public static void Run(SavePageForm parentForm, string url)
{
Form = parentForm;
Url = url;
c = new Crawler(new Uri(url),
new HtmlDocumentProcessor(), // Process html
new SaveFileStep());
c.MaximumThreadCount = 1;
c.MaximumCrawlDepth = 1;
c.ExcludeFilter = CrawlUtil.ExtensionsToSkip;
c.AdhereToRobotRules = false;
c.CrawlFinished += new EventHandler <NCrawler.Events.CrawlFinishedEventArgs>(c_CrawlFinished);
string ua = CrawlUtil.GetRandomUnblockedUserAgent(UserAgentTracker);
//if there are no unblocked user agents left then reset the tracker and retry
if (ua == null)
{
UserAgentTracker = CrawlUtil.InitUserAgentTracker();
ua = CrawlUtil.GetRandomUnblockedUserAgent(UserAgentTracker);
}
c.UserAgent = ua;
// Begin crawl
c.Crawl();
}
/*
* [numFiveStarRatings][numFourStarRatings][numThreeStarRatings][numTwoStarRatings][numOneStarRatings]
* <tr>
* <td align="left" style="padding-right:0.5em;padding-bottom:1px;white-space:nowrap;font-size:10px;">
* <a href="http://www.amazon.com/Already-Dead-Joe-Pitt-Novel/product-reviews/034547824X/ref=cm_cr_pr_hist_5?ie=UTF8&filterBy=addFiveStar&showViewpoints=0" style="font-family:Verdana,Arial,Helvetica,Sans-serif;">
* 5 star
* </a>:
* </td>
* <td style="min-width:60; background-color: #eeeecc" width="60" align="left" class="tiny" title="49%">
* <...>
* </td>
* <td align="right" style="font-family:Verdana,Arial,Helvetica,Sans-serif;;font-size:10px;">
* (82)
* </td>
* </tr>
*/
internal static int GetStarRating(HtmlDocument htmlDoc, Book b, string starNum)
{
int rating = 0;
var nStarNode = htmlDoc.DocumentNode.SelectSingleNode("//a[contains(@href,'filterBy=add" + starNum + "Star')]/../..");
if (nStarNode != null)
{
try
{
//this method works for a review page
rating = CrawlUtil.ExtractRatingFromStarString(CrawlUtil.TidyHtmlText(nStarNode.InnerText));
//try another method (for detail page)
if (rating == -1)
{
var n = nStarNode.SelectSingleNode("//a[contains(@href,'filterBy=add" + starNum + "Star')]");
rating = CrawlUtil.ExtractRatingFromStarStringFromDetailPage(CrawlUtil.TidyHtmlText(n.InnerText));
}
//if it failed, try another method (for detail page)
if (rating == -1)
{
rating = CrawlUtil.ExtractRatingFromStarStringFromDetailPage(CrawlUtil.TidyHtmlText(nStarNode.InnerText));
}
}
catch (Exception ex)
{
//old code
//ReviewHtmlCrawler.form.appendLineToLog("***ERROR*** " + "\t** [" + starNum + " star rating] not found (" + ex.Message + ")");
}
}
return(rating);
}
public static void Run(MainForm parentForm, Book book, bool getRatingStats, bool getReviews)
{
form = parentForm;
IsolatedStorageModule.Setup(false);
GetRatingStats = getRatingStats;
GetReviews = getReviews;
currentBook = book;
baseUri = book.reviewPageURL;
/*
* 140185852 We
* http://www.amazon.com/We-Yevgeny-Zamyatin/product-reviews/0140185852
* http://www.amazon.com/We-Yevgeny-Zamyatin/product-reviews/0140185852/ref=cm_cr_pr_btm_link_4?pageSize=50&pageNumber=4&sortBy=recent
*
* 618260307 The Hobbit http://www.amazon.com/The-Hobbit-J-R-R-Tolkien/product-reviews/0618260307
*/
baseUri += "/ref=cm_cr_pr_btm_link_1?pageSize=50&pageNumber=1";
if (!currentBook.reviewPageURL.Contains("/ref=cm_cr_pr_btm_link"))
{
currentBook.reviewPageURL = baseUri; //hack to make isFirstPage() work [2016-02-04]
}
c = new Crawler(new Uri(baseUri),
new HtmlDocumentProcessor(), // Process html
