extentreports

Overview

ExtentReports is a Java / .NET test reporting library - ExtentReports holds the run; pluggable reporters (ExtentSparkReporter for HTML, ExtentKlovReporter for the Klov server) consume the events and produce the artifact (extent-wiki).

Project status note (extent-readme): "ExtentReports is being sunset and will be replaced by ChainTest Framework." Teams already on ExtentReports v5 should plan a future migration; new projects should evaluate ChainTest first.

This skill covers the v5 API as documented in the official wiki.

When to use

• A JVM (Java / Kotlin / Scala) test suite needs richer reporting than JUnit XML - screenshots inline, log levels, hierarchical parent/child tests.
• A .NET suite uses the sister extentreports-dotnet project (same API shape).
• The team has invested in ExtentReports tooling (Klov server, custom themes) and isn't ready to migrate to ChainTest.

If the team is starting fresh, evaluate ChainTest (the announced successor per extent-readme) before adopting ExtentReports. For framework-agnostic richer reporting, allure-reports covers similar ground with broader language support.

How to use

1. Add the extentreports dependency (or extentreports-dotnet for .NET) and pin the version.
2. Create an ExtentReports, attach an ExtentSparkReporter("<path>"), and call extent.flush() at the end of the run - without flush() no file is written.
3. Per test, createTest(name) then chain log levels (info / pass / warning / skip / fail); attach failure screenshots via MediaEntityBuilder.
4. Wire the reporter into your JUnit 5 / TestNG lifecycle and upload the HTML as a CI artifact - the lifecycle extension, the richer report API (hierarchical tests, category / author / device labels, exception capture, code blocks), and CI upload are in references/api-and-ci-wiring.md.

Install

Maven:

<dependency>
  <groupId>com.aventstack</groupId>
  <artifactId>extentreports</artifactId>
  <version>5.1.2</version>
</dependency>

Per extent-readme, v5.1.2 is the latest release as of 2024-06-26. Pin a version explicitly; the project is in maintenance mode pending ChainTest.

Initialize the report

Per the extent-wiki complete example:

import com.aventstack.extentreports.ExtentReports;
import com.aventstack.extentreports.MediaEntityBuilder;
import com.aventstack.extentreports.Status;
import com.aventstack.extentreports.markuputils.MarkupHelper;
import com.aventstack.extentreports.reporter.ExtentSparkReporter;

public class Main {
    public static void main(String[] args) {
        ExtentReports extent = new ExtentReports();
        ExtentSparkReporter spark = new ExtentSparkReporter("target/Spark/Spark.html");
        extent.attachReporter(spark);
        // ... createTest calls ...
        extent.flush();
    }
}

Three load-bearing pieces:

1. new ExtentReports() - the run aggregator. Holds tests + metadata; emits to all attached reporters on flush().
2. new ExtentSparkReporter("<path>") - the HTML reporter. Path is the output file.
3. extent.attachReporter(spark) - wires reporter to the run.

extent.flush() at the end of the run writes the report to disk. Without flush(), no file is written - the most common new-user mistake.

Create tests and log levels

Per extent-wiki:

extent.createTest("LogLevels")
        .info("info")
        .pass("pass")
        .warning("warn")
        .skip("skip")
        .fail("fail");

The chain reads top-to-bottom in the report. info is neutral; pass / fail / skip set the test's overall status; warning surfaces a yellow flag without changing status.

Each createTest returns an ExtentTest; calls on it accumulate log entries. Multiple createTest calls on the same extent produce multiple test entries in the report.

Screenshots and media

Per extent-wiki:

extent.createTest("ScreenCapture")
        .addScreenCaptureFromPath("extent.png")
        .pass(MediaEntityBuilder.createScreenCaptureFromPath("extent.png").build());

Two patterns:

• .addScreenCaptureFromPath("...") - attaches the screenshot to the test as a top-level media entity.
• MediaEntityBuilder.createScreenCaptureFromPath("...").build() passed to a status method - attaches the screenshot to the specific log entry (the .pass(...) call here).

The latter pattern is preferred for failure screenshots: capture the screenshot inside the test framework's @AfterEach failure hook and pass it to .fail(...) so the failure log includes the visual evidence inline.

Worked example

A minimal runnable report with one passing test, the log-level chain, and a screenshot inlined on the pass entry:

import com.aventstack.extentreports.ExtentReports;
import com.aventstack.extentreports.MediaEntityBuilder;
import com.aventstack.extentreports.reporter.ExtentSparkReporter;

public class Demo {
    public static void main(String[] args) {
        ExtentReports extent = new ExtentReports();
        extent.attachReporter(new ExtentSparkReporter("target/Spark/Spark.html"));

        extent.createTest("checkout adds item")
              .info("navigated to cart")
              .pass("item added",
                    MediaEntityBuilder.createScreenCaptureFromPath("cart.png").build());

        extent.flush();   // writes target/Spark/Spark.html - omit it and nothing is written
    }
}

Open target/Spark/Spark.html in a browser: the test shows the info and pass log lines top-to-bottom with the screenshot inlined on the pass entry.

Operating in CI

ExtentReports HTML is a human artifact, not a CI gate - keep emitting JUnit XML for the gate and upload the Spark HTML alongside it. Register the reporter through a JUnit 5 TestWatcher (or TestNG listener) so every pass/fail is logged and flush() runs in the suite teardown, then upload target/Spark/ with if: always() (Extent matters most on failure runs). The lifecycle extension and the artifact-upload step are in references/api-and-ci-wiring.md.

Anti-patterns

Limitations

• Sunset. Per extent-readme, the framework is being replaced by ChainTest. v5 will receive bug fixes but not new features.
• HTML output is the artifact. Programmatic post-processing requires reading the extent.json companion file (auto-emitted) rather than parsing the HTML.
• No native PR-comment integration. The HTML is a CI artifact; posting a summary to the PR requires a separate step (e.g. parse extent.json, format markdown, post via gh pr comment).
• Klov server is a separate runtime with its own deploy / storage requirements. Most teams stick with file-based Spark.

References

• extent-readme - repo overview, latest version (5.1.2), sunset notice (replacing with ChainTest).
• extent-wiki - wiki home with the "complete example" code cited throughout this skill.
• references/api-and-ci-wiring.md - hierarchical tests, category / author / device labels, exception capture, code blocks, and the JUnit 5 / TestNG + CI wiring.
• allure-reports - framework-agnostic alternative with broader language support (and not sunset).
• junit-xml-analysis - pair with ExtentReports for CI gating; the JUnit XML is the gate, the Spark HTML is the human-readable artifact.