torrents.csv/server/service/src/main.rs

extern crate actix_web;
extern crate serde;
#[macro_use]
extern crate serde_derive;
extern crate grep;
extern crate time;

use actix_web::{fs, fs::NamedFile, http, server, App, HttpRequest, HttpResponse, Query};

use std::error::Error;

use grep::regex::RegexMatcherBuilder;
use grep::searcher::sinks::Lossy;
use grep::searcher::{BinaryDetection, SearcherBuilder};
use std::env;
use std::fs::File;

fn main() {

  
  server::new(|| {
    App::new()
      .route("/service/search", http::Method::GET, search)
      .resource("/favicon.ico", |r| r.f(favicon))
      .resource("/", |r| r.f(index))
      .handler(
        "/static",
        fs::StaticFiles::new(front_end_dir())
          .unwrap()
          // .index_file("index.html"),
      )
      .finish()
  }).bind("0.0.0.0:8080")
    .unwrap()
    .run();
}

fn index(_req: &HttpRequest) -> Result<NamedFile, actix_web::error::Error> {
  Ok(NamedFile::open(front_end_dir() + "/index.html")?)
}

fn favicon(_req: &HttpRequest) -> Result<NamedFile, actix_web::error::Error> {
  Ok(NamedFile::open(front_end_dir() + "/favicon.ico")?)
}

fn front_end_dir() -> String {
  env::var("TORRENTS_CSV_FRONT_END_DIR").unwrap_or("../ui/dist".to_string())
}

fn torrents_csv_file() -> String {
  env::var("TORRENTS_CSV_FILE").unwrap_or("../../torrents.csv".to_string())
}

#[derive(Deserialize)]
struct SearchQuery {
  q: String,
  page: Option<usize>,
  size: Option<usize>,
}

fn search(query: Query<SearchQuery>) -> HttpResponse {
  HttpResponse::Ok()
    .header("Access-Control-Allow-Origin", "*")
    .content_type("text/csv")
    .body(ripgrep(query))
}

fn ripgrep(query: Query<SearchQuery>) -> String {
  let page = query.page.unwrap_or(1);
  let size = query.size.unwrap_or(10);
  let offset = size * (page - 1);
  let csv_file = File::open(torrents_csv_file());

  println!("query = {} , page = {}, size = {}", query.q, page, size);

  let results = search_file(csv_file.unwrap(), &query.q).unwrap();

  let last: usize = if offset + size < results.len() {
    offset + size
  } else {
    results.len()
  };

  results[offset..last].join("")
}

fn search_file(file: File, query: &str) -> Result<Vec<String>, Box<Error>> {
  let pattern = query.replace(" ", ".*");

  let matcher = RegexMatcherBuilder::new()
    .case_insensitive(true)
    .build(&pattern)
    .unwrap();

  let mut matches: Vec<String> = vec![];

  let mut searcher = SearcherBuilder::new()
    .binary_detection(BinaryDetection::quit(b'\x00'))
    // .line_number(false)
    .build();

  searcher.search_file(
    &matcher,
    &file,
    Lossy(|_lnum, line| {
      matches.push(line.to_string());
      Ok(true)
    }),
  )?;

  Ok(matches)
}

#[cfg(test)]
mod tests {
  use std::fs::File;
  use time::PreciseTime;

  #[test]
  fn test() {
    let start = PreciseTime::now();
    let results =
      super::search_file(File::open(super::torrents_csv_file()).unwrap(), "sherlock").unwrap();
    assert!(results.len() > 2);
    let end = PreciseTime::now();
    println!("Query took {} seconds.", start.to(end));
  }
}