This commit is contained in:
2019-02-07 16:18:59 +01:00
parent 3edc4e411a
commit 70f89cc35c
9 changed files with 129 additions and 101 deletions

View File

@@ -1,8 +1,18 @@
use lazy_static::lazy_static;
use scraper::{Html, Selector};
use crate::entry::Entry;
use crate::html::SelectExt;
use crate::probe::Probe;
lazy_static! {
static ref MESSAGE: Selector = Selector::parse(".CompanyResultListItem h3.name > a").unwrap();
static ref HISTORY_1: Selector =
Selector::parse("div.PhoneNoHit div.search-info-container p").unwrap();
static ref HISTORY_2: Selector =
Selector::parse("div.feedback-types div.feedback-type-item").unwrap();
}
fn from_html(document: &str) -> Result<Entry, ()> {
let html = Html::parse_document(document);
@@ -10,40 +20,23 @@ fn from_html(document: &str) -> Result<Entry, ()> {
let mut history = Vec::new();
let comments = Vec::new();
let selector = Selector::parse(".CompanyResultListItem h3.name > a").unwrap();
if let Some(element) = html.select(&selector).next() {
let message = element.inner_html();
let message = htmlescape::decode_html(&message).unwrap();
if let Some(message) = html
.select(&MESSAGE)
.next()
.map(|element| element.easy_text())
{
messages.push(message);
}
let selector = Selector::parse("div.PhoneNoHit div.search-info-container p").unwrap();
if let Some(element) = html.select(&selector).next() {
let message = element
.text()
.map(str::trim)
.filter(|s| !s.is_empty())
.collect::<Vec<_>>()
.join(" ");
let message = htmlescape::decode_html(&message).unwrap();
if let Some(message) = html
.select(&HISTORY_1)
.next()
.map(|element| element.easy_text())
{
history.push(message);
}
let selector = Selector::parse("div.feedback-types div.feedback-type-item").unwrap();
for element in html.select(&selector) {
let message = element
.text()
.map(str::trim)
.filter(|s| !s.is_empty())
.collect::<Vec<_>>()
.join(" ");
for message in html.select(&HISTORY_2).map(|element| element.easy_text()) {
history.push(message);
}