Merge pull request #124 from ogajduse/pre-release-0.2.0-preparations

.github/workflows/pull_request.yml

@@ -12,6 +12,9 @@ jobs:
       matrix:
         python-version: ["3.11"]
     steps:
+      - name: Install xmllint
+        run: sudo apt -y install libxml2-utils
+
       - name: Checkout Repository
         uses: actions/checkout@v4
 

.gitignore

@@ -3,3 +3,11 @@
 build/
 *.egg-info/
 custom_components/hacs
+test_hass/.HA_VERSION
+test_hass/.storage*
+test_hass/blueprints/
+test_hass/custom_components
+test_hass/home-assistant_v2.db*
+test_hass/home-assistant.log*
+test_hass/sensors.yaml
+test_hass/www/

.pre-commit-config.yaml

@@ -8,16 +8,22 @@ repos:
       - id: end-of-file-fixer
       - id: check-yaml
         exclude: ^test_hass/configuration.yaml$
+      - id: check-toml
       - id: debug-statements
-  - repo: https://github.com/asottile/pyupgrade
-    rev: v3.3.0
+  - repo: https://github.com/lsst-ts/pre-commit-xmllint
+    rev: v1.0.0
     hooks:
-      - id: pyupgrade
-        args: [--py311-plus]
-  - repo: https://github.com/psf/black
-    rev: "22.10.0"
+      - id: format-xmllint
+  - repo: https://github.com/macisamuele/language-formatters-pre-commit-hooks
+    rev: v2.12.0
     hooks:
-      - id: black
+      - id: pretty-format-toml
+        args: ["--autofix", "--no-sort"]
+  - repo: https://github.com/abravalheri/validate-pyproject
+    rev: v0.16
+    hooks:
+      - id: validate-pyproject
+        additional_dependencies: ["validate-pyproject-schema-store[all]"]
   - repo: https://github.com/pre-commit/mirrors-mypy
     rev: "v1.4.1"
     hooks:
@@ -30,6 +36,10 @@ repos:
             types-PyYAML,
             types-requests,
           ]
+  - repo: https://github.com/psf/black
+    rev: "22.10.0"
+    hooks:
+      - id: black
   - repo: https://github.com/astral-sh/ruff-pre-commit
     # Ruff version.
     rev: v0.0.278

.vscode/launch.json

@@ -0,0 +1,20 @@
+{
+    // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
+    "version": "0.2.0",
+    "configurations": [
+        {
+            // Example of attaching to local debug server
+            "name": "Python: Attach Local",
+            "type": "python",
+            "request": "attach",
+            "port": 5678,
+            "host": "localhost",
+            "pathMappings": [
+                {
+                    "localRoot": "${workspaceFolder}",
+                    "remoteRoot": "."
+                }
+            ],
+        }
+    ]
+}

.vscode/settings.json

@@ -1,3 +1,8 @@
 {
-    "python.formatting.provider": "black"
+    "python.formatting.provider": "black",
+    "python.testing.pytestArgs": [
+        "tests"
+    ],
+    "python.testing.unittestEnabled": false,
+    "python.testing.pytestEnabled": true
 }

README.md

@@ -15,29 +15,54 @@ Hey dude! Help me out for a couple of :beers: or a :coffee:!
 
 [![coffee](https://www.buymeacoffee.com/assets/img/custom_images/black_img.png)](https://www.buymeacoffee.com/zJtVxUAgH)
 
-To get started put `/custom_components/feedparser/` here:
-`<config directory>/custom_components/feedparser/`
+
+## Installation
+[![hacs_badge](https://img.shields.io/badge/HACS-Custom-41BDF5.svg)](https://github.com/hacs/integration)
+
+1. Open HACS Settings and add this repository (https://github.com/custom-components/feedparser/)
+   as a Custom Repository (use **Integration** as the category).
+2. The `feedparser` page should automatically load (or find it in the HACS Store)
+3. Click `Install`
+
+Alternatively, click on the button below to add the repository:
+
+[![Open your Home Assistant instance and open a repository inside the Home Assistant Community Store.](https://my.home-assistant.io/badges/hacs_repository.svg)](https://my.home-assistant.io/redirect/hacs_repository/?category=Integration&repository=feedparser&owner=custom-components)
+
+
+## Configuration
 
