" vcs="Git" />
   </component>
 </project>

requirements.txt

@@ -1,4 +1,4 @@
-requests~=2.28.1
+requests~=2.31.0
 mutagen~=1.46.0
 musicbrainzngs~=0.7.1
 jellyfish~=0.9.0
@@ -20,4 +20,5 @@ pathvalidate~=2.5.2
 guppy3~=3.1.3
 
 toml~=0.10.2
-typing_extensions~=4.7.1
+typing_extensions~=4.7.1
+responses~=0.24.1

src/music_kraken/connection/cache.py

@@ -64,6 +64,18 @@ class Cache:
         r.mkdir(exist_ok=True)
         return r
 
+    def _write_index(self, indent: int = 4):
+        _json = []
+        for c in self.cached_attributes:
+            d = c.__dict__
+            for key in self._time_fields:
+                d[key] = d[key].isoformat()
+
+            _json.append(d)
+
+        with self.index.open("w") as f:
+            f.write(json.dumps(_json, indent=indent))
+
     def _write_attribute(self, cached_attribute: CacheAttribute, write: bool = True) -> bool:
         existing_attribute: Optional[CacheAttribute] = self._id_to_attribute.get(cached_attribute.id)
         if existing_attribute is not None:
@@ -80,16 +92,7 @@ class Cache:
             self._id_to_attribute[cached_attribute.id] = cached_attribute
 
         if write:
-            _json = []
-            for c in self.cached_attributes:
-                d = c.__dict__
-                for key in self._time_fields:
-                    d[key] = d[key].isoformat()
-
-                _json.append(d)
-
-            with self.index.open("w") as f:
-                f.write(json.dumps(_json, indent=4))
+            self._write_index()
 
         return True
 
@@ -132,3 +135,59 @@ class Cache:
 
         with path.open("rb") as f:
             return f.read()
+
+    def clean(self):
+        keep = set()
+
+        for ca in self.cached_attributes.copy():
+            file = Path(self._dir, ca.module, ca.name)
+
+            if not ca.is_valid:
+                self.logger.debug(f"deleting cache {ca.id}")
+                file.unlink()
+                self.cached_attributes.remove(ca)
+                del self._id_to_attribute[ca.id]
+
+            else:
+                keep.add(file)
+
+        # iterate through every module (folder)
+        for module_path in self._dir.iterdir():
+            if not module_path.is_dir():
+                continue
+
+            # delete all files not in keep
+            for path in module_path.iterdir():
+                if path not in keep:
+                    self.logger.info(f"Deleting cache {path}")
+                    path.unlink()
+
+            # delete all empty directories
+            for path in module_path.iterdir():
+                if path.is_dir() and not list(path.iterdir()):
+                    self.logger.debug(f"Deleting cache directory {path}")
+                    path.rmdir()
+
+        self._write_index()
+
+    def clear(self):
+        """
+        delete every file in the cache directory
+        :return:
+        """
+
+        for path in self._dir.iterdir():
+            if path.is_dir():
+                for file in path.iterdir():
+                    file.unlink()
+                path.rmdir()
+            else:
+                path.unlink()
+
+        self.cached_attributes.clear()
+        self._id_to_attribute.clear()
+
+        self._write_index()
+
+    def __repr__(self):
+        return f"<Cache {self.module}>"

src/music_kraken/connection/connection.py

@@ -1,19 +1,18 @@
-import time
-from typing import List, Dict, Callable, Optional, Set
-from urllib.parse import urlparse, urlunsplit, ParseResult
 import logging
-
 import threading
+import time
+from typing import List, Dict, Optional, Set
+from urllib.parse import urlparse, urlunsplit, ParseResult
+
 import requests
 import responses
-from responses import matchers
 from tqdm import tqdm
 
-from .rotating import RotatingProxy
 from .cache import Cache
+from .rotating import RotatingProxy
+from ..objects import Target
 from ..utils.config import main_settings
 from ..utils.support_classes.download_result import DownloadResult
-from ..objects import Target
 
 
 class Connection:

src/music_kraken/pages/bandcamp.py

@@ -53,7 +53,8 @@ class Bandcamp(Page):
     def __init__(self, *args, **kwargs):
         self.connection: Connection = Connection(
             host="https://bandcamp.com/",
-            logger=self.LOGGER
+            logger=self.LOGGER,
+            module="bandcamp",
         )
 
         super().__init__(*args, **kwargs)