Implement Jinja2 templating for notifications body and title.

Using Jinja2 templating actually simplifies slightly the code, but, most importantly, enables a much greater flexibility
to customize the notifications' content.

This change is breaking: the tokens like `{base_url}` should instead be written s `{{ base_url }}`. It would be very easy
to make it non-breaking, but I think it would be preferable to stick to the default Jinja2 syntax. To make it non-breaking,
the Jinja2 environment should be configured this way:
```
jinja2_env = Environment(
    loader=BaseLoader,
    variable_start_string='{',
    variable_end_string='}',
)
```

NB: This change could enable a few more follow-up enhancements:
* It could make sense as a later evolution to implement the same for the URL list. This could enable features like:
  - Customizing the sender based on the watch parameters through global settings
  - Sending the notification to a different email/channel/group based on a watch attribute, like its tag
* Jinja2 plays fine with complex variables. It would be relatively easy to expose many more variables that could be used
  in the notification.

Changes
=======
* Use Jinja2 template rendering to parse tokens in the notification body and title.
* Replace the settings validator to check that the body and title are both valid Jinja2 templates
  and that they do not contain any invalid token/variable.
* Update the corresponding documentation, setting pages and tests.

changedetectionio/forms.py

@@ -133,7 +133,7 @@ class ValidateAppRiseServers(object):
                 message = field.gettext('\'%s\' is not a valid AppRise URL.' % (server_url))
                 raise ValidationError(message)
 
-class ValidateTokensList(object):
+class ValidateJinja2Template(object):
     """
     Validates that a {token} is from a valid set
     """
@@ -142,11 +142,22 @@ class ValidateTokensList(object):
 
     def __call__(self, form, field):
         from changedetectionio import notification
-        regex = re.compile('{.*?}')
-        for p in re.findall(regex, field.data):
-            if not p.strip('{}') in notification.valid_tokens:
-                message = field.gettext('Token \'%s\' is not a valid token.')
-                raise ValidationError(message % (p))
+        from jinja2 import Environment, BaseLoader, TemplateSyntaxError
+        from jinja2.meta import find_undeclared_variables
+
+        try:
+            jinja2_env = Environment(loader=BaseLoader)
+            jinja2_env.globals.update(notification.valid_tokens)
+            rendered = jinja2_env.from_string(field.data).render()
+        except TemplateSyntaxError as e:
+            raise ValidationError(f"This is not a valid Jinja2 template: {e}") from e
+
+        ast = jinja2_env.parse(field.data)
+        undefined = ", ".join(find_undeclared_variables(ast))
+        if undefined:
+            raise ValidationError(
+                f"The following tokens used in the notification are not valid: {undefined}"
+            )
 
 class ValidateListRegex(object):
     """
@@ -201,8 +212,8 @@ class quickWatchForm(Form):
 class commonSettingsForm(Form):
 
     notification_urls = StringListField('Notification URL List', validators=[validators.Optional(), ValidateAppRiseServers()])
-    notification_title = StringField('Notification Title', default='ChangeDetection.io Notification - {watch_url}', validators=[validators.Optional(), ValidateTokensList()])
-    notification_body = TextAreaField('Notification Body', default='{watch_url} had a change.', validators=[validators.Optional(), ValidateTokensList()])
+    notification_title = StringField('Notification Title', default='ChangeDetection.io Notification - {{ watch_url }}', validators=[validators.Optional(), ValidateJinja2Template()])
+    notification_body = TextAreaField('Notification Body', default='{{ watch_url }} had a change.', validators=[validators.Optional(), ValidateJinja2Template()])
     trigger_check = BooleanField('Send test notification on save')
     fetch_backend = RadioField(u'Fetch Method', choices=content_fetcher.available_fetchers(), validators=[ValidateContentFetcherIsReady()])
     extract_title_as_title = BooleanField('Extract <title> from document and use as watch title', default=False)