Loosen dependency restriction (#7)

* Loosen dependency

* Move to new release scheme

* Change caching in CI

* Save data cache after running tests

* Don't use download tests in CI

.github/workflows/main.yml

@@ -12,13 +12,12 @@ on:
 
 
 jobs:
-  build:
-
+  not_slow:
     runs-on: ${{ matrix.os }}
     strategy:
       fail-fast: false
       matrix:
-        python-version: [3.8, "3.10"]
+        python-version: [3.8, "3.11"]
         os: [ubuntu-latest, macos-latest, windows-latest]
 
     steps:
@@ -33,12 +32,5 @@ jobs:
     - name: Install dependencies
       run: |
         python -m pip install nox_poetry
-    - name: Cache dataset downloads
-      uses: actions/cache@v3
-      with:
-        # datasets are stored in `~/.data/ea-dataset-provider/`
-        path: ~/.data/ea-dataset-provider/
-        key: alwaysthesame
     - name: Test with nox
-      run: nox -s tests
-
+      run: nox -s tests -- -m "not slow"

CHANGELOG.md

@@ -0,0 +1,16 @@
+# Changelog
+
+All notable changes to this project will be documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
+and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+## [Unreleased]
+
+## [0.1.1] - 2023-03-27
+
+### Fixed
+
+- Loosen python dependency
+
+[0.1.1]: https://github.com/dobraczka/sylloge/releases/tag/v1.0.0

noxfile.py

@@ -5,6 +5,7 @@
 def tests(session: Session) -> None:
     args = session.posargs or ["--cov", "--cov-report=xml"]
     session.install(".[all]")
+    session.install("strawman")
     session.install("pytest")
     session.install("pytest-cov")
     session.install("pytest-mock")

pyproject.toml

@@ -14,7 +14,7 @@ repository = "https://github.com/dobraczka/sylloge"
 "Documentation" = "https://sylloge.readthedocs.io"
 
 [tool.poetry.dependencies]
-python = ">=3.8,<3.11"
+python = ">=3.8,<4.0"
 Sphinx = {version = "^5.0.0", optional = true}
 insegel = {version = "^1.3.1", optional = true}
 pystow = "^0.4.6"
@@ -46,6 +46,9 @@ darglint = "^1.8.1"
 [tool.poetry.extras]
 docs = ["sphinx", "insegel", "sphinx-automodapi", "sphinx-autodoc-typehints"]
 
+[tool.poetry.group.dev.dependencies]
+strawman = "^0.1.2"
+
 [build-system]
 requires = ["poetry-core>=1.0.0"]
 build-backend = "poetry.core.masonry.api"
@@ -85,3 +88,7 @@ line_length = 88
 include_trailing_comma = true
 reverse_relative = true
 
+[tool.pytest.ini_options]
+markers = [
+    "slow: mark test as slow"
+]

tests/mocks.py

@@ -0,0 +1,57 @@
+from pathlib import Path
+from typing import Iterable, Union
+
+import pandas as pd
+from strawman import dummy_df, dummy_triples
+from util import DatasetStatistics
+
+from sylloge.base import EA_SIDE_LEFT, EA_SIDE_RIGHT, EA_SIDES
+
+
+class DataPathMocker:
+    def __init__(self, data_path):
+        self.data_path = data_path
+
+    def mock_data_path(self):
+        return self.data_path
+
+
+class ResourceMocker:
+    def __init__(self, statistic: DatasetStatistics, fraction: float):
+        self.statistic = statistic
+        self.fraction = fraction
+
+    def mock_read(self, path: str, names: Iterable):
+        return self.mock_read_zipfile_csv(path="", inner_path=path)
+
+    def mock_read_zipfile_csv(
+        self, path: Union[str, Path], inner_path: str, sep: str = "\t", **kwargs
+    ) -> pd.DataFrame:
+        if "rel_triples_1" in inner_path:
+            return dummy_triples(
+                int(self.statistic.num_rel_triples_left * self.fraction),
+                entity_prefix=EA_SIDE_LEFT,
+            )
+        elif "rel_triples_2" in inner_path:
+            return dummy_triples(
+                int(self.statistic.num_rel_triples_right * self.fraction),
+                entity_prefix=EA_SIDE_RIGHT,
+            )
+        elif "attr_triples_1" in inner_path:
+            return dummy_triples(
+                int(self.statistic.num_attr_triples_left * self.fraction),
+                entity_prefix=EA_SIDE_LEFT,
+                relation_triples=False,
+            )
+        elif "attr_triples_2" in inner_path:
+            return dummy_triples(
+                int(self.statistic.num_attr_triples_right * self.fraction),
+                entity_prefix=EA_SIDE_RIGHT,
+                relation_triples=False,
+            )
+        elif "_links" in inner_path:
+            return dummy_df(
+                shape=(int(self.statistic.num_ent_links * self.fraction), 2),
+                content_length=100,
+                columns=list(EA_SIDES),
+            )