new ReviewPageProcessStep(),
new SaveFileStep());
// Custom step to visualize crawl
c.MaximumThreadCount = 1;
c.MaximumCrawlDepth = 1;
c.ExcludeFilter = CrawlUtil.ExtensionsToSkip;
c.AdhereToRobotRules = false;
c.BeforeDownload += new EventHandler <NCrawler.Events.BeforeDownloadEventArgs>(c_BeforeDownload);
string ua = CrawlUtil.GetRandomUnblockedUserAgent(UserAgentTracker);
//if there are no unblocked user agents left then reset the tracker and retry
if (ua == null)
{
UserAgentTracker = CrawlUtil.InitUserAgentTracker();
ua = CrawlUtil.GetRandomUnblockedUserAgent(UserAgentTracker);
}
c.UserAgent = ua;
// Begin crawl
c.Crawl();
}
public static void Run(MainForm parentForm, Book book, bool getDetailsAndAuthor, bool getRanks)
{
form = parentForm;
IsolatedStorageModule.Setup(false);
GetDetailsAndAuthor = getDetailsAndAuthor;
GetRanks = getRanks;
currentBook = book;
baseUri = book.detailPageURL;
Uri u = new Uri(Uri.EscapeUriString(baseUri));
c = new Crawler(u,
new HtmlDocumentProcessor(), // Process html
new DetailPageDumperStep());
// Custom step to visualize crawl
c.MaximumThreadCount = 1;
c.MaximumCrawlDepth = 1;
c.ExcludeFilter = CrawlUtil.ExtensionsToSkip;
c.AdhereToRobotRules = false;
string ua = CrawlUtil.GetRandomUnblockedUserAgent(UserAgentTracker);
//if there are no unblocked user agents left then reset the tracker and retry
if (ua == null)
{
UserAgentTracker = CrawlUtil.InitUserAgentTracker();
ua = CrawlUtil.GetRandomUnblockedUserAgent(UserAgentTracker);
}
c.UserAgent = ua;
// Begin crawl
c.Crawl();
}
public void Process(Crawler crawler, PropertyBag propertyBag)
{
//using c.MaximumCrawlDepth = 1 means only the details page will be processed (no other links are crawled as they're depth 2)
DateTime currentDateTime = DateTime.Now;
HtmlDocument htmlDoc = propertyBag["HtmlDoc"].Value as HtmlDocument;
if (htmlDoc == null)
{
return; //this happens with HTTP errors etc. We don't bother with retrying or anything like that :(
}
BookHtmlCrawler.form.appendLineToLog("Crawling 'details page' for " + BookHtmlCrawler.currentBook.DisplayString);
BookHtmlCrawler.form.appendLineToLog(propertyBag.ResponseUri.OriginalString);
lock (this)
{
//check if the correct page is loaded (sometimes Amazon redirects to a robot check
if (propertyBag.Title.ToLower() == "robot check")
{
BookHtmlCrawler.form.appendLineToLog(propertyBag.Title + Environment.NewLine + propertyBag.Text.Substring(0, 100));
//block the user agent that's currently in use
BookHtmlCrawler.UserAgentTracker[crawler.UserAgent] = true;
//sleep between 30 secs and 3 minutes
TimeSpan duration = new TimeSpan(0, 0, RandomNumber.Next(30, 3 * 60));
BookHtmlCrawler.form.appendLineToLog(string.Format("Sleeping for {0}:{1} mins and secs before trying next book.", duration.Minutes, duration.Seconds));
System.Threading.Thread.Sleep(duration);
return;
}
string updated = "";
using (AmazonCrawlerEntities context = DbUtil.getNewContext())
{
Book b = DbUtil.getBook(context, BookHtmlCrawler.currentBook.ISBN);
Author a = null;
if (b == null) //this should never happen
{
BookHtmlCrawler.form.appendLineToLog("[error] ISBN " + BookHtmlCrawler.currentBook.ISBN + " not found in database");
return;
}
if (BookHtmlCrawler.GetDetailsAndAuthor)
{
//author, language, reading level, format, sales ranks not stored.
