Add test workflows

.github/workflows/run_testsuite_manual.yml

@@ -0,0 +1,36 @@
+name: Run Testsuite (manual) # Unfortunately, we cannot modify the name: https://github.community/t/github-actions-dynamic-name-of-the-workflow-with-workflow-dispatch/150327
+on:
+  workflow_dispatch:
+    inputs:
+      suites:
+        description: 'Comma seperated testsuites to execute'
+        required: true
+        type: string
+      build_args:
+        description: 'Build arguments, if not specified defaults will be taken'
+        required: false
+        type: string
+      systests_branch:
+        description: 'Branch to take the systest from'
+        default: 'develop'
+        required: true
+        type: string
+      loglevel:
+        description: 'loglevel used for the systemtests'
+        default: 'INFO'
+        required: true
+        type: choice
+        options:
+            - 'DEBUG'
+            - 'INFO'
+            - 'WARNING'
+            - 'ERROR'
+            - 'CRITICAL'
+jobs:
+  run_testsuite_manual:
+    uses: ./.github/workflows/run_testsuite_workflow.yml
+    with: 
+      suites: ${{ inputs.suites }}
+      build_args: ${{ inputs.build_args }}
+      systests_branch: ${{ inputs.systests_branch }}
+      loglevel: ${{ inputs.loglevel }}

.github/workflows/run_testsuite_workflow.yml

@@ -0,0 +1,63 @@
+name: Run Testsuite Workflow 
+on:
+  workflow_call:
+    inputs:
+      suites:
+        description: 'Comma seperated testsuites to execute'
+        required: true
+        type: string
+      build_args:
+        description: 'Build arguments'
+        required: false
+        type: string
+      systests_branch:
+        description: 'Branch to take the systest from'
+        default: 'develop'
+        required: true
+        type: string
+      loglevel:
+        description: 'loglevel used for the systemtests'
+        default: 'INFO'
+        required: false
+        type: string
+jobs:
+  run_testsuite:
+    runs-on: ubuntu-latest
+    steps:
+    - name: Report log
+      run: |
+        echo "Initiated by: ${{ github.actor }}"
+        echo "Running systemtests --build_args=${{github.event.inputs.build_args}} --suites=${{github.event.inputs.suites}}"
+        echo "Systemtests branch: ${{ github.event.inputs.systests_branch }}"
+    - name: Check out Tutorials for systest
+      uses: actions/checkout@v4
+      with:
+        ref: ${{ github.event.inputs.systests_branch }}
+        lfs: true
+        fetch-depth: 0
+    - name: Log directory
+      run: |
+        ls -al
+        pwd
+    - uses: actions/setup-python@v4
+      with:
+        python-version: '3.11'
+    - name: Set up Docker Buildx
+      uses: docker/setup-buildx-action@v2
+    - name: install python depencies
+      run: |
+        python -m pip install -r ./tools/tests/requirements.txt
+    - name: Expose GitHub Runtime (needed for accessing the gha cache)
+      uses: crazy-max/ghaction-github-runtime@v3
+    - name: Run tests
+      run: |
+        cd ./tools/tests
+        python systemtests.py --build_args=${{github.event.inputs.build_args}} --suites=${{github.event.inputs.suites}} --log-level=${{github.event.inputs.loglevel}} 
+        cd ../../
+    - name: Archive run files
+      if: success() || failure()
+      uses: actions/upload-artifact@v3
+      with:
+        name: runs
+        path: |
+          runs/*

tools/tests/README.md

@@ -19,6 +19,22 @@ python3 systemtests.py --suites=openfoam-adapter-release,<someothersuite>
 To discover all tests, use `python print_test_suites.py`.
 To be able to fill in the right case tuple into the `tests.yaml`, you can use the `python3 print_case_combinations.py` script.
 
