Store entry source

[SorenDalsgaard]

Hi

First of thanks for building this great tool.

I am having a small issue with understanding how to store elements from a feed item. For this feed https://www.newsdesk.lexisnexis.com/feed/aeb9dae6799dcbd4.rss I am interested in storing the source element.

How could I store this for each entry in the feed? Thanks

[lemon24]

Hi, thanks for reaching out!

source doesn't seems to be stored at the moment; reader started out using a subset of the stuff feedparser exposes, and I kept adding more and more fields as I needed them.

I think source should be relatively easy to add. I can't promise I can get to it anytime soon, but I will prioritize reviewing a pull request for it.

[lemon24]

Some notes on implementing this.

The RSS source just points to the original feed:

<source url="http://example.com/feed.xml">Some Feed</source>

The Atom source is way more complicated, containing a subset of the feed elements.

feedparser represents the RSS source element like this:

{'href': 'http://example.com/feed.xml', 'title': 'Some Feed'}

Note the href is not a documented attribute of the feedparser entries[i].source (but is accessible as source.url, because source is a FeedParserDict). The Atom equivalent of href would then be link[rel=self], and the reader equivalent url (per #153, and for similarity with Feed).

It should be pretty safe to make Entry.source have a subset of the Feed attributes:

class Entry:
    ...
    source: Optional[EntrySource]

class EntrySource:
    url: str  # RSS href/url; Atom link[rel=self]
    updated: Optional[datetime] = None
    title: Optional[str] = None
    link: Optional[str] = None
    author: Optional[str] = None
    subtitle: Optional[str] = None

[lemon24]

2024 implementation notes:

• We probably want change_feed_url() to set Entry.source with the details of the old feed, if the entry does not already have a source (in this case, Entry.source.url would be the same as Entry.original_feed_url).
• However, if the entry in the parsed feed has a source, that should always overwrite Entry.source.

[lemon24]

Additional implementation notes (in light of #290):

• We should probably also store feed.user_title with .source (problem: what if the user wants to change it?).
• It'd be nice if source.title / source.user_title were added to the search index (ideally as part of / default for feed).