 **Example configuration.yaml:**
 
 ```yaml
 sensor:
-  platform: feedparser
-  name: Engineering Feed
-  feed_url: 'https://www.sciencedaily.com/rss/matter_energy/engineering.xml'
-  date_format: '%a, %d %b %Y %H:%M:%S %Z'
-  scan_interval:
-    hours: 3
-  inclusions:
-    - title
-    - link
-    - description
-    - image
-    - pubDate
-  exclusions:
-    - language
+  - platform: feedparser
+    name: Engineering Feed
+    feed_url: 'https://www.sciencedaily.com/rss/matter_energy/engineering.xml'
+    date_format: '%a, %d %b %Y %H:%M:%S %Z'
+    scan_interval:
+      hours: 3
+    inclusions:
+      - title
+      - link
+      - description
+      - image
+      - published
+    exclusions:
+      - language
+
+  # Configuration of the second sensor tracking a different RSS feed
+  - platform: feedparser
+    name: Algemeen
+    feed_url: https://www.nu.nl/rss/Algemeen
+    local_time: true
+    show_topn: 1
 ```
 
+If you wish the integration to look for enclosures in the feed entries, add `image` to `inclusions` list. Do not use `enclosure`.
+The integration tries to get the link to an image for the given feed item and stores it under the attribute named `image`. If it fails to find it, it assigns the Home Assistant logo to it instead.
+
+Note that the original `pubDate` field is available under `published` attribute for the given feed entry. Other date-type values that can be available are `updated`, `created` and `expired`. Please refer to [the documentation of the original feedparser](https://feedparser.readthedocs.io/en/latest/date-parsing.html) library.
+
 **Configuration variables:**
 
 key | description
@@ -65,6 +90,6 @@ Due to how `custom_components` are loaded, it is normal to see a `ModuleNotFound
 [forum-shield]: https://img.shields.io/badge/community-forum-brightgreen.svg?style=for-the-badge
 [forum]: https://community.home-assistant.io/t/custom-component-rss-feed-parser/64637
 [license-shield]: https://img.shields.io/github/license/custom-components/feedparser.svg?style=for-the-badge
-[maintenance-shield]: https://img.shields.io/badge/maintainer-Ian%20Richardson%20%40iantrich-blue.svg?style=for-the-badge
+[maintenance-shield]: https://img.shields.io/badge/maintainer-Ondrej%20Gajdusek%20%40ogajduse-blue.svg?style=for-the-badge
 [releases-shield]: https://img.shields.io/github/release/custom-components/feedparser.svg?style=for-the-badge
 [releases]: https://github.com/custom-components/feedparser/releases

custom_components/feedparser/manifest.json

@@ -6,6 +6,6 @@
   "documentation": "https://github.com/custom-components/feedparser/blob/master/README.md",
   "iot_class": "cloud_polling",
   "issue_tracker": "https://github.com/custom-components/feedparser/issues",
-  "requirements": ["feedparser==6.0.10", "python-dateutil"],
+  "requirements": ["feedparser==6.0.11", "python-dateutil", "requests-file", "requests"],
   "version": "0.1.11"
 }

custom_components/feedparser/sensor.py

@@ -35,12 +35,14 @@
 CONF_INCLUSIONS = "inclusions"
 CONF_EXCLUSIONS = "exclusions"
 CONF_SHOW_TOPN = "show_topn"
+CONF_REMOVE_SUMMARY_IMG = "remove_summary_image"
 
 DEFAULT_DATE_FORMAT = "%a, %b %d %I:%M %p"
 DEFAULT_SCAN_INTERVAL = timedelta(hours=1)
 DEFAULT_THUMBNAIL = "https://www.home-assistant.io/images/favicon-192x192-full.png"
 DEFAULT_TOPN = 9999
 USER_AGENT = f"Home Assistant Feed-parser Integration {__version__}"
+IMAGE_REGEX = r"<img.+?src=\"(.+?)\".+?>"
 