#region average rating
var averageRatingNode = htmlDoc.DocumentNode.SelectSingleNode(@"//div[@id='avgRating']/span/a");
if (averageRatingNode != null)
{
string ratingText = averageRatingNode.InnerText;
decimal rating = CrawlUtil.ExtractRatingFromSummaryString(CrawlUtil.TidyHtmlText(ratingText));
b.avgCustomerReview = rating;
BookHtmlCrawler.form.appendLineToLog("\tavg: " + rating);
}
#endregion
#region review page URL
string url = null;
var reviewUrlNode = htmlDoc.DocumentNode.SelectSingleNode(@"//a[@class='a-link-emphasis a-nowrap']/@href");
if (reviewUrlNode != null)
{
var working_url = reviewUrlNode.GetAttributeValue("href", null);
if (!string.IsNullOrEmpty(working_url) && working_url.IndexOf("/ref=") > 0)
{
url = working_url.Substring(0, working_url.IndexOf("/ref="));
}
}
try //save for rating and reviewURL
{
BookHtmlCrawler.form.appendLineToLog("\treview URL added: " + url);
b.reviewPageURL = url ?? "-";
context.SaveChanges();
}
catch (Exception ex)
{
//ignore :(
BookHtmlCrawler.form.appendLineToLog(ex.Message);
}
#endregion
//2015-12-01: disabled this feature. It looks like the ratings are now in %s rather than numbers.
#region rating stats
//TEMP 2014-06-12: get star rating from details page
//b.statsCollectedAt = currentDateTime;
//b.numFiveStarRatings = CrawlUtil.getStarRating(htmlDoc, b, CrawlUtil.FIVE);
//b.numFourStarRatings = CrawlUtil.getStarRating(htmlDoc, b, CrawlUtil.FOUR);
//b.numThreeStarRatings = CrawlUtil.getStarRating(htmlDoc, b, CrawlUtil.THREE);
//b.numTwoStarRatings = CrawlUtil.getStarRating(htmlDoc, b, CrawlUtil.TWO);
//b.numOneStarRatings = CrawlUtil.getStarRating(htmlDoc, b, CrawlUtil.ONE);
//BookHtmlCrawler.form.appendLineToLog("\t" + CrawlUtil.FIVE + " star: " + b.numFiveStarRatings);
//BookHtmlCrawler.form.appendLineToLog("\t" + CrawlUtil.FOUR + " star: " + b.numFourStarRatings);
//BookHtmlCrawler.form.appendLineToLog("\t" + CrawlUtil.THREE + " star: " + b.numThreeStarRatings);
//BookHtmlCrawler.form.appendLineToLog("\t" + CrawlUtil.TWO + " star: " + b.numTwoStarRatings);
//BookHtmlCrawler.form.appendLineToLog("\t" + CrawlUtil.ONE + " star: " + b.numOneStarRatings);
//try //save for rating stats
//{
// context.SaveChanges();
// updated += "RATING STATS;";
//}
//catch (Exception ex)
//{
// //ignore :(
// ReviewHtmlCrawler.form.appendLineToLog(ex.Message);
//}
#endregion
updated += "DETAILS (avg rating, reviewURL)";
}
}
BookHtmlCrawler.form.appendLineToLog("Updated " + updated + " for " + BookHtmlCrawler.currentBook.DisplayString);
}
}
private void ProcessPage(string filePath)
{
DateTime currentDateTime = DateTime.Today;
HtmlAgilityPack.HtmlDocument html = new HtmlAgilityPack.HtmlDocument();
html.Load(filePath);
HtmlNode doc = html.DocumentNode;
appendLineToLog("Crawling a 'review page' for " + filePath);
AmazonCrawlerEntities context = DbUtil.getNewContext();
HtmlNode isbnNode = doc.SelectSingleNode(".//link[@rel='canonical']/@href");
//<link rel="canonical" href="http://www.amazon.com/Breaking-Dawn-Twilight-Saga-Book/product-reviews/031606792X?pageNumber=113">
Match isbnMatch = Regex.Match(isbnNode.OuterHtml, @".+/product-reviews/(.+)\?");
string isbn = "";
if (isbnMatch == Match.Empty)
{
isbnMatch = Regex.Match(isbnNode.OuterHtml, @".+/product-reviews/(.+)(/|"")");
