Moved dataset to have access to benchmarks

via pip install.
leojklarner · Nov 8, 2023 · 2bf761a · 2bf761a
1 parent 191cb80
commit 2bf761a
Show file tree

Hide file tree

Showing 20 changed files with 11 additions and 160 deletions.
diff --git a/MANIFEST.in b/MANIFEST.in
@@ -0,0 +1,4 @@
+include README.md
+include .requirements/*
+recursive-include gauche/datasets *
+
diff --git a/data/property_prediction/petase_151_mutants.csv b/data/property_prediction/petase_151_mutants.csv
diff --git a/gauche/__init__.py b/gauche/__init__.py
@@ -2,4 +2,4 @@
 
 __all__ = ["SIGP", "NonTensorialInputs"]
 
-__version__ = "1.0.0"
+__version__ = "0.1.4"
diff --git a/gauche/dataloader/molprop_loader.py b/gauche/dataloader/molprop_loader.py
@@ -246,8 +246,7 @@ def load_benchmark(
                     os.path.abspath(__file__),
                     "..",
                     "..",
-                    "..",
-                    "data",
+                    "datasets",
                     "property_prediction",
                     benchmark + ".csv",
                 )

diff --git a/gauche/dataloader/reaction_loader.py b/gauche/dataloader/reaction_loader.py
@@ -249,8 +249,7 @@ def load_benchmark(
                     os.path.abspath(__file__),
                     "..",
                     "..",
-                    "..",
-                    "data",
+                    "datasets",
                     "reactions",
                     benchmark.removesuffix("RXN") + ".csv",
                 )

diff --git a/tests/__init__.py → gauche/datasets/__init__.py b/tests/__init__.py → gauche/datasets/__init__.py
diff --git a/data/property_prediction/ESOL.csv → gauche/datasets/property_prediction/ESOL.csv b/data/property_prediction/ESOL.csv → gauche/datasets/property_prediction/ESOL.csv
diff --git a/data/property_prediction/FreeSolv.csv → ...datasets/property_prediction/FreeSolv.csv b/data/property_prediction/FreeSolv.csv → ...datasets/property_prediction/FreeSolv.csv
diff --git a/data/property_prediction/Lipophilicity.csv → ...ets/property_prediction/Lipophilicity.csv b/data/property_prediction/Lipophilicity.csv → ...ets/property_prediction/Lipophilicity.csv
diff --git a/data/property_prediction/Photoswitch.csv → ...asets/property_prediction/Photoswitch.csv b/data/property_prediction/Photoswitch.csv → ...asets/property_prediction/Photoswitch.csv
diff --git a/tests/test_dataloaders/__init__.py → .../datasets/property_prediction/__init__.py b/tests/test_dataloaders/__init__.py → .../datasets/property_prediction/__init__.py
diff --git a/tests/test_kernels/__init__.py → gauche/datasets/proteins/__init__.py b/tests/test_kernels/__init__.py → gauche/datasets/proteins/__init__.py
diff --git a/data/proteins/petase_151_mutants.csv → .../datasets/proteins/petase_151_mutants.csv b/data/proteins/petase_151_mutants.csv → .../datasets/proteins/petase_151_mutants.csv
diff --git a/data/reactions/DreherDoyle.csv → gauche/datasets/reactions/DreherDoyle.csv b/data/reactions/DreherDoyle.csv → gauche/datasets/reactions/DreherDoyle.csv
diff --git a/data/reactions/SuzukiMiyaura.csv → gauche/datasets/reactions/SuzukiMiyaura.csv b/data/reactions/SuzukiMiyaura.csv → gauche/datasets/reactions/SuzukiMiyaura.csv
diff --git a/...nels/test_fingerprint_kernels/__init__.py → gauche/datasets/reactions/__init__.py b/...nels/test_fingerprint_kernels/__init__.py → gauche/datasets/reactions/__init__.py
diff --git a/pyproject.toml b/pyproject.toml
@@ -26,7 +26,7 @@ force_grid_wrap = 0
 include_trailing_comma = true
 line_length = 79 # match our custom config above
 multi_line_output = 3
-src_paths = ["gauche", "test"]
+src_paths = ["gauche"]
 float_to_top = true
 use_parentheses = true
 

diff --git a/setup.py b/setup.py
@@ -85,7 +85,6 @@ def find_version(*file_paths):
 readme = open("README.md").read()
 packages = find_packages(".", exclude=["tests"])
 
-
 setup(
     name="gauche",
     version=version,
@@ -95,6 +94,7 @@ def find_version(*file_paths):
     license="MIT",
     keywords="machine-learning gaussian-processes kernels pytorch chemistry biology protein ligand",
     packages=packages,
+    include_package_data=True,
     install_requires=INSTALL_REQUIRES,
     extras_require=EXTRA_REQUIRES,
     python_requires=">=3.8",

diff --git a/tests/test_dataloaders/test_molprop_loader.py b/tests/test_dataloaders/test_molprop_loader.py
@@ -56,7 +56,8 @@ def test_benchmark_loader(dataset, representation, kwargs):
             "..",
             "..",
             "..",
-            "data",
+            "gauche",
+            "datasets",
             "property_prediction",
         )
     )

diff --git a/tests/test_kernels/test_graph_kernels/__init__.py b/tests/test_kernels/test_graph_kernels/__init__.py
Original file line number	Diff line number	Diff line change
Expand Up		@@ -2,4 +2,4 @@

		__all__ = ["SIGP", "NonTensorialInputs"]

		__version__ = "1.0.0"
		__version__ = "0.1.4"
-Original file line number
+Diff line change
@@ Expand Up / @@ -56,7 +56,8 @@ def test_benchmark_loader(dataset, representation, kwargs): @@
                 "..",
                 "..",
                 "..",
-                "data",
+                "gauche",
+                "datasets",
                 "property_prediction",
             )
         )
@@ Expand Down @@