protected void Page_Load(object sender, EventArgs e)
{
string oldText = @"<p><i>This is</i> some sample text to <strong>demonstrate</strong> the capability of the <strong>HTML diff tool</strong>.</p>
<p>It is based on the <b>Ruby</b> implementation found <a href='http://github.com/myobie/htmldiff'>here</a>. Note how the link has no tooltip</p>
<p>What about a number change: 123456?</p>
<table cellpadding='0' cellspacing='0'>
<tr><td>Some sample text</td><td>Some sample value</td></tr>
<tr><td>Data 1 (this row will be removed)</td><td>Data 2</td></tr>
</table>
Here is a number 2 32";
string newText = @"<p>This is some sample <strong>text to</strong> demonstrate the awesome capabilities of the <strong>HTML <u>diff</u> tool</strong>.</p><br/><br/>Extra spacing here that was not here before.
<p>It is <i>based</i> on the Ruby implementation found <a title='Cool tooltip' href='http://github.com/myobie/htmldiff'>here</a>. Note how the link has a tooltip now and the HTML diff algorithm has preserved formatting.</p>
<p>What about a number change: 123356?</p>
<table cellpadding='0' cellspacing='0'>
<tr><td>Some sample <strong>bold text</strong></td><td>Some sample value</td></tr>
</table>
Here is a number 2 <sup>32</sup>";
HtmlDiff diffHelper = new HtmlDiff(oldText, newText);
litOldText.Text = oldText;
litNewText.Text = newText;
litDiffText.Text = diffHelper.Build();
}
private void ShowVsHistory()
{
HtmlContent html = repository.Fetch(moduleId);
ContentHistory history = new ContentHistory(historyGuid);
if (history.ContentGuid != html.ModuleGuid) { return; }
litCurrentHeading.Text = string.Format(Resource.CurrentVersionHeadingFormat,
DateTimeHelper.Format(html.LastModUtc, timeZone, "g", timeOffset));
if ((HtmlConfiguration.UseHtmlDiff)&&(highlightDiff))
{
HtmlDiff diffHelper = new HtmlDiff(history.ContentText, html.Body);
litCurrentVersion.Text = diffHelper.Build();
}
else
{
litCurrentVersion.Text = html.Body;
}
litHistoryHead.Text = string.Format(Resource.VersionAsOfHeadingFormat,
DateTimeHelper.Format(history.CreatedUtc, timeZone, "g", timeOffset));
litHistoryVersion.Text = history.ContentText;
string onClick = "top.window.LoadHistoryInEditor('" + historyGuid.ToString() + "'); return false;";
btnRestore.Attributes.Add("onclick", onClick);
}
private void ShowVsDraft()
{
HtmlContent html = repository.Fetch(moduleId);
if (html == null) { return; }
ContentWorkflow draftContent = ContentWorkflow.GetWorkInProgress(html.ModuleGuid);
if (draftContent.Guid != workflowGuid) { return; }
if ((HtmlConfiguration.UseHtmlDiff)&&(highlightDiff))
{
HtmlDiff diffHelper = new HtmlDiff(html.Body, draftContent.ContentText);
litCurrentVersion.Text = diffHelper.Build();
}
else
{
litCurrentVersion.Text = draftContent.ContentText;
}
litCurrentHeading.Text = string.Format(Resource.CurrentDraftHeadingFormat,
DateTimeHelper.Format(draftContent.RecentActionOn, timeZone, "g", timeOffset));
litHistoryHead.Text = string.Format(Resource.CurrentVersionHeadingFormat,
DateTimeHelper.Format(html.LastModUtc, timeZone, "g", timeOffset));
litHistoryVersion.Text = html.Body;
//string onClick = "top.window.LoadHistoryInEditor('" + historyGuid.ToString() + "'); return false;";
//btnRestore.Attributes.Add("onclick", onClick);
btnRestore.Visible = false;
}
private void PopulateControls()
{
if (moduleId == -1) { return; }
if (itemId == -1) { return; }
//if (module == null) { return; }
if (historyGuid == Guid.Empty) { return; }
Blog blog = new Blog(itemId);
if (blog.ModuleId != moduleId)
{
SiteUtils.RedirectToAccessDeniedPage(this);
return;
}
ContentHistory history = new ContentHistory(historyGuid);
if (history.ContentGuid != blog.BlogGuid) { return; }
