Easy report generation

[karanjitsingh]

• Expose ReportConfigurationBuilder
  • Refactor code to be able to create ReportConfiguration from a dictionary
• Null check for RiskHotspotAnalysisResult in ReportBuilderBase
• Expose ReportContext ReportGenerator CachingFileReader IFileReader

Creating reports using the SDK becomes as simple as:

using ...

namespace htmlreport
{
    class Program
    {
        static void Main(string[] args)
        {
            var result = ParseCoverageFiles(new List<string>() { @"D:\CodeCoverageReport_8440364\coverage.xml" });
            CreateHTMLReport(result, DateTime.Now);
        }

        private static ParserResult ParseCoverageFiles(List<string> coverageFiles)
        {
            CoverageReportParser parser = new CoverageReportParser(
                1,
                new string[] { },
                new DefaultFilter(new string[] { }),
                new DefaultFilter(new string[] { }),
                new DefaultFilter(new string[] { }));

            ReadOnlyCollection<string> collection = new ReadOnlyCollection<string>(coverageFiles);
            return parser.ParseFiles(collection);
        }

        private static void CreateHTMLReport(ParserResult parserResult, DateTime executionTime)
        {
            var summaryResult = new SummaryResult(parserResult);
            var builder = new HtmlInlineAzurePipelinesReportBuilder();

            var folder = Path.Combine(Path.GetTempPath(), Guid.NewGuid().ToString());
            Directory.CreateDirectory(folder);

            var config = new ReportConfigurationBuilder().Create(new Dictionary<string, string>() {
                { "targetdir", folder },
                { "reporttypes", "HtmlInline_AzurePipelines" }
            });

            var reportContext = new ReportContext(config, new Settings());

            builder.ReportContext = reportContext;

            new ReportGenerator(
                new CachingFileReader(reportContext.Settings.CachingDuringOfRemoteFilesInMinutes),
                parserResult,
                new List<IReportBuilder>() { builder } )
                    .CreateReport(false, null, executionTime, reportContext.ReportConfiguration.Tag);
        }
    }
}

[karanjitsingh]

Gtihub diff seems weird

[danielpalme]

I think you can achieve the same result, with much less code and fewer changes in ReportGenerator.
Only the changes to ReportConfigurationBuilder are required:

The following sample demonstrates how to use the higher level APIs:

using ...

namespace htmlreport
{
    class Program
    {
        static void Main(string[] args)
        {
            CreateHTMLReport(new List<string>() { @"D:\CodeCoverageReport_8440364\coverage.xml" });
        }

        private static void CreateHTMLReport(IEnumerable<string> coverageFiles)
        {
            var folder = Path.Combine(Path.GetTempPath(), Guid.NewGuid().ToString());
            Directory.CreateDirectory(folder);

            var reportConfigurationBuilder = new ReportConfigurationBuilder();
            ReportConfiguration configuration = reportConfigurationBuilder.Create(new Dictionary<string, string>() {
                { "reports", string.Join(";", coverageFiles) }
                { "targetdir", folder },
                { "reporttypes", "HtmlInline_AzurePipelines" }
            });

            new Generator().GenerateReport(configuration);
        }
    }
}

What do you think?

[karanjitsingh]

My original idea was to modify Generator.cs itself. I wanted to separate parsing and report creation for scenarios where only parsing is required.

[danielpalme]

My recommendation is to:

• Use CoverageReportParser to parse coverage files and then do you own processing based on the ParserResult.
• Use the approach from above to generate HTML reports.

Don't use the (currently internal classes). If they become public and you start using them, it would become very hard for me to change the internal behavior without breaking your code.

[karanjitsingh]

@danielpalme, is there a way to generate HTML reports based on the ParserResult that I have? For example if I simply go through Generator it will parse the files for me instead of using my ParserResult instance.

My ask here is to decouple these two operations.

I can think of two ways of doing this without exposing internal classes,

• Separate Parsing and report generation in Generator itself, and GenerateReport calls into both of these to continue working the same way. This way Generator can expose methods for decoupled parsing and report generation as well.
• Another way of doing this is by having an overloaded implementation of GenerateReport to accept an instance of ParserResult and use that instead of parsing on its own. (I personally prefer this one)

It's possible in my scenario that first we parse the files and decide later on in time whether to create HTML reports or not, hence decoupling the two operations would make doing that simpler.

Let me know if this makes sense and I can raise a PR.

[danielpalme]

I will have a look at this tomorrow.

[danielpalme]

I added a two new overloads of GenerateReport:

ReportGenerator/src/ReportGenerator.Core/Generator.cs

Line 160 in c5a98de

public void GenerateReport(

ReportGenerator/src/ReportGenerator.Core/Generator.cs

Line 192 in c5a98de

public void GenerateReport(

This (pre-)release contains the new methods:
https://www.nuget.org/packages/ReportGenerator.Core/4.1.9-rc2