Update Docker Image

* Explain baseline/models folder
* Reduce Docker Image Size
* Build ardoco-cli in docker image

Author: dfuchss

Dockerfile

@@ -41,9 +41,7 @@ RUN . venv/bin/activate && \
 
 # Copy FASTTEXT Models
 WORKDIR /replication/baselines/models
-ADD https://dl.fbaipublicfiles.com/fasttext/vectors-crawl/cc.en.300.bin.gz .
-ADD https://dl.fbaipublicfiles.com/fasttext/vectors-crawl/cc.it.300.bin.gz .
-RUN gunzip cc.en.300.bin.gz && gunzip cc.it.300.bin.gz
+RUN curl -O https://dl.fbaipublicfiles.com/fasttext/vectors-crawl/cc.en.300.bin.gz && curl -O https://dl.fbaipublicfiles.com/fasttext/vectors-crawl/cc.it.300.bin.gz && gunzip cc.en.300.bin.gz && gunzip cc.it.300.bin.gz
 
 # Copy CodeBERT Models
 RUN git lfs install && git clone https://huggingface.co/kit-mcse/CodeBERT-Java && rm -r CodeBERT-Java/.git
@@ -77,5 +75,7 @@ RUN echo "Cloning Source Code of Benchmarks" && \
 RUN cd ardoco+arcotl && mvn -P tlr clean test -Dsurefire.failIfNoSpecifiedTests=false -Dtest=TraceLinkEvaluationIT
 RUN cd baselines/TAROT && mvn -B compile exec:java -Dexec.mainClass="Start"
 
-ENTRYPOINT [ "/bin/bash", "-c", "cat README.md && bash" ]
+# Build ArDoCo CLI
+RUN cd ardoco+arcotl && mvn -P tlr package -DskipTests -DskipITs && cp cli/target/ardoco-cli.jar ardoco-cli.jar && mvn -P tlr clean
 
+ENTRYPOINT [ "/bin/bash", "-c", "cat README.md && bash" ]

README.md

@@ -325,6 +325,7 @@ In the folder [ardoco+arcotl/cli](./ardoco+arcotl/cli), there is the source code
 You can easily compile the CLI by using the command `mvn package` within the directory [ardoco+arcotl](./ardoco+arcotl). 
 To speed up the build, you can skip tests using the following command: `mvn package -DskipTests -DskipITs`. 
 The compiled CLI can then be found in the folder [ardoco+arcotl/cli/target/](./ardoco+arcotl/cli/target/) as `ardoco-cli.jar`.
+Note: If you are using the provided Docker image, the CLI is already compiled and can be found at `/replication/ardoco+arcotl/ardoco-cli.jar`.
 
 You can call the CLI to execute the evaluation with the following command, where OUT is the path to a folder where the output should be written to: <br>
 `java -jar ardoco-cli.jar -e -o OUT`

baselines/models/README.txt

@@ -1 +1,5 @@
-Place your models here.
+# baselines/models folder
+You can place the ML models here.
+If you use the docker image, you will find all needed models in this directory.
+Otherwise, follow the steps in the top-level readme to add the ML models here.
+