litCurrentHeading.Text = string.Format(BlogResources.CurrentVersionHeadingFormat,
DateTimeHelper.Format(blog.LastModUtc, timeZone, "g", timeOffset));
if (BlogConfiguration.UseHtmlDiff)
{
HtmlDiff diffHelper = new HtmlDiff(history.ContentText, blog.Description);
litCurrentVersion.Text = diffHelper.Build();
}
else
{
litCurrentVersion.Text = blog.Description;
}
litHistoryHead.Text = string.Format(BlogResources.VersionAsOfHeadingFormat,
DateTimeHelper.Format(history.CreatedUtc, timeZone, "g", timeOffset));
litHistoryVersion.Text = history.ContentText;
string onClick = "top.window.LoadHistoryInEditor('" + historyGuid.ToString() + "'); return false;";
btnRestore.Attributes.Add("onclick", onClick);
}
public ActionResult Compare(string path, int first, int second)
{
var page = _repository.Get(new PagePath(path));
var id1 = Math.Min(first, second);
var id2 = Math.Max(first, second);
var diff1 = page.Revisions.First(k => k.Id == id1).HtmlBody;
var diff2 = page.Revisions.First(k => k.Id == id2).HtmlBody;
var differ = new HtmlDiff(diff1, diff2);
return Json(new
{
success = true,
content = differ.Build()
}, JsonRequestBehavior.AllowGet);
}
static void Main(string[] args)
{
Log.Logger = new LoggerConfiguration()
.ReadFrom.AppSettings()
.CreateLogger();
var log = Log.Logger.ForContext<Crufty.CruftyConsole>();
// get list of all courts to process
var client = new MongoClient(ConfigurationManager.AppSettings["MongoServer"]);
var database = client.GetDatabase("CourtCruft");
var collection = database.GetCollection<CourtWebsite>("CourtWebsites");
List<CourtWebsite> documents = new List<CourtWebsite>();
Task.Run(async () =>
{
documents = await collection.Find(new BsonDocument()).ToListAsync();
}).Wait();
// loop over list of courts and scrape them.
foreach (var courtWebsite in documents)
{
try
{
// wrap in error handling so we catch 404s and "network down"
ScrapingBrowser browser = new ScrapingBrowser();
WebPage homePage = browser.NavigateToPage(new Uri(courtWebsite.Url));
//PageWebForm form = homePage.FindFormById("sb_form");
//form["q"] = "scrapysharp";
//form.Method = HttpVerb.Get;
//WebPage resultsPage = form.Submit();
// if no xpath filter, just grab the whole page
var currentPage = (String.IsNullOrEmpty(courtWebsite.SelectionXPathString)) ?
homePage.Html.InnerHtml
:homePage.Html.SelectSingleNode(courtWebsite.SelectionXPathString).InnerHtml;
if (String.IsNullOrWhiteSpace(currentPage))
{
log.Error("Unable to find {XPath} in {Url}", courtWebsite.SelectionXPathString, courtWebsite.Url);
continue;
}
if (courtWebsite.NewPageHtml != currentPage)
{
HtmlDiff diffHelper = new HtmlDiff(courtWebsite.NewPageHtml, currentPage);
courtWebsite.DiffedHtml = diffHelper.Build().Insert(0, "<style>ins {background-color: #cfc;text-decoration: none;} del { color: #999; background-color:#FEC8C8;</style>");
courtWebsite.OldPageHtml = courtWebsite.NewPageHtml;
courtWebsite.NewPageHtml = currentPage;
courtWebsite.LastChangedDateTime = DateTime.Now;
courtWebsite.Checked = false;
log.Information("Changes found for {CourtName}", courtWebsite.CourtName);
}
courtWebsite.LastRunDateTime = DateTime.Now;
var filter = Builders<CourtWebsite>.Filter.Eq(s => s.Id, courtWebsite.Id);
Task.Run(async () =>
{
await collection.ReplaceOneAsync(filter, courtWebsite);
}).Wait();
log.Verbose("Successfully scraped site {CourtName}", courtWebsite.CourtName);
}
catch (WebException exception)
{
log.Error(exception, "Web Exception for {CourtName} (check the url) {ExceptionMessage} {url}", courtWebsite.CourtName, exception.Message, courtWebsite.Url);
}
catch (Exception exception)
{
log.With("CourtWebsite", courtWebsite)
.Error(exception, "General Exception for {CourtName} {ExceptionMessage}", courtWebsite.CourtName, exception.Message);
}
}
}
