reimplement tests

.github/workflows/ci.yml

@@ -6,7 +6,7 @@ on:
       - master
 
 jobs:
-  default:
+  test-data-access:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout repository
@@ -20,10 +20,10 @@ jobs:
       - name: Install dependencies
         run: pip install -r requirements.txt
 
-      - name: Default test
-        run: python3 af_benchmark/benchmark.py tests/test-default.yaml
+      - name: Run data access tests
+        run: python3 tests/test-data-access.py
 
-  read-from-dir:
+  test-executors:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout repository
@@ -38,9 +38,9 @@ jobs:
         run: pip install -r requirements.txt
 
       - name: 
-        run: python3 af_benchmark/benchmark.py tests/test-read-from-dir.yaml
+        run: python3 tests/test-executors.py
 
-  futures-executor:
+  test-processing:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout repository
@@ -54,75 +54,6 @@ jobs:
       - name: Install dependencies
         run: pip install -r requirements.txt
 
-      - name: Test futures executor
-        run: python3 af_benchmark/benchmark.py tests/test-futures.yaml
-
-  dask-local-executor:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@v2
-
-      - name: Setup Python
-        uses: actions/setup-python@v2
-        with:
-          python-version: '3.10'
-
-      - name: Install dependencies
-        run: pip install -r requirements.txt
-
-      - name: Test Dask executor (local)
-        run: python3 af_benchmark/benchmark.py tests/test-dask-local.yaml
-
-  explicit-columns:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@v2
-
-      - name: Setup Python
-        uses: actions/setup-python@v2
-        with:
-          python-version: '3.10'
-
-      - name: Install dependencies
-        run: pip install -r requirements.txt
-
-      - name: Test explicit list of columns
-        run: python3 af_benchmark/benchmark.py tests/test-explicit-columns.yaml
-
-  read-by-file:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@v2
-
-      - name: Setup Python
-        uses: actions/setup-python@v2
-        with:
-          python-version: '3.10'
-
-      - name: Install dependencies
-        run: pip install -r requirements.txt
-
-      - name: Test parallelization over files
-        run: python3 af_benchmark/benchmark.py tests/test-read-by-file.yaml
-
-  read-by-column:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@v2
-
-      - name: Setup Python
-        uses: actions/setup-python@v2
-        with:
-          python-version: '3.10'
-
-      - name: Install dependencies
-        run: pip install -r requirements.txt
-
-      - name: Test parallelization over columns
-        run: python3 af_benchmark/benchmark.py tests/test-read-by-column.yaml
-
+      - name: Test processing
+        run: tests/test-processing.py
 

tests/test-explicit-columns.yaml → tests/config-default.yaml

@@ -1,12 +1,10 @@
-executor:
-  backend: sequential
 data-access:
   mode: explicit-files
   files: 
     - tests/data/nano_dimuon.root
+executor:
+  backend: sequential
 processor:
   parallelize_over: files
-  columns:
-    - event
-    - Muon_pt
+  columns: 5
   operation: sum

tests/test-data-access.py

@@ -0,0 +1,25 @@
+from utils import run_tests
+
+
+def test_data_access_explicit_files(b):
+    b.config["data-access"]["mode"] = "explicit-files"
+    b.config["data-access"]["files"] = ["tests/data/nano_dimuon.root"]
+    b.run()
+    print(f"Successfully tested accessing data via explicing list of files")
+
+def test_data_access_explicit_dirs(b):
+    b.config["data-access"]["mode"] = "explicit-dirs"
+    b.config["data-access"]["directories"] = ["tests/data/"]
+    b.run()
+    print(f"Successfully tested accessing data via explicing list of directories")
+
+
+if __name__=='__main__':
+    run_tests(
+        config="tests/config-default.yaml",
+        functions=[
+            test_data_access_explicit_files,
+            test_data_access_explicit_dirs,
+        ]
+    )
+