+## Adding new tests
+
+### Adding tutorials
+
+In order for the systemtests to pick up the tutorial we need to define a `metadata.yaml` in the folder of the tutorial. There are a few `metadata.yaml` already present to get inspiration from. You can also have a look at the implementation details but normally the currently available ones should be easy to adopt. You can check your metadata parsing by `python print_metadata.py` and `python print_case_combinations.py`
+
+### Adding Testsuites
+
+To add a testsuite just open the `tests.yaml` file and use the output of `python print_case_combinations.py` to add the right case combinations you want to test. Note that you can specify a `reference_result` which is not yet present. The `generate_reference_data.py` will pick that up and create it for you.
+Note that its important to carefully check the paths of the `reference_result` in order to not have typos in there. Also note that same cases in different testsuites should use the same `reference_result`.
+
+### Generate reference results
+
+Since we need data to compare against, you need to run `python generate_reference_data.py`. This process might take a while.
+Please include the generated reference results in the pull request as they are strongly connected to the new testsuites.
+
 ## Implementation details
 
 Each tutorial contains automation scripts (mainly `run.sh` and `clean.sh`), as well as metadata (`metadata.yaml`). The metadata file describes the available cases, how to run them, as well as their dependencies. A central `tests.yaml` file in this directory defines test suites, which execute different combinations of cases. The Python script `systemtests.py` executes the tests, allowing to filter for specific components or test suites.

tools/tests/component-templates/calculix-adapter.yaml

@@ -5,6 +5,10 @@ build:
       - {{key}}={{value}}
     {% endfor %}
   target: calculix_adapter
+  cache_from:
+  - type=gha
+  cache_to:
+  - type=gha,mode=min,scope=calculix_adapter
 depends_on:    
   prepare:
     condition: service_completed_successfully

tools/tests/component-templates/fenics-adapter.yaml

@@ -5,6 +5,10 @@ build:
       - {{key}}={{value}}
     {% endfor %}
   target: fenics_adapter
+  cache_from:
+  - type=gha
+  cache_to:
+  - type=gha,mode=min,scope=fenics_adapter
 depends_on:    
   prepare:
     condition: service_completed_successfully

tools/tests/component-templates/nutils-adapter.yaml

@@ -5,6 +5,10 @@ build:
       - {{key}}={{value}}
     {% endfor %}
   target: nutils_adapter
+  cache_from:
+  - type=gha
+  cache_to:
+  - type=gha,mode=min,scope=nutils_adapter
 depends_on:    
   prepare:
     condition: service_completed_successfully

tools/tests/component-templates/openfoam-adapter.yaml

@@ -5,6 +5,10 @@ build:
       - {{key}}={{value}}
     {% endfor %}
   target: openfoam_adapter
+  cache_from:
+  - type=gha
+  cache_to:
+  - type=gha,mode=min,scope=openfoam_adapter
 depends_on:    
   prepare:
     condition: service_completed_successfully

tools/tests/docker-compose.field_compare.template.yaml

@@ -7,3 +7,7 @@ services:
     command:
       - /runs/{{ tutorial_folder }}/{{ precice_output_folder }}
       - /runs/{{ tutorial_folder }}/{{ reference_output_folder }}
+      - "-rtol 3e-7"
+
+# Currently its really hard to estimate the impact of compiling and executing in a different platform (like github actions)
+# 3e-7 might not be the thightest we can afford and we want to have but its an okayish guestimation for now. 

tools/tests/docker-compose.template.yaml

@@ -4,9 +4,16 @@ services:
     build: 
       context: {{ dockerfile_context }}
       target: base_image
+      cache_from:
+      - type=gha
+      cache_to:
+      - type=gha,mode=min,scope=prepare
+      args:
+        {% for key, value in build_arguments.items() %}
+          - {{key}}={{value}}
+        {% endfor %}
     volumes:
       - {{ run_directory }}:/runs
-
     command: >
       /bin/bash -c "id &&
       cd '/runs/{{ tutorial_folder }}' && 
@@ -18,4 +25,4 @@ services:
   {% for service in services %}
   {{ service }}:
     {{ services[service] |indent(4) }}
-  {% endfor %}
+  {% endfor %}