}
isbn = isbnMatch.Groups[1].Value;
Book b = DbUtil.getBook(context, isbn);
if (b == null) //this should never happen
{
appendLineToLog("[error] ISBN " + isbn + " not found in database");
return;
}
HtmlNodeCollection reviews = doc.SelectNodes(".//div[@class='a-section review']");
int numProcessed = 0;
int numNotProcessed = 0;
int numBadges = 0;
foreach (HtmlNode r in reviews)
{
//id
string reviewId = r.Id;
Review reviewToAdd = DbUtil.getOrCreateReview(context, reviewId);
reviewToAdd.Book = b;
//starRating
HtmlNode ratingStringNode = r.SelectSingleNode(".//i[starts-with(@class, 'a-icon a-icon-star')]");
short rating = CrawlUtil.GetStarRatingFromString(ratingStringNode.InnerText);
reviewToAdd.starRating = rating;
//reviewTitle
HtmlNode titleNode = r.SelectSingleNode(".//a[contains(@class, 'review-title')]");
string title = CrawlUtil.RemoveExcessWhitespaceFromString(titleNode.InnerText.Trim());
reviewToAdd.reviewTitle = title;
//reviewerId and reviewerName and badges
HtmlNode reviewerNode = r.SelectSingleNode(".//a[contains(@class, 'author')]");
if (reviewerNode != null)
{
string profileUrl = reviewerNode.GetAttributeValue("href", "");
string reviewerId = "";
Match m = Regex.Match(profileUrl, @"profile/(.+)/ref");
if (m.Groups.Count == 2)
{
reviewerId = m.Groups[1].Value;
}
string reviewerName = CrawlUtil.RemoveExcessWhitespaceFromString(reviewerNode.InnerText);
Reviewer reviewer = DbUtil.getOrCreateReviewer(context, reviewerId);
reviewer.reviewerName = reviewerName;
reviewer.profileURL = profileUrl;
reviewToAdd.reviewerId = reviewerId;
reviewToAdd.reviewerName = reviewerName;
//badges
HtmlNodeCollection badgeNodes = r.SelectNodes(".//a[contains(@class, 'c7y-badge-link')]");
if (badgeNodes != null)
{
AmazonCrawlerEntities tmp = DbUtil.getNewContext();
foreach (HtmlNode badgeNode in badgeNodes)
{
try
{
string badgeText = CrawlUtil.RemoveExcessWhitespaceFromString(badgeNode.InnerText);
ReviewerBadge newBadge = new ReviewerBadge();
newBadge.reviewerId = reviewerId;
newBadge.badge = badgeText;
newBadge.statsCollectedAt = currentDateTime;
tmp.ReviewerBadges.AddObject(newBadge);
tmp.SaveChanges();
numBadges += 1;
}
catch (Exception ex)
{
//ignore :(
}
}
}
}
//publishedDate
HtmlNode reviewDateNode = r.SelectSingleNode(".//span[contains(@class, 'review-date')]");
string dateText = CrawlUtil.RemoveExcessWhitespaceFromString(reviewDateNode.InnerText);
dateText = dateText.Replace("on", "");
DateTime publishedDate = DateTime.Parse(dateText);
reviewToAdd.publishedDate = publishedDate;
//reviewContent
HtmlNode reviewTextNode = r.SelectSingleNode(".//span[contains(@class, 'review-text')]");
string reviewContent = CrawlUtil.RemoveExcessWhitespaceFromString(reviewTextNode.InnerText);
reviewToAdd.reviewContent = reviewContent;
int chars = reviewContent.Length > 200 ? 200 : reviewContent.Length;
//numHelpful and numTotal
//as at 2016-03-02 the 'helpful' text forrmat is "3 people found this helpful." (numTotal doesn't exist any more)
HtmlNode helpfulNode = r.SelectSingleNode(".//span[contains(@class, 'review-votes')]");
if (helpfulNode != null)
{
if (helpfulNode.InnerText.Contains("of"))
{
int[] helpful = CrawlUtil.GetNumbersFromHelpfulText(helpfulNode.InnerText);
reviewToAdd.numHelpful = helpful[0];
