Initial commit

env/lib/python3.10/site-packages/wagtail/embeds/finders/__init__.py

@@ -0,0 +1,48 @@
+from functools import lru_cache
+from importlib import import_module
+
+from django.conf import settings
+from django.utils.module_loading import import_string
+
+
+def import_finder_class(dotted_path):
+    """
+    Imports a finder class from a dotted path. If the dotted path points to a
+    module, that module is imported and its "embed_finder_class" class returned.
+
+    If not, this will assume the dotted path points to directly a class and
+    will attempt to import that instead.
+    """
+    try:
+        finder_module = import_module(dotted_path)
+        return finder_module.embed_finder_class
+    except ImportError as e:
+        try:
+            return import_string(dotted_path)
+        except ImportError:
+            raise ImportError from e
+
+
+def _get_config_from_settings():
+    if hasattr(settings, "WAGTAILEMBEDS_FINDERS"):
+        return settings.WAGTAILEMBEDS_FINDERS
+    else:
+        # Default to the oembed backend
+        return [
+            {
+                "class": "wagtail.embeds.finders.oembed",
+            }
+        ]
+
+
+@lru_cache(maxsize=None)
+def get_finders():
+    finders = []
+
+    for finder_config in _get_config_from_settings():
+        finder_config = finder_config.copy()
+        cls = import_finder_class(finder_config.pop("class"))
+
+        finders.append(cls(**finder_config))
+
+    return finders

env/lib/python3.10/site-packages/wagtail/embeds/finders/base.py

@@ -0,0 +1,6 @@
+class EmbedFinder:
+    def accept(self, url):
+        return False
+
+    def find_embed(self, url, max_width=None, max_height=None):
+        raise NotImplementedError

env/lib/python3.10/site-packages/wagtail/embeds/finders/embedly.py

@@ -0,0 +1,76 @@
+from django.utils.html import format_html
+
+from wagtail.embeds.exceptions import EmbedException, EmbedNotFoundException
+
+from .base import EmbedFinder
+
+
+class EmbedlyException(EmbedException):
+    pass
+
+
+class AccessDeniedEmbedlyException(EmbedlyException):
+    pass
+
+
+class EmbedlyFinder(EmbedFinder):
+    key = None
+
+    def __init__(self, key=None):
+        if key:
+            self.key = key
+
+    def get_key(self):
+        return self.key
+
+    def accept(self, url):
+        # We don't really know what embedly supports so accept everything
+        return True
+
+    def find_embed(self, url, max_width=None, key=None):
+        from embedly import Embedly
+
+        # Get embedly key
+        if key is None:
+            key = self.get_key()
+
+        # Get embedly client
+        client = Embedly(key=key)
+
+        # Call embedly
+        if max_width is not None:
+            oembed = client.oembed(url, maxwidth=max_width, better=False)
+        else:
+            oembed = client.oembed(url, better=False)
+
+        # Check for error
+        if oembed.get("error"):
+            if oembed["error_code"] in [401, 403]:
+                raise AccessDeniedEmbedlyException
+            elif oembed["error_code"] == 404:
+                raise EmbedNotFoundException
+            else:
+                raise EmbedlyException
+
+        # Convert photos into HTML
+        if oembed["type"] == "photo":
+            html = format_html('<img src="{}" alt="">', oembed["url"])
+        else:
+            html = oembed.get("html")
+
+        # Return embed as a dict
+        return {
+            "title": oembed["title"] if "title" in oembed else "",
+            "author_name": oembed["author_name"] if "author_name" in oembed else "",
+            "provider_name": oembed["provider_name"]
+            if "provider_name" in oembed
+            else "",
+            "type": oembed["type"],
+            "thumbnail_url": oembed.get("thumbnail_url"),
+            "width": oembed.get("width"),
+            "height": oembed.get("height"),
+            "html": html,
+        }
+
+
+embed_finder_class = EmbedlyFinder

