单击链接后获取噩梦，等待下一页加载

Question

I'm using nightmare.js to scrape public records and am just trying to get the scraper to wait for the next page to load. I'm crawling search results which I press a next button to (obviously) get to the next page. I can't use nightmare.wait(someConstTime) to accurately wait for the next page to load because sometimes someConstTime is shorter than the time it takes for the next page to load (although it's always under 30 seconds). I also can't use nightmare.wait(selector) because the same selectors are always present on all result pages. In that case nightmare basically doesn't wait at all because the selector is already present (on the page I already scraped) so it it will proceed to scrape the same page several times unless the new page loads before the next loop.

How can I conditionally wait for the next page to load after I click on the next button?

If I could figure out how - I would compare the "Showing # to # of ## entries" indicator of the current page ( currentPageStatus ) to the last known value ( lastPageStatus ) and wait until they're different (hence the next page loaded).

(ignore that the example image only has one search result page)

I'd do that using this code from https://stackoverflow.com/a/36734481/3491991 but that would require passing lastPageStatus into deferredWait (which I can't figure out).

Here's the code I've got so far:

// Load dependencies
//const { csvFormat } = require('d3-dsv');
const Nightmare = require('nightmare');
const fs = require('fs');
var vo = require('vo');

const START = 'http://propertytax.peoriacounty.org';
var parcelPrefixes = ["01","02","03","04","05","06","07","08","09","10",
                      "11","12","13","14","15","16","17","18","19"]

vo(main)(function(err, result) {
  if (err) throw err;
});

function* main() {
  var nightmare = Nightmare(),
    currentPage = 0;
    // Go to Peoria Tax Records Search
    try {
      yield nightmare
        .goto(START)
        .wait('input[name="property_key"]')
        .insert('input[name="property_key"]', parcelPrefixes[0])
        // Click search button (#btn btn-success)
        .click('.btn.btn-success')
    } catch(e) {
      console.error(e)
    }
    // Get parcel numbers ten at a time
    try {
      yield nightmare
        .wait('.sorting_1')
        isLastPage = yield nightmare.visible('.paginate_button.next.disabled')
        while (!isLastPage) {
          console.log('The current page should be: ', currentPage); // Display page status
          try {
            const result = yield nightmare
              .evaluate(() => {
                return [...document.querySelectorAll('.sorting_1')]
                  .map(el => el.innerText);
              })
              // Save property numbers
              // fs.appendFile('parcels.txt', result, (err) => {
              //   if (err) throw err;
              //   console.log('The "data to append" was appended to file!');
              // });
          } catch(e) {
            console.error(e);
            return undefined;
          }
          yield nightmare
            // Click next page button
            .click('.paginate_button.next')
            // ************* THIS IS WHERE I NEED HELP *************** BEGIN
            // Wait for next page to load before continue while loop
            try {
              const currentPageStatus = yield nightmare
                .evaluate(() => {
                  return document.querySelector('.dataTables_info').innerText;
                })
              console.log(currentPageStatus);
            } catch(e) {
              console.error(e);
              return undefined;
            }
            // ************* THIS IS WHERE I NEED HELP *************** END
          currentPage++;
          isLastPage = yield nightmare.visible('.paginate_button.next.disabled')
        }
    } catch(e) {
      console.error(e)
    }
  yield nightmare.end();
}

Answer 1

I had a similar issue that I managed to fix. Basically I had to navigate to a search page, select the '100 per page' option and then wait for the refresh. Only problem was, it was a crapshoot as to whether a manual wait time allowed the AJAX to fire and repopulate with more than 10 results (the default).

I ended up doing this:

nightmare
.goto(url)
.wait('input.button.primary')
.click('input.button.primary')
.wait('#searchresults')
.select('#resultsPerPage',"100")
.click('input.button.primary')
.wait('.searchresult:nth-child(11)')
.evaluate(function() {
    ...
}
.end()

With this, the evaluate won't fire until it detects at least 11 divs with the class of .searchresult. Given that the default is 10, it has to wait for the reload for this to complete.

You could extend this to scrape the total number of available results from the first page to ensure that there are - in my case - more than 10 available. But the foundation of the concept works.

Answer 2

From what I could understand, basically you need the DOM change to be completed before you start extracting from the page being loaded.

In your case, the element for DOM changes is table with CSS selector: '#search-results'

I think MutationObserver is what you need.

I have used Mutation Summary library which provides a nice wrapper on raw functionality of MutationObservers, to achieve something similar

var observer = new MutationSummary({
  callback: updateWidgets,
  queries: [{
    element: '[data-widget]'
  }]
});

:From Tutorial

First register MutationSummary observer when the search results are loaded.

Then, after clicking 'Next' use nightmare.evaluate to wait for mutationSummary callback to return extracted values.