rss-cli 0.1.0

An AI-friendly, cache-backed RSS / Atom / JSON Feed CLI that also runs as an MCP server
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196

//! Feed autodiscovery from a website homepage. **Owner: `cli` agent.**
//!
//! ## Requirements
//! - GET the `site_url` HTML via [`HttpClient::get_bytes`].
//! - Parse the `<head>` for `<link rel="alternate" type="application/rss+xml">` and
//!   `type="application/atom+xml"` (also accept `application/json` / `feed+json`). Use the
//!   lightweight `tl` HTML parser — do not pull a heavyweight DOM stack.
//! - Resolve each `href` to an absolute URL against `site_url` (the `url` crate).
//! - Map the `type` to `feed_type` (`"rss" | "atom" | "json"`), carry the `title` attr.
//! - If the page *is itself* a feed (content sniffing optional), or if no `<link>` tags are
//!   found, return an empty list rather than erroring (the CLI decides exit behavior).

use std::collections::HashSet;

use url::Url;

use crate::error::RssError;
use crate::fetch::HttpClient;
use crate::model::{DiscoverOutput, DiscoveredFeed};

/// Discover feeds advertised on `site_url`. **Owner: `cli` agent.**
pub async fn discover(site_url: &str, http: &HttpClient) -> Result<DiscoverOutput, RssError> {
    let (bytes, final_url) = http.get_bytes(site_url).await?;
    let html = String::from_utf8_lossy(&bytes);
    // Resolve relative hrefs against the post-redirect URL when available.
    let feeds = extract_feeds(&html, &final_url);
    Ok(DiscoverOutput::new(site_url, feeds))
}

/// Extract `<link rel="alternate">` feed references from an HTML document, resolving each
/// `href` to an absolute URL against `base_url`. Factored out so it can be unit-tested on
/// inline HTML without the network. De-dup is order-preserving on the resolved URL.
fn extract_feeds(html: &str, base_url: &str) -> Vec<DiscoveredFeed> {
    let dom = match tl::parse(html, tl::ParserOptions::default()) {
        Ok(dom) => dom,
        // A malformed homepage should not error the whole command; treat it as no feeds.
        Err(_) => return Vec::new(),
    };
    let base = Url::parse(base_url).ok();

    let mut feeds: Vec<DiscoveredFeed> = Vec::new();
    let mut seen: HashSet<String> = HashSet::new();

    for node in dom.nodes() {
        let Some(tag) = node.as_tag() else { continue };
        if !tag.name().as_bytes().eq_ignore_ascii_case(b"link") {
            continue;
        }
        let attrs = tag.attributes();
        let rel = attr(attrs, "rel");
        // Only consider `<link rel="alternate">` (rel may carry multiple space-separated tokens).
        if !rel
            .as_deref()
            .is_some_and(|r| r.to_ascii_lowercase().contains("alternate"))
        {
            continue;
        }

        let href = match attr(attrs, "href") {
            Some(h) if !h.trim().is_empty() => h,
            _ => continue,
        };
        let ty = attr(attrs, "type");
        let feed_type = match classify(ty.as_deref(), &href) {
            Some(t) => t,
            None => continue,
        };

        // Resolve to an absolute URL; skip entries we cannot resolve deterministically.
        let resolved = match &base {
            Some(b) => match b.join(href.trim()) {
                Ok(u) => u.to_string(),
                Err(_) => continue,
            },
            None => href.trim().to_string(),
        };

        if !seen.insert(resolved.clone()) {
            continue;
        }
        feeds.push(DiscoveredFeed {
            url: resolved,
            feed_type: feed_type.to_string(),
            title: attr(attrs, "title"),
        });
    }

    feeds
}

/// Read an attribute as an owned `String` (HTML entities decoded by `tl`).
fn attr(attrs: &tl::Attributes<'_>, key: &str) -> Option<String> {
    attrs
        .get(key)
        .flatten()
        .map(|b| b.as_utf8_str().into_owned())
}

/// Map a `<link>`'s `type` (and `href` as a fallback) to a feed type, or `None` to reject.
fn classify(ty: Option<&str>, href: &str) -> Option<&'static str> {
    if let Some(ty) = ty {
        // Strip any `; charset=…` parameter and normalize case.
        let mime = ty
            .split(';')
            .next()
            .unwrap_or(ty)
            .trim()
            .to_ascii_lowercase();
        return match mime.as_str() {
            "application/rss+xml" => Some("rss"),
            "application/atom+xml" => Some("atom"),
            "application/json" | "application/feed+json" => Some("json"),
            _ => None,
        };
    }
    // No `type`: fall back to a coarse extension heuristic on the href.
    let lower = href
        .split(['?', '#'])
        .next()
        .unwrap_or(href)
        .to_ascii_lowercase();
    if lower.ends_with(".rss") || lower.ends_with(".atom") || lower.ends_with(".xml") {
        Some("unknown")
    } else {
        None
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn extracts_rss_and_atom_links() {
        let html = r#"
            <html><head>
              <link rel="alternate" type="application/rss+xml" title="RSS Feed" href="/feed.rss">
              <link rel="alternate" type="application/atom+xml" title="Atom Feed" href="https://other.example/atom.xml">
            </head><body></body></html>
        "#;
        let feeds = extract_feeds(html, "https://example.com/blog/");

        assert_eq!(feeds.len(), 2);

        assert_eq!(feeds[0].feed_type, "rss");
        assert_eq!(feeds[0].title.as_deref(), Some("RSS Feed"));
        // Relative href resolved against the base URL.
        assert_eq!(feeds[0].url, "https://example.com/feed.rss");

        assert_eq!(feeds[1].feed_type, "atom");
        // Absolute href left intact.
        assert_eq!(feeds[1].url, "https://other.example/atom.xml");
    }

    #[test]
    fn accepts_json_and_strips_charset() {
        let html = r#"<link rel="alternate" type="application/feed+json; charset=utf-8" href="/feed.json">"#;
        let feeds = extract_feeds(html, "https://example.com/");
        assert_eq!(feeds.len(), 1);
        assert_eq!(feeds[0].feed_type, "json");
        assert_eq!(feeds[0].url, "https://example.com/feed.json");
    }

    #[test]
    fn rejects_non_feed_links() {
        // stylesheet, html alternate, and a bare anchor must all be ignored.
        let html = r#"
            <link rel="stylesheet" href="/style.css">
            <link rel="alternate" type="text/html" href="/index.html">
            <link rel="alternate" hreflang="fr" href="/fr/">
        "#;
        let feeds = extract_feeds(html, "https://example.com/");
        assert!(feeds.is_empty(), "expected no feeds, got {feeds:?}");
    }

    #[test]
    fn extension_heuristic_when_type_missing() {
        let html = r#"<link rel="alternate" href="/posts.atom">"#;
        let feeds = extract_feeds(html, "https://example.com/");
        assert_eq!(feeds.len(), 1);
        assert_eq!(feeds[0].feed_type, "unknown");
    }

    #[test]
    fn dedup_is_order_preserving() {
        let html = r#"
            <link rel="alternate" type="application/rss+xml" href="/feed.rss">
            <link rel="alternate" type="application/rss+xml" href="/feed.rss">
            <link rel="alternate" type="application/atom+xml" href="/feed.atom">
        "#;
        let feeds = extract_feeds(html, "https://example.com/");
        assert_eq!(feeds.len(), 2);
        assert_eq!(feeds[0].url, "https://example.com/feed.rss");
        assert_eq!(feeds[1].url, "https://example.com/feed.atom");
    }
}