assistant: Strip out general website chrome in /fetch
command (#13264)
This PR updates the `/fetch` command to strip out general website chrome that likely won't contain content on any websites. Release Notes: - N/A
This commit is contained in:
parent
aff7a83815
commit
ddf07253c4
2 changed files with 25 additions and 0 deletions
|
@ -1,6 +1,30 @@
|
|||
use crate::html_element::HtmlElement;
|
||||
use crate::markdown_writer::{HandleTag, HandlerOutcome, MarkdownWriter, StartTagOutcome};
|
||||
|
||||
pub struct WebpageChromeRemover;
|
||||
|
||||
impl HandleTag for WebpageChromeRemover {
|
||||
fn should_handle(&self, tag: &str) -> bool {
|
||||
match tag {
|
||||
"head" | "script" | "style" | "nav" => true,
|
||||
_ => false,
|
||||
}
|
||||
}
|
||||
|
||||
fn handle_tag_start(
|
||||
&mut self,
|
||||
tag: &HtmlElement,
|
||||
_writer: &mut MarkdownWriter,
|
||||
) -> StartTagOutcome {
|
||||
match tag.tag() {
|
||||
"head" | "script" | "style" | "nav" => return StartTagOutcome::Skip,
|
||||
_ => {}
|
||||
}
|
||||
|
||||
StartTagOutcome::Continue
|
||||
}
|
||||
}
|
||||
|
||||
pub struct ParagraphHandler;
|
||||
|
||||
impl HandleTag for ParagraphHandler {
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue