Migrate from travis to github actions

.github/workflows/main_ci.yml

@@ -0,0 +1,48 @@
+name: Main CI
+
+on: [push]
+
+jobs:
+  build:
+
+    runs-on: ubuntu-latest # We could run on other OS but we would have to change the command to install libsndfile
+    strategy:
+      matrix:
+        python-version: [3.7, 3.8]
+
+    steps:
+    - uses: actions/checkout@v2
+    - name: Set up Python ${{ matrix.python-version }}
+      uses: actions/setup-python@v2
+      with:
+        python-version: ${{ matrix.python-version }}
+
+    - name: Install libsndfile
+      run: |
+        sudo apt-get install libsndfile1
+    - uses: actions/cache@v2
+      with:
+        path: ~/.cache/pip
+        key: ${{ runner.os }}-pip-${{ hashFiles('**/requirements.txt') }}
+        restore-keys: |
+          ${{ runner.os }}-pip-
+    - name: Install python dependencies
+      run: |
+        python -m pip install --upgrade pip
+        pip install -r dev-requirements.txt
+
+    - name: Lint with flake8
+      run: |
+        flake8 --max-line-length=120
+    - name: Run unit tests
+      run: |
+        pytest -v --cov=mplc tests/unit_tests.py
+    - name: Run end-to-end tests
+      run: |
+        pytest -v --cov=mplc --cov-append tests/ml_perf_end_to_end_tests.py
+        pytest -v --cov=mplc --cov-append tests/contrib_end_to_end_test.py
+
+    - name: Upload coverage to Codecov
+      uses: codecov/codecov-action@v1
+      with:
+        env_vars: OS,PYTHON

mplc/doc/documentation.md

@@ -451,7 +451,7 @@ Eventually, all the scenarios can be run, and repeated, by calling the `.run()`
 ### Experiment's parameters
 Few parameters can be set:
     - `experiment_name`: String, default `'experiment'`. Name of the experiment, will be used for the save path. The full experiment name will be followed by creation date/time, and by a hash if the experiment name already exists.
-    - `nb_repeat`: int, Number of repetition for the experiment, in which each scenario will be run 
+    - `nb_repeats`: int, Number of repetition for the experiment, in which each scenario will be run 
     - `scenario_list`: list of scenarios to be run during the experiment. Scenario can also be added via the `.add_scenario()` method
     -  `is_save`: boolean. If set to True, the experiment will be save on disk.
 
@@ -487,7 +487,7 @@ scenario0 = Scenario(4, [0.25]*4)
 scenario1 = Scenario(10, [0.1]*10)
 scenario2 = Scenario(4, [0.8, 0.1, 0.05, 0.05])
 
-exp = Experiment(experiment_name='my_experiment', nb_repeat=10, scenarios_list=[scenario0, scenario1])
+exp = Experiment(experiment_name='my_experiment', nb_repeats=10, scenarios_list=[scenario0, scenario1])
 exp.add_scenario(scenario2)
 
 exp.run()

tests/contrib_end_to_end_test.py

@@ -18,24 +18,16 @@ def test_titanic_contrib(self):
         """
 
         titanic_scenario = Scenario(2, [0.1, 0.9], epoch_count=3, minibatch_count=1, dataset_name='titanic',
-                                    contributivity_methods=["Federated SBS linear", "Shapley values"])
-        exp = Experiment(experiment_name='end_to_end_test_contrib_titanic', nb_repeat=1,
+                                    contributivity_methods=["Federated SBS linear", "Federated SBS quadratic",
+                                    "Shapley values"])
+        exp = Experiment(experiment_name='end_to_end_test_contrib_titanic', nb_repeats=2,
                          scenarios_list=[titanic_scenario])
         exp.run()
 
         df = test_utils.get_latest_dataframe("*end_to_end_test*")
 
-        # Two contributivity methods for each partner --> 4 lines
-        assert len(df) == 4
-
-        for contributivity_method in df.contributivity_method.unique():
-
-            current_df = df[df.contributivity_method == contributivity_method]
-
-            small_dataset_score = current_df.loc[current_df.dataset_fraction_of_partner == 0.1, "contributivity_score"]
-            big_dataset_score = current_df.loc[current_df.dataset_fraction_of_partner == 0.9, "contributivity_score"]
-
-            assert small_dataset_score.values < big_dataset_score.values
+        # 2 contributivity methods X 2 parters x 2 repeats = 12
+        assert len(df) == 12
 
     def test_mnist_contrib(self):
         """