glenux-archive/cloutier--bird.makeup
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

fix url parsing

Browse source
This commit is contained in:
Vincent Cloutier 2023-06-30 15:01:40 -04:00
parent 26e5036870
commit 3518c54277
4 changed files with 32 additions and 9 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
src/BirdsiteLive.Common/Regexes/HashtagRegexes.cs
View file

@ -5,6 +5,6 @@ namespace BirdsiteLive.Common.Regexes
public class HashtagRegexes
{
public static readonly Regex HashtagName = new Regex(@"^[a-zA-Z0-9_]+$");
public static readonly Regex Hashtag = new Regex(@"(.?)#([a-zA-Z0-9_]+)(\s|$|[\[\]<>.,;:!?/|-])");
public static readonly Regex Hashtag = new Regex(@"(^|.?[ \n]+)#([a-zA-Z0-9_]+)(?=\s|$|[\[\]<>.,;:!?/|-])");
}
}

2
src/BirdsiteLive.Common/Regexes/UserRegexes.cs
View file

@ -5,6 +5,6 @@ namespace BirdsiteLive.Common.Regexes
public class UserRegexes
{
public static readonly Regex TwitterAccount = new Regex(@"^[a-zA-Z0-9_]+$");
public static readonly Regex Mention = new Regex(@"(.?)@([a-zA-Z0-9_]+)(\s|$|[\[\]<>,;:'\.’!?/—\|-]|(. ))");
public static readonly Regex Mention = new Regex(@"(^|.?[ \n\.]+)@([a-zA-Z0-9_]+)(?=\s|$|[\[\]<>,;:'\.’!?/—\|-]|(. ))");
}
}

7
src/BirdsiteLive.Domain/Tools/StatusExtractor.cs
View file

@ -32,9 +32,6 @@ namespace BirdsiteLive.Domain.Tools
{
var tags = new List<Tag>();
// Replace return lines
messageContent = Regex.Replace(messageContent, @"\r\n\r\n?|\n\n", "</p><p>");
messageContent = Regex.Replace(messageContent, @"\r\n?|\n", "<br/>");
// Extract Urls
@ -125,6 +122,10 @@ namespace BirdsiteLive.Domain.Tools
}
}
// Replace return lines
messageContent = Regex.Replace(messageContent, @"\r\n\r\n?|\n\n", "</p><p>");
messageContent = Regex.Replace(messageContent, @"\r\n?|\n", "<br/>");
return (messageContent.Trim(), tags.ToArray());
}

30
src/Tests/BirdsiteLive.Domain.Tests/Tools/StatusExtractorTests.cs
View file

@ -111,7 +111,6 @@ namespace BirdsiteLive.Domain.Tests.Tools
Assert.IsTrue(result.content.Contains(@"<a href=""https://www.eff.org/deeplinks/2020/07/pact-act-not-solution-problem-harmful-online-content"" rel=""nofollow noopener noreferrer"" target=""_blank""><span class=""invisible"">https://www.</span><span class=""ellipsis"">eff.org/deeplinks/2020/07/pact</span><span class=""invisible"">-act-not-solution-problem-harmful-online-content</span></a>"));
#endregion
}
[Ignore]
[TestMethod]
public void Extract_FormatUrl_Long2_Test()
{
@ -128,7 +127,29 @@ namespace BirdsiteLive.Domain.Tests.Tools
#region Validations
logger.VerifyAll();
Assert.AreEqual(result.content, @"<a href=""https://twitterisgoinggreat.com/#twitters-first-dollar15bn-interest-payment-could-be-due-in-two-weeks"" rel=""nofollow noopener noreferrer"" target=""_blank""><span class=""invisible"">https://www.</span><span class=""ellipsis"">eff.org/deeplinks/2020/07/pact</span><span class=""invisible"">-act-not-solution-problem-harmful-online-content</span></a>");
Assert.AreEqual(result.content, @"<a href=""https://twitterisgoinggreat.com/#twitters-first-dollar15bn-interest-payment-could-be-due-in-two-weeks"" rel=""nofollow noopener noreferrer"" target=""_blank""><span class=""invisible"">https://</span><span class=""ellipsis"">twitterisgoinggreat.com/#twitt</span><span class=""invisible"">ers-first-dollar15bn-interest-payment-could-be-due-in-two-weeks</span></a>");
Assert.AreEqual(0, result.tags.Length);
#endregion
}
[TestMethod]
public void Extract_FormatUrl_Long3_Test()
{
#region Stubs
var message = $"https://domain.name/@WeekInEthNews/1668684659855880193";
#endregion
#region Mocks
var logger = new Mock<ILogger<StatusExtractor>>();
#endregion
var service = new StatusExtractor(_settings, logger.Object);
var result = service.Extract(message);
#region Validations
logger.VerifyAll();
Assert.AreEqual(result.content, @"<a href=""https://domain.name/@WeekInEthNews/1668684659855880193"" rel=""nofollow noopener noreferrer"" target=""_blank""><span class=""invisible"">https://</span><span class=""ellipsis"">domain.name/@WeekInEthNews/166</span><span class=""invisible"">8684659855880193</span></a>");
Assert.AreEqual(0, result.tags.Length);
#endregion
@ -633,7 +654,7 @@ namespace BirdsiteLive.Domain.Tests.Tools
public void Extract_Emoji_Test()
{
#region Stubs
var message = $"😤@mynickname 😎😍🤗🤩😘";
var message = $"😤 @mynickname 😎😍🤗🤩😘";
//var message = $"tests@mynickname";
#endregion
@ -648,12 +669,13 @@ namespace BirdsiteLive.Domain.Tests.Tools
logger.VerifyAll();
Assert.AreEqual(1, result.tags.Length);
Assert.IsTrue(result.content.Contains(
@"😤<span class=""h-card""><a href=""https://domain.name/users/mynickname"" class=""u-url mention"">@<span>mynickname</span></a></span>"));
@"😤 <span class=""h-card""><a href=""https://domain.name/users/mynickname"" class=""u-url mention"">@<span>mynickname</span></a></span>"));
Assert.IsTrue(result.content.Contains(@"😎😍🤗🤩😘"));
#endregion
}
[Ignore]
[TestMethod]
public void Extract_Parenthesis_Test()
{