Feature: add http caching mechanism (#309)

This PR aims to improve plugin's performance regarding HTTP requests
which are performed to retrieve remote image length:

- add a new option `cache_dir`
- rely on https://pypi.org/project/CacheControl/ to manage local cache
requests

For now, it works only for GET requests. See upstream issue:
psf/cachecontrol#337

docs/configuration.md

@@ -195,6 +195,28 @@ Default: `<!-- more -->`
 
 ----
 
+### :material-recycle: `cache_dir`: folder where to store plugin's cached files { #cache_dir }
+
+The plugin implements a caching mechanism, ensuring that a remote media is only get once during its life-cycle on remote HTTP server (using [Cache Control](https://pypi.org/project/CacheControl/) under the hood). It is normally not necessary to specify this setting, except for when you want to change the path within your root directory where HTTP body and metadata files are cached.
+
+If you want to change it, use:
+
+``` yaml
+plugins:
+  - rss:
+      cache_dir: my/custom/dir
+```
+
+It's strongly recommended to add the path to your `.gitignore` file in the root of your project:
+
+``` title=".gitignore"
+.cache
+```
+
+Default: `.cache/plugins/rss`.
+
+----
+
 ### :material-tag-multiple: `categories`: item categories { #categories }
 
 `categories`: list of page metadata values to use as [RSS item categories](https://www.w3schools.com/xml/rss_tag_category_item.asp).

mkdocs_rss_plugin/config.py

@@ -8,6 +8,9 @@
 from mkdocs.config import config_options
 from mkdocs.config.base import Config
 
+# package
+from mkdocs_rss_plugin.constants import DEFAULT_CACHE_FOLDER
+
 # ############################################################################
 # ########## Classes ###############
 # ##################################
@@ -42,6 +45,7 @@ class RssPluginConfig(Config):
     categories = config_options.Optional(
         config_options.ListOfItems(config_options.Type(str))
     )
+    cache_dir = config_options.Type(str, default=f"{DEFAULT_CACHE_FOLDER.resolve()}")
     comments_path = config_options.Optional(config_options.Type(str))
     date_from_meta = config_options.SubConfig(_DateFromMeta)
     enabled = config_options.Type(bool, default=True)

mkdocs_rss_plugin/constants.py

@@ -14,6 +14,7 @@
 # ########## Globals #############
 # ################################
 
+DEFAULT_CACHE_FOLDER = Path(".cache/plugins/rss")
 DEFAULT_TEMPLATE_FOLDER = Path(__file__).parent / "templates"
 DEFAULT_TEMPLATE_FILENAME = DEFAULT_TEMPLATE_FOLDER / "rss.xml.jinja2"
 MKDOCS_LOGGER_NAME = "[RSS-plugin]"

mkdocs_rss_plugin/plugin.py

@@ -92,6 +92,11 @@ def on_config(self, config: MkDocsConfig) -> MkDocsConfig:
             self.config.enabled = False
             return config
 
+        # cache dir
+        self.cache_dir = Path(self.config.cache_dir)
+        self.cache_dir.mkdir(parents=True, exist_ok=True)
+        logger.debug(f"Caching HTTP requests to: {self.cache_dir.resolve()}")
+
         # integrations - check if theme is Material and if social cards are enabled
         self.integration_material_social_cards = IntegrationMaterialSocialCards(
             mkdocs_config=config,
@@ -100,6 +105,7 @@ def on_config(self, config: MkDocsConfig) -> MkDocsConfig:
 
         # instantiate plugin tooling
         self.util = Util(
+            cache_dir=self.cache_dir,
             use_git=self.config.use_git,
             integration_material_social_cards=self.integration_material_social_cards,
         )
@@ -169,10 +175,6 @@ def on_config(self, config: MkDocsConfig) -> MkDocsConfig:
             self.config.date_from_meta.default_time = datetime.strptime(
                 self.config.date_from_meta.default_time, "%H:%M"
             )
-            print(
-                self.config.date_from_meta.default_time,
-                type(self.config.date_from_meta.default_time),
-            )
         except (TypeError, ValueError) as err:
             logger.warning(
                 "Config error: `date_from_meta.default_time` value "

mkdocs_rss_plugin/util.py

@@ -19,6 +19,8 @@
 # 3rd party
 import markdown
 import urllib3
+from cachecontrol import CacheControl
+from cachecontrol.caches.file_cache import SeparateBodyFileCache
 from git import (
     GitCommandError,
     GitCommandNotFound,
@@ -34,7 +36,11 @@
 from requests.exceptions import ConnectionError, HTTPError
 
 # package
-from mkdocs_rss_plugin.constants import MKDOCS_LOGGER_NAME, REMOTE_REQUEST_HEADERS
+from mkdocs_rss_plugin.constants import (
+    DEFAULT_CACHE_FOLDER,
+    MKDOCS_LOGGER_NAME,
+    REMOTE_REQUEST_HEADERS,
+)
 from mkdocs_rss_plugin.git_manager.ci import CiHandler
 from mkdocs_rss_plugin.integrations.theme_material_social_plugin import (
     IntegrationMaterialSocialCards,
@@ -67,6 +73,7 @@ class Util:
     def __init__(
         self,
         path: str = ".",
+        cache_dir: Path = DEFAULT_CACHE_FOLDER,
         use_git: bool = True,
         integration_material_social_cards: Optional[
             IntegrationMaterialSocialCards
@@ -122,8 +129,13 @@ def __init__(
         self.social_cards = integration_material_social_cards
 
         # http/s session
-        self.req_session = Session()
-        self.req_session.headers.update(REMOTE_REQUEST_HEADERS)
+        session = Session()
+        session.headers.update(REMOTE_REQUEST_HEADERS)
+        self.req_session = CacheControl(
+            sess=session,
+            cache=SeparateBodyFileCache(directory=cache_dir),
+            cacheable_methods=("GET", "HEAD"),
+        )
 
     def build_url(
         self, base_url: str, path: str, args_dict: Optional[dict] = None

requirements/base.txt

@@ -1,7 +1,7 @@
 # Common requirements
 # -----------------------
 
-
+cachecontrol[filecache] >=0.14,<1
 GitPython>=3.1,<3.2
 mkdocs>=1.5,<2
 requests>=2.31,<3

tests/dev/dev_cached_http.py

@@ -0,0 +1,45 @@
+import http.client
+import logging
+from pathlib import Path
+
+import requests
+from cachecontrol import CacheControl
+from cachecontrol.caches.file_cache import FileCache
+
+http.client.HTTPConnection.debuglevel = 1
+logging.basicConfig()
+logging.getLogger().setLevel(logging.DEBUG)
+req_log = logging.getLogger("requests.packages.urllib3")
+req_log.setLevel(logging.DEBUG)
+req_log.propagate = True
+
+
+sess = CacheControl(
+    requests.Session(), cache=FileCache(".web_cache"), cacheable_methods=("HEAD", "GET")
+)
+
+
+# get requests
+resp = sess.get("https://geotribu.fr")
+resp_img = sess.get(
+    "https://cdn.geotribu.fr/img/articles-blog-rdp/capture-ecran/kevish_Air-Traffic.png"
+)
+
+# try again, cache hit expected
+resp = sess.get("https://geotribu.fr")
+resp_img = sess.get(
+    "https://cdn.geotribu.fr/img/articles-blog-rdp/capture-ecran/kevish_Air-Traffic.png"
+)
+
+# head requests
+resp_img = sess.head(
+    "https://cdn.geotribu.fr/img/articles-blog-rdp/capture-ecran/kevish_Air-Traffic.png"
+)
+
+
+# try again, cache hit expected
+resp_img = sess.head(
+    "https://cdn.geotribu.fr/img/articles-blog-rdp/capture-ecran/kevish_Air-Traffic.png"
+)
+
+print(list(Path(".web_cache").iterdir()))

tests/test_config.py

@@ -21,8 +21,10 @@
 # 3rd party
 from mkdocs.config.base import Config
 
-# plugin target
 from mkdocs_rss_plugin.config import RssPluginConfig
+
+# plugin target
+from mkdocs_rss_plugin.constants import DEFAULT_CACHE_FOLDER
 from mkdocs_rss_plugin.plugin import GitRssPlugin
 
 # test suite
@@ -62,6 +64,7 @@ def test_plugin_config_defaults(self):
             "abstract_chars_count": 160,
             "abstract_delimiter": "<!-- more -->",
             "categories": None,
+            "cache_dir": f"{DEFAULT_CACHE_FOLDER.resolve()}",
             "comments_path": None,
             "date_from_meta": {
                 "as_creation": "git",
@@ -105,6 +108,7 @@ def test_plugin_config_image(self):
         expected = {
             "abstract_chars_count": 160,
             "abstract_delimiter": "<!-- more -->",
+            "cache_dir": f"{DEFAULT_CACHE_FOLDER.resolve()}",
             "categories": None,
             "comments_path": None,
             "date_from_meta": {