env/lib/python3.10/site-packages/wagtail/embeds/finders/facebook.py

@@ -0,0 +1,103 @@
+import json
+from urllib import request as urllib_request
+from urllib.error import HTTPError, URLError
+from urllib.parse import urlencode
+from urllib.request import Request
+
+from wagtail.embeds.exceptions import EmbedException, EmbedNotFoundException
+
+from .oembed import OEmbedFinder
+
+
+class AccessDeniedFacebookOEmbedException(EmbedException):
+    pass
+
+
+FACEBOOK_PROVIDERS = [
+    # Videos
+    {
+        "endpoint": "https://graph.facebook.com/v11.0/oembed_video",
+        "urls": [
+            r"^https://(?:www\.)?facebook\.com/.+?/videos/.+$",
+            r"^https://(?:www\.)?facebook\.com/video\.php\?(?:v|id)=.+$",
+            r"^https://fb.watch/.+$",
+        ],
+    },
+    # Posts
+    {
+        "endpoint": "https://graph.facebook.com/v11.0/oembed_post",
+        "urls": [
+            r"^https://(?:www\.)?facebook\.com/.+?/(?:posts|activity)/.+$",
+            r"^https://(?:www\.)?facebook\.com/photo\.php\?fbid=.+$",
+            r"^https://(?:www\.)?facebook\.com/(?:photos|questions)/.+$",
+            r"^https://(?:www\.)?facebook\.com/permalink\.php\?story_fbid=.+$",
+            r"^https://(?:www\.)?facebook\.com/media/set/?\?set=.+$",
+            r"^https://(?:www\.)?facebook\.com/notes/.+?/.+?/.+$",
+            # At the moment, not documented on https://developers.facebook.com/docs/plugins/oembed-endpoints
+            # Works for posts with a single photo
+            r"^https://(?:www\.)?facebook\.com/.+?/photos/.+$",
+        ],
+    },
+]
+
+
+class FacebookOEmbedFinder(OEmbedFinder):
+    """
+    An embed finder that supports the authenticated Facebook oEmbed Endpoint.
+    https://developers.facebook.com/docs/plugins/oembed
+    """
+
+    def __init__(self, omitscript=False, app_id=None, app_secret=None):
+        # {settings.facebook_APP_ID}|{settings.facebook_APP_SECRET}
+        self.app_id = app_id
+        self.app_secret = app_secret
+        self.omitscript = omitscript
+
+        super().__init__(providers=FACEBOOK_PROVIDERS)
+
+    def find_embed(self, url, max_width=None, max_height=None):
+        # Find provider
+        endpoint = self._get_endpoint(url)
+        if endpoint is None:
+            raise EmbedNotFoundException
+
+        params = {"url": url, "format": "json"}
+        if max_width:
+            params["maxwidth"] = max_width
+        if max_height:
+            params["maxheight"] = max_height
+        if self.omitscript:
+            params["omitscript"] = "true"
+
+        # Configure request
+        request = Request(endpoint + "?" + urlencode(params))
+        request.add_header("Authorization", f"Bearer {self.app_id}|{self.app_secret}")
+
+        # Perform request
+        try:
+            r = urllib_request.urlopen(request)
+        except (HTTPError, URLError) as e:
+            if isinstance(e, HTTPError) and e.code == 404:
+                raise EmbedNotFoundException
+            elif isinstance(e, HTTPError) and e.code in [400, 401, 403]:
+                raise AccessDeniedFacebookOEmbedException
+            else:
+                raise EmbedNotFoundException
+        oembed = json.loads(r.read().decode("utf-8"))
+
+        # Return embed as a dict
+        return {
+            "title": oembed["title"] if "title" in oembed else "",
+            "author_name": oembed["author_name"] if "author_name" in oembed else "",
+            "provider_name": oembed["provider_name"]
+            if "provider_name" in oembed
+            else "Facebook",
+            "type": oembed["type"],
+            "thumbnail_url": oembed.get("thumbnail_url"),
+            "width": oembed.get("width"),
+            "height": oembed.get("height"),
+            "html": oembed.get("html"),
+        }
+
+
+embed_finder_class = FacebookOEmbedFinder