reviewToAdd.numTotal = helpful[1];
}
else
{
Match m = Regex.Match(helpfulNode.InnerText, @"(\d+) p.+", RegexOptions.IgnoreCase);
if (m.Groups.Count == 2)
{
reviewToAdd.numHelpful = Int32.Parse(m.Groups[1].Value);
}
reviewToAdd.numTotal = null;
}
}
//numComments
HtmlNode numCommentsNode = r.SelectSingleNode(".//span[contains(@class, 'review-comment-total')]");
int numComments = 0;
Int32.TryParse(numCommentsNode.InnerText, out numComments);
reviewToAdd.numComments = numComments;
//isAmazonVerifiedPurchase
HtmlNode verifiedPurchaseNode = r.SelectSingleNode(".//span[@class = 'a-size-mini a-color-state a-text-bold']");
bool isVerifiedPurchase = false;
if (verifiedPurchaseNode != null)
{
isVerifiedPurchase = true;
}
reviewToAdd.isAmazonVerifiedPurchase = isVerifiedPurchase;
//format
HtmlNode formatNode = r.SelectSingleNode(".//div[contains(@class, 'review-data')]/a");
if (formatNode != null)
{
string formatText = CrawlUtil.RemoveExcessWhitespaceFromString(formatNode.InnerText).Replace("Format:", "").Trim();
reviewToAdd.format = formatText;
}
reviewToAdd.statsCollectedAt = currentDateTime;
reviewToAdd.reviewType = "R";
////debugging output
//appendLineToLog("Processing review " + reviewId);
//appendLineToLog(string.Format("\trating: {0}", rating));
//appendLineToLog(string.Format("\ttitle: {0}", title));
//appendLineToLog(string.Format("\tReviewer: {0} | {1}", reviewToAdd.reviewerId, reviewToAdd.reviewerName));
////badges not output
//appendLineToLog(string.Format("\tPublished: {0}", publishedDate));
//appendLineToLog(string.Format("\tContent: {0}", reviewContent.Substring(0, chars)));
//appendLineToLog(string.Format("\tHelpful: {0} of {1}", reviewToAdd.numHelpful, reviewToAdd.numTotal));
//appendLineToLog(string.Format("\tCommments: {0}", numComments));
//appendLineToLog(string.Format("\tisVerifiedPurchase: {0}", isVerifiedPurchase));
//appendLineToLog(string.Format("\tFormat: {0}", reviewToAdd.format));
try
{
context.SaveChanges();
numProcessed += 1;
}
catch (Exception ex)
{
appendLineToLog(ex.Message);
appendLineToLog(ex.StackTrace);
numNotProcessed += 1;
}
}
appendLineToLog(string.Format("{0} reviews saved/updated on page; {1} not saved; {2} badges.", numProcessed, numNotProcessed, numBadges));
}
public void Process(Crawler crawler, PropertyBag propertyBag)
{
DateTime currentDateTime = DateTime.Now;
HtmlDocument htmlDoc = propertyBag["HtmlDoc"].Value as HtmlDocument;
if (htmlDoc == null)
{
return; //this happens with HTTP errors etc. We don't bother with retrying or anything like that :(
}
AuthorHtmlCrawler.form.appendLineToLog("Crawling 'details (rankings) page' for " + AuthorHtmlCrawler.currentAuthor.DisplayString + " [ " + propertyBag.ResponseUri.OriginalString + " ]");
lock (this)
{
AmazonCrawlerEntities context = DbUtil.getNewContext();
Author a = DbUtil.getAuthor(context, AuthorHtmlCrawler.currentAuthor.id);
string updated = "rankings";
if (a == null) //this should never happen
{
AuthorHtmlCrawler.form.appendLineToLog("[error] author id " + AuthorHtmlCrawler.currentAuthor.id + " not found in database");
return;
}
//get rankings
var rankingNodes = htmlDoc.DocumentNode.SelectNodes(".//div[@class='nodeRank']");
if (rankingNodes != null)
{
foreach (var rankNode in rankingNodes)
{
try
{
Ranking r = new Ranking();