tools/tests/dockerfiles/ubuntu_2204/Dockerfile

@@ -2,8 +2,13 @@ FROM ubuntu:22.04 as base_image
 USER root
 SHELL ["/bin/bash", "-c"]
 ENV DEBIAN_FRONTEND=noninteractive
-
-RUN useradd -ms /bin/bash precice 
+# We set a sensical value, but still have the possibilty to influence this via the build time arguments. 
+# When the dockerfile is built using the systemtests.py we set the PRECICE_UID and PRECICE_GID to the user executing the systemtests.
+# This ensures no file ownership problems down the line and is the most easy fix, as we normally built the containers locally
+# If not built via the systemtests.py its either possible to specify manually but 1000 would be the default anyway.
+ARG PRECICE_UID=1000
+ARG PRECICE_GID=1000
+RUN groupadd -g ${PRECICE_GID} precice && useradd -u ${PRECICE_UID} -g ${PRECICE_GID} -ms /bin/bash precice
 ENV PATH="${PATH}:/home/precice/.local/bin" 
 ENV LD_LIBRARY_PATH="/home/precice/.local/lib:${LD_LIBRARY_PATH}" 
 ENV CPATH="/home/precice/.local/include:$CPATH"

tools/tests/systemtests/Systemtest.py

@@ -16,6 +16,7 @@
 import unicodedata
 import re
 import logging
+import os
 
 
 GLOBAL_TIMEOUT = 360
@@ -194,6 +195,7 @@ def __get_docker_compose_file(self):
             'tutorial_folder': self.tutorial_folder,
             'tutorial': self.tutorial.path.name,
             'services': rendered_services,
+            'build_arguments': self.params_to_use,
             'dockerfile_context': self.dockerfile_context,
             'precice_output_folder': PRECICE_REL_OUTPUT_DIR,
         }
@@ -217,7 +219,7 @@ def _get_git_ref(self, repository: Path, abbrev_ref=False) -> Optional[str]:
         try:
             result = subprocess.run([
                 "git",
-                "-C", repository.resolve(),
+                "-C", os.fspath(repository.resolve()),
                 "rev-parse",
                 "--abbrev-ref" if abbrev_ref else
                 "HEAD"], stdout=subprocess.PIPE,
@@ -227,13 +229,28 @@ def _get_git_ref(self, repository: Path, abbrev_ref=False) -> Optional[str]:
         except Exception as e:
             raise RuntimeError(f"An error occurred while getting the current Git ref: {e}") from e
 
+    def _fetch_ref(self, repository: Path, ref: str):
+        try:
+            result = subprocess.run([
+                "git",
+                "-C", os.fspath(repository.resolve()),
+                "fetch",
+                "origin",
+                f"{ref}:{ref}"
+            ], check=True, timeout=60)
+            if result.returncode != 0:
+                raise RuntimeError(f"git command returned code {result.returncode}")
+
+        except Exception as e:
+            raise RuntimeError(f"An error occurred while fetching origin '{ref}':  {e}")
+
     def _checkout_ref_in_subfolder(self, repository: Path, subfolder: Path, ref: str):
         try:
             result = subprocess.run([
                 "git",
-                "-C", repository.resolve(),
+                "-C", os.fspath(repository.resolve()),
                 "checkout", ref,
-                "--", subfolder.resolve()
+                "--", os.fspath(subfolder.resolve())
             ], check=True, timeout=60)
             if result.returncode != 0:
                 raise RuntimeError(f"git command returned code {result.returncode}")
@@ -251,6 +268,7 @@ def __copy_tutorial_into_directory(self, run_directory: Path):
         ref_requested = self.params_to_use.get("TUTORIALS_REF")
         if ref_requested:
             logging.debug(f"Checking out tutorials {ref_requested} before copying")
+            self._fetch_ref(PRECICE_TUTORIAL_DIR, ref_requested)
             self._checkout_ref_in_subfolder(PRECICE_TUTORIAL_DIR, self.tutorial.path, ref_requested)
 
         self.tutorial_folder = slugify(f'{self.tutorial.path.name}_{self.case_combination.cases}_{current_time_string}')
@@ -461,10 +479,9 @@ def __prepare_for_run(self, run_directory: Path):
         self.__copy_tutorial_into_directory(run_directory)
         self.__copy_tools(run_directory)
         self.__put_gitignore(run_directory)
-        uid, gid = self.__get_uid_gid()
-        self.env["UID"] = uid
-        self.env["GID"] = gid
-        self.__write_env_file()
+        host_uid, host_gid = self.__get_uid_gid()
+        self.params_to_use['PRECICE_UID'] = host_uid
+        self.params_to_use['PRECICE_GID'] = host_gid
 
     def run(self, run_directory: Path):
         """