env/lib/python3.10/site-packages/wagtail/embeds/finders/instagram.py

@@ -0,0 +1,91 @@
+import json
+from urllib import request as urllib_request
+from urllib.error import HTTPError, URLError
+from urllib.parse import urlencode
+from urllib.request import Request
+
+from wagtail.embeds.exceptions import EmbedException, EmbedNotFoundException
+
+from .oembed import OEmbedFinder
+
+
+class AccessDeniedInstagramOEmbedException(EmbedException):
+    pass
+
+
+INSTAGRAM_PROVIDER = {
+    "endpoint": "https://graph.facebook.com/v11.0/instagram_oembed",
+    "urls": [
+        r"^https?://(?:www\.)?instagram\.com/p/.+$",
+        r"^https?://(?:www\.)?instagram\.com/tv/.+$",
+        r"^https?://(?:www\.)?instagram\.com/reel/.+$",
+    ],
+}
+
+
+class InstagramOEmbedFinder(OEmbedFinder):
+    """
+    An embed finder that supports the authenticated Instagram oEmbed Endpoint.
+    https://developers.facebook.com/docs/instagram/oembed
+    """
+
+    def __init__(self, omitscript=False, app_id=None, app_secret=None):
+        # {settings.INSTAGRAM_APP_ID}|{settings.INSTAGRAM_APP_SECRET}
+        self.app_id = app_id
+        self.app_secret = app_secret
+        self.omitscript = omitscript
+
+        super().__init__(providers=[INSTAGRAM_PROVIDER])
+
+    def find_embed(self, url, max_width=None, max_height=None):
+        # Find provider
+        endpoint = self._get_endpoint(url)
+        if endpoint is None:
+            raise EmbedNotFoundException
+
+        params = {"url": url, "format": "json"}
+        if max_width:
+            params["maxwidth"] = max_width
+        if max_height:
+            params["maxheight"] = max_height
+        if self.omitscript:
+            params["omitscript"] = "true"
+
+        # Configure request
+        request = Request(endpoint + "?" + urlencode(params))
+        request.add_header("Authorization", f"Bearer {self.app_id}|{self.app_secret}")
+
+        # Perform request
+        try:
+            r = urllib_request.urlopen(request)
+        except (HTTPError, URLError) as e:
+            if isinstance(e, HTTPError) and e.code == 404:
+                raise EmbedNotFoundException
+            elif isinstance(e, HTTPError) and e.code in [400, 401, 403]:
+                raise AccessDeniedInstagramOEmbedException
+            else:
+                raise EmbedNotFoundException
+        oembed = json.loads(r.read().decode("utf-8"))
+
+        # Convert photos into HTML
+        if oembed["type"] == "photo":
+            html = '<img src="{}" alt="">'.format(oembed["url"])
+        else:
+            html = oembed.get("html")
+
+        # Return embed as a dict
+        return {
+            "title": oembed["title"] if "title" in oembed else "",
+            "author_name": oembed["author_name"] if "author_name" in oembed else "",
+            "provider_name": oembed["provider_name"]
+            if "provider_name" in oembed
+            else "Instagram",
+            "type": oembed["type"],
+            "thumbnail_url": oembed.get("thumbnail_url"),
+            "width": oembed.get("width"),
+            "height": oembed.get("height"),
+            "html": html,
+        }
+
+
+embed_finder_class = InstagramOEmbedFinder

env/lib/python3.10/site-packages/wagtail/embeds/finders/oembed.py

@@ -0,0 +1,99 @@
+import json
+import re
+from datetime import timedelta
+from urllib import request as urllib_request
+from urllib.error import URLError
+from urllib.parse import urlencode
+from urllib.request import Request
+
+from django.utils import timezone
+
+from wagtail.embeds.exceptions import EmbedNotFoundException
+from wagtail.embeds.oembed_providers import all_providers
+
+from .base import EmbedFinder
+
+
+class OEmbedFinder(EmbedFinder):
+    options = {}
+    _endpoints = None
+
+    def __init__(self, providers=None, options=None):
+        self._endpoints = {}
+
+        for provider in providers or all_providers:
+            patterns = []
+
+            endpoint = provider["endpoint"].replace("{format}", "json")
+
+            for url in provider["urls"]:
+                patterns.append(re.compile(url))
+
+            self._endpoints[endpoint] = patterns
+
+        if options:
+            self.options = self.options.copy()
+            self.options.update(options)
+
+    def _get_endpoint(self, url):
+        for endpoint, patterns in self._endpoints.items():
+            for pattern in patterns:
+                if re.match(pattern, url):
+                    return endpoint
+
+    def accept(self, url):
+        return self._get_endpoint(url) is not None
+
+    def find_embed(self, url, max_width=None, max_height=None):
+        # Find provider
+        endpoint = self._get_endpoint(url)
+        if endpoint is None:
+            raise EmbedNotFoundException
+
+        # Work out params
+        params = self.options.copy()
+        params["url"] = url
+        params["format"] = "json"
+        if max_width:
+            params["maxwidth"] = max_width
+        if max_height:
+            params["maxheight"] = max_height
+
+        # Perform request
+        request = Request(endpoint + "?" + urlencode(params))
+        request.add_header("User-agent", "Mozilla/5.0")
+        try:
+            r = urllib_request.urlopen(request)
+            oembed = json.loads(r.read().decode("utf-8"))
+        except (URLError, json.decoder.JSONDecodeError):
+            raise EmbedNotFoundException
+
+        # Convert photos into HTML
+        if oembed["type"] == "photo":
+            html = '<img src="{}" alt="">'.format(oembed["url"])
+        else:
+            html = oembed.get("html")
+
+        # Return embed as a dict
+        result = {
+            "title": oembed.get("title", ""),
+            "author_name": oembed.get("author_name", ""),
+            "provider_name": oembed.get("provider_name", ""),
+            "type": oembed["type"],
+            "thumbnail_url": oembed.get("thumbnail_url"),
+            "width": oembed.get("width"),
+            "height": oembed.get("height"),
+            "html": html,
+        }
+
+        try:
+            cache_age = int(oembed["cache_age"])
+        except (KeyError, TypeError, ValueError):
+            pass
+        else:
+            result["cache_until"] = timezone.now() + timedelta(seconds=cache_age)
+
+        return result
+
+
+embed_finder_class = OEmbedFinder