mirror of
https://github.com/neon-mmd/websurfx.git
synced 2024-11-23 06:28:23 -05:00
9f23a1c70b
* 🧑💻 fix: add closing curly brace in the nix build step * ✨ feat: add code to provide search results from `mojeek` engine --------- Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
152 lines
4.7 KiB
Rust
152 lines
4.7 KiB
Rust
//! The `mojeek` module handles the scraping of results from the mojeek search engine
|
|
//! by querying the upstream mojeek search engine with user provided query and with a page
|
|
//! number if provided.
|
|
|
|
use std::collections::HashMap;
|
|
|
|
use reqwest::header::HeaderMap;
|
|
use reqwest::Client;
|
|
use scraper::Html;
|
|
|
|
use crate::models::aggregation_models::SearchResult;
|
|
|
|
use crate::models::engine_models::{EngineError, SearchEngine};
|
|
|
|
use error_stack::{Report, Result, ResultExt};
|
|
|
|
use super::search_result_parser::SearchResultParser;
|
|
|
|
/// A new Mojeek engine type defined in-order to implement the `SearchEngine` trait which allows to
|
|
/// reduce code duplication as well as allows to create vector of different search engines easily.
|
|
pub struct Mojeek {
|
|
/// The parser, used to interpret the search result.
|
|
parser: SearchResultParser,
|
|
}
|
|
|
|
impl Mojeek {
|
|
/// Creates the Mojeek parser.
|
|
pub fn new() -> Result<Self, EngineError> {
|
|
Ok(Self {
|
|
parser: SearchResultParser::new(
|
|
".result-col",
|
|
".results-standard li",
|
|
"a span.url",
|
|
"h2 a.title",
|
|
"p.s",
|
|
)?,
|
|
})
|
|
}
|
|
}
|
|
|
|
#[async_trait::async_trait]
|
|
impl SearchEngine for Mojeek {
|
|
async fn results(
|
|
&self,
|
|
query: &str,
|
|
page: u32,
|
|
user_agent: &str,
|
|
client: &Client,
|
|
safe_search: u8,
|
|
) -> Result<HashMap<String, SearchResult>, EngineError> {
|
|
// Mojeek uses `start results from this number` convention
|
|
// So, for 10 results per page, page 0 starts at 1, page 1
|
|
// starts at 11, and so on.
|
|
let results_per_page = 10;
|
|
let start_result = results_per_page * page + 1;
|
|
|
|
let results_per_page = results_per_page.to_string();
|
|
let start_result = start_result.to_string();
|
|
|
|
let search_engines = vec![
|
|
"Bing",
|
|
"Brave",
|
|
"DuckDuckGo",
|
|
"Ecosia",
|
|
"Google",
|
|
"Lilo",
|
|
"Metager",
|
|
"Qwant",
|
|
"Startpage",
|
|
"Swisscows",
|
|
"Yandex",
|
|
"Yep",
|
|
"You",
|
|
];
|
|
let qss = search_engines.join("%2C");
|
|
let safe = if safe_search == 0 { "0" } else { "1" };
|
|
|
|
// Mojeek detects automated requests, these are preferences that are
|
|
// able to circumvent the countermeasure. Some of these are
|
|
// not documented in their Search API
|
|
let query_params: Vec<(&str, &str)> = vec![
|
|
("t", results_per_page.as_str()),
|
|
("theme", "dark"),
|
|
("arc", "none"),
|
|
("date", "1"),
|
|
("cdate", "1"),
|
|
("tlen", "100"),
|
|
("ref", "1"),
|
|
("hp", "minimal"),
|
|
("lb", "en"),
|
|
("qss", &qss),
|
|
("safe", safe),
|
|
];
|
|
|
|
let mut query_params_string = String::new();
|
|
for (k, v) in &query_params {
|
|
query_params_string.push_str(&format!("&{k}={v}"));
|
|
}
|
|
|
|
let url: String = match page {
|
|
0 => {
|
|
format!("https://www.mojeek.com/search?q={query}{query_params_string}")
|
|
}
|
|
_ => {
|
|
format!(
|
|
"https://www.mojeek.com/search?q={query}&s={start_result}{query_params_string}"
|
|
)
|
|
}
|
|
};
|
|
|
|
let mut cookie_string = String::new();
|
|
for (k, v) in &query_params {
|
|
cookie_string.push_str(&format!("{k}={v}; "));
|
|
}
|
|
|
|
let header_map = HeaderMap::try_from(&HashMap::from([
|
|
("USER_AGENT".to_string(), user_agent.to_string()),
|
|
("REFERER".to_string(), "https://google.com/".to_string()),
|
|
(
|
|
"CONTENT_TYPE".to_string(),
|
|
"application/x-www-form-urlencoded".to_string(),
|
|
),
|
|
("COOKIE".to_string(), cookie_string),
|
|
]))
|
|
.change_context(EngineError::UnexpectedError)?;
|
|
|
|
let document: Html = Html::parse_document(
|
|
&Mojeek::fetch_html_from_upstream(self, &url, header_map, client).await?,
|
|
);
|
|
|
|
if let Some(no_result_msg) = self.parser.parse_for_no_results(&document).nth(0) {
|
|
if no_result_msg
|
|
.inner_html()
|
|
.contains("No pages found matching:")
|
|
{
|
|
return Err(Report::new(EngineError::EmptyResultSet));
|
|
}
|
|
}
|
|
|
|
// scrape all the results from the html
|
|
self.parser
|
|
.parse_for_results(&document, |title, url, desc| {
|
|
Some(SearchResult::new(
|
|
title.inner_html().trim(),
|
|
url.inner_html().trim(),
|
|
desc.inner_html().trim(),
|
|
&["mojeek"],
|
|
))
|
|
})
|
|
}
|
|
}
|