 PLATFORM_SCHEMA = PLATFORM_SCHEMA.extend(
     {
@@ -49,6 +51,7 @@
         vol.Required(CONF_DATE_FORMAT, default=DEFAULT_DATE_FORMAT): cv.string,
         vol.Optional(CONF_LOCAL_TIME, default=False): cv.boolean,
         vol.Optional(CONF_SHOW_TOPN, default=DEFAULT_TOPN): cv.positive_int,
+        vol.Optional(CONF_REMOVE_SUMMARY_IMG, default=False): cv.boolean,
         vol.Optional(CONF_INCLUSIONS, default=[]): vol.All(cv.ensure_list, [cv.string]),
         vol.Optional(CONF_EXCLUSIONS, default=[]): vol.All(cv.ensure_list, [cv.string]),
         vol.Optional(CONF_SCAN_INTERVAL, default=DEFAULT_SCAN_INTERVAL): cv.time_period,
@@ -72,6 +75,7 @@ async def async_setup_platform(
                 name=config[CONF_NAME],
                 date_format=config[CONF_DATE_FORMAT],
                 show_topn=config[CONF_SHOW_TOPN],
+                remove_summary_image=config[CONF_REMOVE_SUMMARY_IMG],
                 inclusions=config[CONF_INCLUSIONS],
                 exclusions=config[CONF_EXCLUSIONS],
                 scan_interval=config[CONF_SCAN_INTERVAL],
@@ -95,6 +99,7 @@ def __init__(
         name: str,
         date_format: str,
         show_topn: int,
+        remove_summary_image: bool,
         exclusions: list[str | None],
         inclusions: list[str | None],
         scan_interval: timedelta,
@@ -106,6 +111,7 @@ def __init__(
         self._attr_icon = "mdi:rss"
         self._date_format = date_format
         self._show_topn: int = show_topn
+        self._remove_summary_image = remove_summary_image
         self._inclusions = inclusions
         self._exclusions = exclusions
         self._scan_interval = scan_interval
@@ -119,7 +125,9 @@ def __repr__(self: FeedParserSensor) -> str:
         """Return the representation."""
         return (
             f'FeedParserSensor(name="{self.name}", feed="{self._feed}", '
-            f"show_topn={self._show_topn}, inclusions={self._inclusions}, "
+            f"show_topn={self._show_topn}, "
+            f"remove_summary_image={self._remove_summary_image}, "
+            f"inclusions={self._inclusions}, "
             f"exclusions={self._exclusions}, scan_interval={self._scan_interval}, "
             f'local_time={self._local_time}, date_format="{self._date_format}")'
         )
@@ -191,14 +199,20 @@ def _generate_sensor_entry(
             else:
                 sensor_entry[key] = value
 
-            if "image" in self._inclusions and "image" not in sensor_entry:
-                sensor_entry["image"] = self._process_image(feed_entry)
-            if (
-                "link" in self._inclusions
-                and "link" not in sensor_entry
-                and (processed_link := self._process_link(feed_entry))
-            ):
-                sensor_entry["link"] = processed_link
+        if "image" in self._inclusions and "image" not in sensor_entry:
+            sensor_entry["image"] = self._process_image(feed_entry)
+        if (
+            "link" in self._inclusions
+            and "link" not in sensor_entry
+            and (processed_link := self._process_link(feed_entry))
+        ):
+            sensor_entry["link"] = processed_link
+        if self._remove_summary_image and "summary" in sensor_entry:
+            sensor_entry["summary"] = re.sub(
+                IMAGE_REGEX,
+                "",
+                sensor_entry["summary"],
+            )
         _LOGGER.debug("Feed %s: Generated sensor entry: %s", self.name, sensor_entry)
         return sensor_entry
 
@@ -239,7 +253,7 @@ def _parse_date(self: FeedParserSensor, date: str) -> datetime:
         return parsed_time
 
     def _process_image(self: FeedParserSensor, feed_entry: FeedParserDict) -> str:
-        if "enclosures" in feed_entry and feed_entry["enclosures"]:
+        if feed_entry.get("enclosures"):
             images = [
                 enc for enc in feed_entry["enclosures"] if enc.type.startswith("image/")
             ]
@@ -248,7 +262,7 @@ def _process_image(self: FeedParserSensor, feed_entry: FeedParserDict) -> str:
                 return images[0]["href"]
         elif "summary" in feed_entry:
             images = re.findall(
-                r"<img.+?src=\"(.+?)\".+?>",
+                IMAGE_REGEX,
                 feed_entry["summary"],
             )
             if images:
@@ -265,7 +279,7 @@ def _process_link(self: FeedParserSensor, feed_entry: FeedParserDict) -> str:
         """Return link from feed entry."""
         if "links" in feed_entry:
             if len(feed_entry["links"]) > 1:
-                _LOGGER.warning(
+                _LOGGER.debug(
                     "Feed %s: More than one link found for %s. Using the first link.",
                     self.name,
                     feed_entry,