r.authorId = a.id;
r.statsCollectedAt = currentDateTime;
r.rankString = CrawlUtil.TidyHtmlText(rankNode.InnerText);
AuthorHtmlCrawler.form.appendLineToLog(r.rankString);
context.AddToRankings(r);
context.SaveChanges();
}
catch (Exception ex)
{
AuthorHtmlCrawler.form.appendLineToLog("**ERROR** " + ex.Message);
if (ex.InnerException != null)
{
AuthorHtmlCrawler.form.appendLineToLog("\t** " + ex.InnerException.Message);
}
}
}
}
else
{
try //add a ranking with null rankstring to mark this author has been processed
{
Ranking r = new Ranking();
r.authorId = a.id;
context.AddToRankings(r);
context.SaveChanges();
}
catch (Exception ex)
{
}
}
AuthorHtmlCrawler.form.appendLineToLog("Updated " + updated + " for " + AuthorHtmlCrawler.currentAuthor.DisplayString);
}
}
public void Process(Crawler crawler, PropertyBag propertyBag)
{
DateTime currentDateTime = DateTime.Now;
HtmlDocument htmlDoc = propertyBag["HtmlDoc"].Value as HtmlDocument;
if (htmlDoc == null)
{
return; //this happens with HTTP errors etc. We don't bother with retrying or anything like that :(
}
//check if the correct page is loaded (sometimes Amazon redirects to a robot check
if (propertyBag.Title.ToLower() == "robot check")
{
ReviewHtmlCrawler.form.appendLineToLog(propertyBag.Title + Environment.NewLine + propertyBag.Text.Substring(0, 100));
//block the user agent that's currently in use
ReviewHtmlCrawler.UserAgentTracker[crawler.UserAgent] = true;
//log the page for retry
File.AppendAllText("REVIEW-to-retry.txt", propertyBag.ResponseUri.OriginalString + System.Environment.NewLine);
//sleep between 30 secs and 1 minute
TimeSpan duration = new TimeSpan(0, 0, RandomNumber.Next(30, 60));
ReviewHtmlCrawler.form.appendLineToLog(string.Format("Sleeping for {0}:{1} mins and secs before trying next book.", duration.Minutes, duration.Seconds));
System.Threading.Thread.Sleep(duration);
return;
}
ReviewHtmlCrawler.form.appendLineToLog("Crawling a 'review page' for " + ReviewHtmlCrawler.currentBook.DisplayString);
ReviewHtmlCrawler.form.appendLineToLog(propertyBag.ResponseUri.ToString());
AmazonCrawlerEntities context = DbUtil.getNewContext();
Book b = DbUtil.getBook(context, ReviewHtmlCrawler.currentBook.ISBN);
if (b == null) //this should never happen
{
ReviewHtmlCrawler.form.appendLineToLog("[error] ISBN " + ReviewHtmlCrawler.currentBook.ISBN + " not found in database");
return;
}
//add all other pages of reviews if crawling reviews and on page 1
//TODO: also get star rating distribution
if (ReviewHtmlCrawler.GetReviews && isFirstPage(propertyBag))
{
ReviewHtmlCrawler.form.appendLineToLog("crawling first page");
int numPages = CrawlUtil.GetReviewLastPageNumber(htmlDoc.DocumentNode);
CrawlUtil.AddReviewPagesToCrawl(crawler, propertyBag.ResponseUri.OriginalString, numPages);
ReviewHtmlCrawler.form.appendLineToLog(string.Format("***************************", numPages));
ReviewHtmlCrawler.form.appendLineToLog(string.Format("*** {0} pages to crawl ***", numPages));
ReviewHtmlCrawler.form.appendLineToLog(string.Format("***************************", numPages));
if (ReviewHtmlCrawler.GetRatingStats)
{
#region rating stats
//b.statsCollectedAt = currentDateTime;
////as of 2015-12-11 the rating break-down is expressed in percentages rather than numbers.