tests/test-executors.py

@@ -0,0 +1,31 @@
+from utils import run_tests
+
+
+def test_executor_sequential(b):
+    b.config["executor"]["backend"] = "sequential"
+    b.run()
+    print(f"Successfully tested sequential executor")
+
+def test_executor_futures(b):
+    b.config["executor"]["backend"] = "futures"
+    b.run()
+    print(f"Successfully tested futures executor")
+
+def test_executor_dask_local(b):
+    b.config["executor"]["backend"] = "dask-local"
+    b.config["executor"]["workers"] = 1
+    b.run()
+    print(f"Successfully tested dask-local executor")
+
+
+if __name__=='__main__':
+    run_tests(
+        config="tests/config-default.yaml",
+        functions=[
+            test_executor_sequential,
+            test_executor_futures,
+            test_executor_dask_local
+        ]
+    )
+
+

tests/test-processing.py

@@ -0,0 +1,73 @@
+from utils import run_tests
+
+
+def test_processor_columns_explicit(b):
+    b.config["processor"]["columns"] = ["Muon_pt", "Muon_eta"]
+    b.run()
+    print(f"Successfully tested processing explicit list of columns")
+
+def test_processor_columns_number(b):
+    b.config["processor"]["columns"] = 5
+    b.run()
+    print(f"Successfully tested processing given number of columns")
+
+def test_processor_collections(b):
+    b.config["processor"]["columns"] = []
+    b.config["processor"]["collections"] = ["Muon"]
+    b.run()
+    print(f"Successfully tested processing a collection of columns")
+
+def test_processor_operation_nothing(b):
+    b.config["processor"]["operation"] = "nothing"
+    b.run()
+    print(f"Successfully tested doing nothing to specified columns")
+
+def test_processor_operation_load(b):
+    b.config["processor"]["operation"] = "load_into_memory"
+    b.run()
+    print(f"Successfully tested loading specified columns into memory")
+
+def test_processor_parallelize_over_files(b):
+    b.config["processor"]["parallelize_over"] = "files"
+    b.config["executor"]["backend"] = "futures"
+    b.config["data-access"]["files"] = [
+        "tests/data/nano_dimuon.root",
+        "tests/data/nano_dimuon.root"
+    ]
+    b.run()
+    print(f"Successfully tested parallelization over files")
+
+def test_processor_parallelize_over_columns(b):
+    b.config["processor"]["parallelize_over"] = "columns"
+    b.config["executor"]["backend"] = "futures"
+    b.config["processor"]["columns"] = 2
+    b.run()
+    print(f"Successfully tested parallelization over files")
+
+def test_processor_parallelize_over_files_and_columns(b):
+    b.config["processor"]["parallelize_over"] = "files_and_columns"
+    b.config["executor"]["backend"] = "futures"
+    b.config["data-access"]["files"] = [
+        "tests/data/nano_dimuon.root",
+        "tests/data/nano_dimuon.root"
+    ]
+    b.config["processor"]["columns"] = 2
+    b.run()
+    print(f"Successfully tested parallelization over files and columns")
+
+
+if __name__=='__main__':
+    run_tests(
+        config="tests/config-default.yaml",
+        functions=[
+            test_processor_columns_explicit,
+            test_processor_columns_number,
+            test_processor_collections,
+            test_processor_operation_nothing,
+            test_processor_operation_load,
+            test_processor_parallelize_over_files,
+            test_processor_parallelize_over_columns,
+            test_processor_parallelize_over_files_and_columns,
+        ]
+    )
+

tests/utils.py

@@ -0,0 +1,13 @@
+import os, sys
+sys.path.append(os.getcwd()+"/af_benchmark")
+import copy
+from benchmark import Benchmark
+
+def run_tests(config, functions):
+    b = Benchmark(config)
+    for func in functions:
+        old_config = copy.deepcopy(b.config)
+        func(b)
+        b.reset()
+        b.config = old_config
+