tests/test_moviebenchmark.py

@@ -1,6 +1,7 @@
 from typing import Dict
 
 import pytest
+from mocks import DataPathMocker, ResourceMocker
 from util import DatasetStatistics
 
 from sylloge import MovieGraphBenchmark
@@ -40,6 +41,7 @@
 ]
 
 
+@pytest.mark.slow
 @pytest.mark.parametrize("params,statistic", statistics_with_params)
 def test_movie_benchmark(params: Dict, statistic: DatasetStatistics):
     ds = MovieGraphBenchmark(**params)
@@ -64,3 +66,26 @@ def test_movie_benchmark(params: Dict, statistic: DatasetStatistics):
             assert len(fold.train) == 496 or len(fold.train) == 497
             assert len(fold.test) == 1738
             assert len(fold.val) == 249 or len(fold.val) == 248
+
+
+@pytest.mark.parametrize("params,statistic", statistics_with_params)
+def test_movie_benchmark_mock(
+    params: Dict, statistic: DatasetStatistics, mocker, tmpdir
+):
+    rm = ResourceMocker(statistic=statistic, fraction=0.1)
+    dp_mocker = DataPathMocker(data_path=tmpdir)
+    mocker.patch("moviegraphbenchmark.loading._read", rm.mock_read)
+    mocker.patch(
+        "moviegraphbenchmark.create_graph._data_path", dp_mocker.mock_data_path
+    )
+    ds = MovieGraphBenchmark(**params)
+    assert ds.rel_triples_left is not None
+    assert ds.rel_triples_right is not None
+    assert ds.attr_triples_left is not None
+    assert ds.attr_triples_right is not None
+    assert ds.ent_links is not None
+    assert ds.folds is not None
+    for fold in ds.folds:
+        assert fold.train is not None
+        assert fold.test is not None
+        assert fold.val is not None

tests/test_open_ea.py

@@ -1,6 +1,7 @@
 from typing import Dict
 
 import pytest
+from mocks import ResourceMocker
 from util import DatasetStatistics
 
 from sylloge import OpenEA
@@ -170,6 +171,7 @@
 ]
 
 
+@pytest.mark.slow
 @pytest.mark.parametrize("params,statistic", statistics_with_params)
 def test_open_ea(params: Dict, statistic: DatasetStatistics):
     ds = OpenEA(**params)
@@ -179,6 +181,7 @@ def test_open_ea(params: Dict, statistic: DatasetStatistics):
     assert len(ds.attr_triples_right) == statistic.num_attr_triples_right
     assert len(ds.ent_links) == statistic.num_ent_links
     assert ds.folds is not None
+    assert len(ds.folds) == 5
     for fold in ds.folds:
         if params["size"] == SIZE_15K:
             assert len(fold.train) == 3000
@@ -188,3 +191,23 @@ def test_open_ea(params: Dict, statistic: DatasetStatistics):
             assert len(fold.train) == 20000
             assert len(fold.test) == 70000
             assert len(fold.val) == 10000
+
+
+@pytest.mark.parametrize("params,statistic", statistics_with_params)
+def test_open_ea_mock(params: Dict, statistic: DatasetStatistics, mocker):
+    left_name, right_name = params["graph_pair"].split("_")
+    fraction = 0.01 if params["size"] == SIZE_15K else 0.001
+    rm = ResourceMocker(statistic=statistic, fraction=fraction)
+    mocker.patch("sylloge.base.read_zipfile_csv", rm.mock_read_zipfile_csv)
+    ds = OpenEA(**params)
+    assert ds.rel_triples_left is not None
+    assert ds.rel_triples_right is not None
+    assert ds.attr_triples_left is not None
+    assert ds.attr_triples_right is not None
+    assert ds.ent_links is not None
+    assert ds.folds is not None
+    assert len(ds.folds) == 5
+    for fold in ds.folds:
+        assert fold.train is not None
+        assert fold.test is not None
+        assert fold.val is not None