//b.numFiveStarRatings = CrawlUtil.getStarRating(htmlDoc, b, CrawlUtil.FIVE);
//b.numFourStarRatings = CrawlUtil.getStarRating(htmlDoc, b, CrawlUtil.FOUR);
//b.numThreeStarRatings = CrawlUtil.getStarRating(htmlDoc, b, CrawlUtil.THREE);
//b.numTwoStarRatings = CrawlUtil.getStarRating(htmlDoc, b, CrawlUtil.TWO);
//b.numOneStarRatings = CrawlUtil.getStarRating(htmlDoc, b, CrawlUtil.ONE);
//try //save for rating stats
//{
// context.SaveChanges();
// updated += "RATING STATS;";
//}
//catch (Exception ex)
//{
// //ignore :(
// ReviewHtmlCrawler.form.appendLineToLog(ex.Message);
//}
#endregion
}
}
if (ReviewHtmlCrawler.GetReviews)
{
HtmlNodeCollection reviews = htmlDoc.DocumentNode.SelectNodes(".//div[@class='a-section review']");
int numProcessed = 0;
int numNotProcessed = 0;
int numBadges = 0;
foreach (HtmlNode r in reviews)
{
//id
string reviewId = r.Id;
Review reviewToAdd = DbUtil.getOrCreateReview(context, reviewId);
reviewToAdd.Book = b;
//starRating
HtmlNode ratingStringNode = r.SelectSingleNode(".//i[starts-with(@class, 'a-icon a-icon-star')]");
short rating = CrawlUtil.GetStarRatingFromString(ratingStringNode.InnerText);
reviewToAdd.starRating = rating;
//reviewTitle
HtmlNode titleNode = r.SelectSingleNode(".//a[contains(@class, 'review-title')]");
string title = CrawlUtil.RemoveExcessWhitespaceFromString(titleNode.InnerText.Trim());
reviewToAdd.reviewTitle = title;
//reviewerId and reviewerName and badges
HtmlNode reviewerNode = r.SelectSingleNode(".//a[contains(@class, 'author')]");
if (reviewerNode != null)
{
string profileUrl = reviewerNode.GetAttributeValue("href", "");
string reviewerId = "";
Match m = Regex.Match(profileUrl, @"profile/(.+)/ref");
if (m.Groups.Count == 2)
{
reviewerId = m.Groups[1].Value;
}
string reviewerName = CrawlUtil.RemoveExcessWhitespaceFromString(reviewerNode.InnerText);
Reviewer reviewer = DbUtil.getOrCreateReviewer(context, reviewerId);
reviewer.reviewerName = reviewerName;
reviewer.profileURL = profileUrl;
reviewToAdd.reviewerId = reviewerId;
reviewToAdd.reviewerName = reviewerName;
//badges
HtmlNodeCollection badgeNodes = r.SelectNodes(".//a[contains(@class, 'c7y-badge-link')]");
if (badgeNodes != null)
{
AmazonCrawlerEntities tmp = DbUtil.getNewContext();
foreach (HtmlNode badgeNode in badgeNodes)
{
try
{
string badgeText = CrawlUtil.RemoveExcessWhitespaceFromString(badgeNode.InnerText);
ReviewerBadge newBadge = new ReviewerBadge();
newBadge.reviewerId = reviewerId;
newBadge.badge = badgeText;
newBadge.statsCollectedAt = currentDateTime;
tmp.ReviewerBadges.AddObject(newBadge);
tmp.SaveChanges();
numBadges += 1;
}
catch (Exception ex)
{
//ignore :(
}
}
}
}
//publishedDate
HtmlNode reviewDateNode = r.SelectSingleNode(".//span[contains(@class, 'review-date')]");
string dateText = CrawlUtil.RemoveExcessWhitespaceFromString(reviewDateNode.InnerText);
dateText = dateText.Replace("on", "");
DateTime publishedDate = DateTime.Parse(dateText);
reviewToAdd.publishedDate = publishedDate;
//reviewContent
HtmlNode reviewTextNode = r.SelectSingleNode(".//span[contains(@class, 'review-text')]");
