* RSS Funnel

The RSS Funnel is a modular RSS processing pipeline. It is designed to be used to modify existing RSS source in various interesting ways such as:

- Fetch full content
- Generate a RSS feed from an HTML page
- Remove unwanted elements from the article (using a CSS selector)
- Keep or remove articles matching keywords or patterns
- Highlight keywords in articles
- Redact or replace text in the article (using a regular expression)
- Split a single RSS article into multiple articles
- Merge multiple feeds into a single feed
- Run arbitrary JS code to transform the feed or articles (with [[https://github.com/shouya/rss-funnel/wiki/JS-DOM-API][DOM API support]])

[[https://rss-funnel-demo.fly.dev/][Try out the live demo!]]

** Installation

You can use the docker image ([[https://github.com/shouya/rss-funnel/pkgs/container/rss-funnel][latest version]]) in your =docker-compose.yaml=:

#+begin_src yaml
version: "3.8"
services:
  rss-funnel:
    image: ghcr.io/shouya/rss-funnel:latest
    ports:
      - 4080:4080
    volumes:
      - ./funnel.yaml:/funnel.yaml
    command: /rss-funnel -c /funnel.yaml server -b 0.0.0.0:4080
#+end_src

Alternatively, you can build it directly from source:

#+begin_src bash
git clone https://github.com/shouya/rss-funnel.git
cd rss-funnel

# first build the front-end assets
cd inspector && npm i && npm run build && cd ..

# then build the binary
cargo build --release
#+end_src

Or if you prefer not to build from source, you can download the pre-built artifacts from [[https://github.com/shouya/rss-funnel/releases][release page]].

** Usage

To use =rss-funnel=, you need to supply a configuration file in YAML. Here is an example configuration.

#+begin_src yaml
endpoints:
  - path: /tokio-blog.xml
    note: Full text of Tokio blog
    source: https://tokio.rs/_next/static/feed.xml
    filters:
      - full_text: {}
      - simplify_html: {}

  - path: /solidot.xml
    note: Solidot news with links
    source: https://www.solidot.org/index.rss
    filters:
      - full_text: {}
      - keep_element: .p_mainnew
      - simplify_html: {}
      - sanitize:
          - replace_regex:
              from: "(?<link>http(s)?://[^< \n]*)"
              to: '<a href="$link">$link</a>'

  - path: /hackernews.xml
    note: Full text of Hacker News
    source: https://news.ycombinator.com/rss
    filters:
      - full_text:
          simplify: true
          append_mode: true
#+end_src

Save above file to =/path/to/funnel.yaml= and run the following command:

#+begin_src
rss-funnel -c /path/to/funnel.yaml server
#+end_src

You can optionally specify the bind address and port (default =127.0.0.1:4080=). Detailed usage can be found in =--help= output.

The endpoints like =http://127.0.0.1:4080/tokio-blog.xml= should be serving the filtered feeds.

** Endpoint

Each of the configuration contains a number of endpoints. Each endpoint correspond to a RSS feed.

Properties:

- =path= (required): The path of the endpoint. The path should start with =/=.
- =note= (optional): A note for the endpoint. Only used for display purpose.
- =source= (optional): The source url of the RSS feed.
  + If not specified, you must specify =?source=<url>= query in the request. This allows for usages like applying same filters for different feeds.
  + If the source points to a HTML page, =rss-funnel= will try to generate a RSS feed from the page with a single article. You can then use =split= filter to split the single article into multiple articles. See [[https://github.com/shouya/rss-funnel/wiki/Cookbook#hacker-news-top-links][Cookbook: Hacker News Top Links]] for an example.
- =filters= (required): A list of filters to apply to the feed.
  + The feed from the =source= goes through the filters in the order specified. You can think of each filter as corresponding to a transformation on the =Feed=.
  + Each filter is specified as an YAML object with the singleton key being the name of the filter and the value being the configuration of the filter.
    - For example, in the filter definition: =- keep_element: .p_mainnew=
      + the filter's name is =keep_element=
      + the configuration is the string value =.p_mainnew=. Depending on the filter, the configuration can have different types.
  + The =Feed= object from the last filter is returned as the response.
- =client= (optional): The configuration for the HTTP client used to fetch the source like the user_agent. See [[https://github.com/shouya/rss-funnel/wiki/Client-config][Client config]] for detail.

** Filters

See [[https://github.com/shouya/rss-funnel/wiki/Filters][Filters]] for the documentations for all available filters.

** Cookbook

See [[https://github.com/shouya/rss-funnel/wiki/Cookbook][Cookbook]] for some examples of using =rss-funnel=.