Micronaut Dependency Finder Test

A sample application that uses the Dependency Finder library.

The application is configured in application.properties.

Running the App

The default configuration in application.properties expects to extract dependencies from ./lib/DependencyFinder-SNAPSHOT.jar. It also expects to load a dependency graph from ./df.xml. You can change these expectations by changing in the configuration in application.properties.

The app will run on port 8080.

In Local Shell

./gradlew run

In Docker

Building the Image

First, build the image. See the Micronaut Docker documentation for more options.

./gradlew dockerBuild

And tag the image with the name in the Docker repository.

docker image tag dependency-finder-micronaut jeantessier/dependency-finder-micronaut:1.4.1-rc1
docker image tag dependency-finder-micronaut jeantessier/dependency-finder-micronaut:latest

Run a Container

Once you have a Docker image, you can create a container.

The default configuration gets dependency graphs from local files. This is suitable when running the app outside Docker, but does not work anymore when inside a container.

Use --env to override the locations for the Extract and Load pages.

Use --volume to mount actual data at the locations you specified with --env.

This example maps the local files from the default configuration to /code in the container.

docker run \
  --name dependencyfinder \
  --detach \
  --publish "8080:8080" \
  --env dependency.finder.load.file=/code/df.xml \
  --env dependency.finder.extract.source=/code \
  --volume ./lib/DependencyFinder-SNAPSHOT.jar:/code/DependencyFinder.jar:ro \
  --volume ./df.xml:/code/df.xml:ro \
  jeantessier/dependency-finder-micronaut

To clean up Docker:

docker stop dependencyfinder
docker rm dependencyfinder
docker rmi dependency-finder-micronaut jeantessier/dependency-finder-micronaut

Docker Compose

Instead of creating a custom container, you can use Docker Compose to store all config in the docker-compose.yml and docker-compose.override.yml files.

The default configuration gets dependency graphs from local files. This is suitable when running the app outside Docker, but does not work anymore when inside a container.

Use environment to override the locations for the Extract and Load pages.

Use volumes to mount actual data at the locations you specified in environment.

The sample docker-compose.yml and docker-compose.override.yml files map the local files from the default configuration to /code in the container.

To run the container:

docker compose up -d

To stop the container:

docker compose stop

To clean up Docker:

docker compose down

Getting a Graph in Memory

You can either extract a graph from compiled Java code, or you can load a graph that has already been extracted.

To Extract a Graph From Code

List your JAR files in the dependency.finder.extract.source config.

A GET request to /extract returns details about the graph currently in memory.

http :8080/extract

{
  "extract": {
    "source": "lib/DependencyFinder-SNAPSHOT.jar",
    "filterIncludes": "//",
    "filterExcludes": "/^java/, /^org/"
  },
  "graph": {
    "name": "Dependency Finder",
    "label": null,
    "extractStart": "2024-09-29T23:30:35.728Z",
    "extractDurationInMillis": 548,
    "updateStart": null,
    "updateDurationInMillis": null,
    "loadStart": null,
    "loadDurationInMillis": null,
    "nbPackages": 23,
    "nbClasses": 691,
    "nbFeatures": 8108
  }
}

A POST request to /extract will extract a graph from the JAR files and redirect to /extract to get statistics

http --follow POST :8080/extract

{
  "extract": {
    "source": "lib/DependencyFinder-SNAPSHOT.jar",
    "filterIncludes": "//",
    "filterExcludes": "/^java/, /^org/"
  },
  "graph": {
    "name": "Dependency Finder",
    "label": null,
    "extractStart": "2024-09-29T23:30:35.728Z",
    "extractDurationInMillis": 548,
    "updateStart": null,
    "updateDurationInMillis": null,
    "loadStart": null,
    "loadDurationInMillis": null,
    "nbPackages": 23,
    "nbClasses": 691,
    "nbFeatures": 8108
  }
}

You can supply an optional label parameter that will be saved along with the graph. You can use this label to qualify the version of the code the graph is being extracted from.

To Update an Extracted Graph After Code Changes

If you make changes and recompile your code, you can update the graph through the /extract endpoint by adding an update parameter.

http --follow POST :8080/extract update=true

{
  "extract": {
    "source": "lib/DependencyFinder-SNAPSHOT.jar",
    "filterIncludes": "//",
    "filterExcludes": "/^java/, /^org/"
  },
  "graph": {
    "name": "Dependency Finder",
    "label": null,
    "extractStart": "2024-09-29T23:30:35.728Z",
    "extractDurationInMillis": 548,
    "updateStart": "2024-09-29T23:32:41.475Z",
    "updateDurationInMillis": 26,
    "loadStart": null,
    "loadDurationInMillis": null,
    "nbPackages": 23,
    "nbClasses": 691,
    "nbFeatures": 8108
  }
}

You can update the label associated with the graph with the label parameter. If you don't supply a label, it will be removed from the graph.

To Load a Graph From a File

List your XML files in the dependency.finder.load.file config.

A GET request to /load returns details about the graph currently in memory.

http :8080/load

{
  "load": {
    "file": "df.xml"
  },
  "graph": {
    "name": "Dependency Finder",
    "label": null,
    "extractStart": null,
    "extractDurationInMillis": null,
    "updateStart": null,
    "updateDurationInMillis": null,
    "loadStart": "2024-09-29T23:51:50.206Z",
    "loadDurationInMillis": 192,
    "nbPackages": 42,
    "nbClasses": 872,
    "nbFeatures": 8710
  }
}

A POST request to /load will load a graph from the XML files and redirect to /load to get statistics

http --follow POST :8080/load

{
  "load": {
    "file": "df.xml"
  },
  "graph": {
    "name": "Dependency Finder",
    "label": null,
    "extractStart": null,
    "extractDurationInMillis": null,
    "updateStart": null,
    "updateDurationInMillis": null,
    "loadStart": "2024-09-29T23:51:50.206Z",
    "loadDurationInMillis": 192,
    "nbPackages": 42,
    "nbClasses": 872,
    "nbFeatures": 8710
  }
}

You can supply an optional label parameter that will be saved along with the graph. You can use this label to qualify the version of the code the graph is being extracted from.

To Query a Graph

A POST request to /query will query the graph and return the resulting subgraph as a JSON object.

http :8080/query \
  packageScope:=true \
  packageFilter:=true \
  scopeIncludes=/^com.jeantessier/ \
  filterIncludes=/^com.jeantessier/ \
  showInbounds:=true \
  showEmptyNodes:=true

In this example, the request will look like:

{
    "filterIncludes": "/^com.jeantessier/",
    "packageFilter": true,
    "packageScope": true,
    "scopeIncludes": "/^com.jeantessier/",
    "showEmptyNodes": true,
    "showInbounds": true
}

And the response will be shaped like:

[
    {
        "type": "package",
        "confirmed": "true",
        "name": "com.jeantessier.classreader",
        "outbound": [
            {
                "type": "package",
                "confirmed": "true",
                "name": "com.jeantessier.classreader.impl"
            },
            {
                "type": "package",
                "confirmed": "true",
                "name": "com.jeantessier.text"
            }
        ],
        "inbound": [],
        "classes": []
    },
    ...
]

Micronaut 4.7.4 Documentation

• User Guide
• API Reference
• Configuration Reference
• Micronaut Guides

---

• Shadow Gradle Plugin
• Micronaut Gradle Plugin documentation

Feature serialization-jackson documentation

• Micronaut Serialization Jackson Core documentation

Feature micronaut-aot documentation

• Micronaut AOT documentation