string reviewContent = CrawlUtil.RemoveExcessWhitespaceFromString(reviewTextNode.InnerText);
reviewToAdd.reviewContent = reviewContent;
int chars = reviewContent.Length > 200 ? 200 : reviewContent.Length;
//numHelpful and numTotal
//as at 2016-03-02 the 'helpful' text forrmat is "3 people found this helpful." (numTotal doesn't exist any more)
HtmlNode helpfulNode = r.SelectSingleNode(".//span[contains(@class, 'review-votes')]");
if (helpfulNode != null)
{
if (helpfulNode.InnerText.Contains("of"))
{
int[] helpful = CrawlUtil.GetNumbersFromHelpfulText(helpfulNode.InnerText);
reviewToAdd.numHelpful = helpful[0];
reviewToAdd.numTotal = helpful[1];
}
else
{
Match m = Regex.Match(helpfulNode.InnerText, @"(\d+) p.+", RegexOptions.IgnoreCase);
if (m.Groups.Count == 2)
{
reviewToAdd.numHelpful = Int32.Parse(m.Groups[1].Value);
}
reviewToAdd.numTotal = null;
}
}
//numComments
HtmlNode numCommentsNode = r.SelectSingleNode(".//span[contains(@class, 'review-comment-total')]");
int numComments = 0;
Int32.TryParse(numCommentsNode.InnerText, out numComments);
reviewToAdd.numComments = numComments;
//isAmazonVerifiedPurchase
HtmlNode verifiedPurchaseNode = r.SelectSingleNode(".//span[@class = 'a-size-mini a-color-state a-text-bold']");
bool isVerifiedPurchase = false;
if (verifiedPurchaseNode != null)
{
isVerifiedPurchase = true;
}
reviewToAdd.isAmazonVerifiedPurchase = isVerifiedPurchase;
//format
HtmlNode formatNode = r.SelectSingleNode(".//div[contains(@class, 'review-data')]/a");
if (formatNode != null)
{
string formatText = CrawlUtil.RemoveExcessWhitespaceFromString(formatNode.InnerText).Replace("Format:", "").Trim();
reviewToAdd.format = formatText;
}
reviewToAdd.statsCollectedAt = currentDateTime;
reviewToAdd.reviewType = "R";
//debugging output
//ReviewHtmlCrawler.form.appendLineToLog("Processing review " + reviewId);
//ReviewHtmlCrawler.form.appendLineToLog(string.Format("\trating: {0}", rating));
//ReviewHtmlCrawler.form.appendLineToLog(string.Format("\ttitle: {0}", title));
//ReviewHtmlCrawler.form.appendLineToLog(string.Format("\tReviewer: {0} | {1}", reviewToAdd.reviewerId, reviewToAdd.reviewerName));
////badges not output
//ReviewHtmlCrawler.form.appendLineToLog(string.Format("\tPublished: {0}", publishedDate));
//ReviewHtmlCrawler.form.appendLineToLog(string.Format("\tContent: {0}", reviewContent.Substring(0, chars)));
//ReviewHtmlCrawler.form.appendLineToLog(string.Format("\tHelpful: {0} of {1}", reviewToAdd.numHelpful, reviewToAdd.numTotal));
//ReviewHtmlCrawler.form.appendLineToLog(string.Format("\tCommments: {0}", numComments));
//ReviewHtmlCrawler.form.appendLineToLog(string.Format("\tisVerifiedPurchase: {0}", isVerifiedPurchase));
//ReviewHtmlCrawler.form.appendLineToLog(string.Format("\tFormat: {0}", reviewToAdd.format));
try
{
context.SaveChanges();
numProcessed += 1;
}
catch (Exception ex)
{
ReviewHtmlCrawler.form.appendLineToLog(ex.Message);
ReviewHtmlCrawler.form.appendLineToLog(ex.StackTrace);
numNotProcessed += 1;
}
}
ReviewHtmlCrawler.form.appendLineToLog(string.Format("{0} reviews saved/updated on page; {1} not saved; {2} badges.", numProcessed, numNotProcessed, numBadges));
}
}
