Adding Dataframe tests

pyproject.toml

@@ -1,9 +1,12 @@
 [build-system]
-requires = ["maturin>=1.0,<2.0", "polars>=1.3.0"]
+requires = ["poetry-core>=1.0.0", "maturin>=1.0,<2.0", "polars>=1.3.0"]
 build-backend = "maturin"
 
 [project]
 name = "polars-bio"
+version = "0.1.0"
+description = "Blazing fast genomic operations on large Python dataframes"
+authors = []
 requires-python = ">=3.8"
 classifiers = [
   "Programming Language :: Rust",
@@ -17,3 +20,26 @@ module-name = "polars_bio"
 [[tool.mypy.overrides]]
 module = "polars.utils.udfs"
 ignore_missing_imports = true
+
+[tool.poetry]
+name = "polars-bio"
+version = "0.1.0"
+description = ""
+authors = ["Marek Wiewiórka <[email protected]>"]
+readme = "README.md"
+
+[tool.poetry.dependencies]
+python = ">=3.9,<3.13"
+polars = "^1.16.0"
+pandas = "^2.2.3"
+
+
+[tool.poetry.dev-dependencies]
+pytest = "^8.3.3"
+pytest-cov = "^6.0.0"
+pre-commit = "^4.0.1"
+jupyter = "^1.1.0"
+ruff = "^0.8.2"
+maturin = "^1.7.5"
+bioframe = "^0.7.2"
+mdpd = "^0.2.1"

tests/_expected.py

@@ -0,0 +1,42 @@
+import mdpd
+from pathlib import Path
+
+import pandas as pd
+
+TEST_DIR = Path(__file__).parent
+DATA_DIR = TEST_DIR / "data"
+EXPECTED_OVERLAP = """
+        "+--------+-----------+---------+--------+-----------+---------+",
+        "| contig_1 | pos_start_1 | pos_end_1 | contig_2 | pos_start_2 | pos_end_2 |",
+        "+--------+-----------+---------+--------+-----------+---------+",
+        "| chr1   | 150       | 250     | chr1   | 100       | 190     |",
+        "| chr1   | 150       | 250     | chr1   | 200       | 290     |",
+        "| chr1   | 190       | 300     | chr1   | 100       | 190     |",
+        "| chr1   | 190       | 300     | chr1   | 200       | 290     |",
+        "| chr1   | 300       | 501     | chr1   | 400       | 600     |",
+        "| chr1   | 500       | 700     | chr1   | 400       | 600     |",
+        "| chr1   | 15000     | 15000   | chr1   | 10000     | 20000   |",
+        "| chr1   | 22000     | 22300   | chr1   | 22100     | 22100   |",
+        "| chr2   | 150       | 250     | chr2   | 100       | 190     |",
+        "| chr2   | 150       | 250     | chr2   | 200       | 290     |",
+        "| chr2   | 190       | 300     | chr2   | 100       | 190     |",
+        "| chr2   | 190       | 300     | chr2   | 200       | 290     |",
+        "| chr2   | 300       | 500     | chr2   | 400       | 600     |",
+        "| chr2   | 500       | 700     | chr2   | 400       | 600     |",
+        "| chr2   | 15000     | 15000   | chr2   | 10000     | 20000   |",
+        "| chr2   | 22000     | 22300   | chr2   | 22100     | 22100   |",
+        "+--------+-----------+---------+--------+-----------+---------+",
+"""
+
+DF_OVERLAP = (mdpd.from_md(EXPECTED_OVERLAP)
+              .astype({'pos_start_1': 'int64'})
+              .astype({'pos_end_1': 'int64'})
+              .astype({'pos_start_2': 'int64'})
+              .astype({'pos_end_2': 'int64'}))
+
+DF_OVERLAP = DF_OVERLAP.sort_values(by=list(DF_OVERLAP.columns)).reset_index(drop=True)
+
+
+
+DF1 = pd.read_csv(f"{DATA_DIR}/reads.csv")
+DF2 = pd.read_csv(f"{DATA_DIR}/targets.csv")

tests/data/reads.csv

@@ -0,0 +1,13 @@
+contig,pos_start,pos_end
+chr1,150,250
+chr1,190,300
+chr1,300,501
+chr1,500,700
+chr1,22000,22300
+chr1,15000,15000
+chr2,150,250
+chr2,190,300
+chr2,300,500
+chr2,500,700
+chr2,22000,22300
+chr2,15000,15000

tests/data/targets.csv

@@ -0,0 +1,11 @@
+contig,pos_start,pos_end
+chr1,100,190
+chr1,200,290
+chr1,400,600
+chr1,10000,20000
+chr1,22100,22100
+chr2,100,190
+chr2,200,290
+chr2,400,600
+chr2,10000,20000
+chr2,22100,22100

tests/test_pandas_overlap.py

@@ -0,0 +1,23 @@
+import pandas as pd
+
+import polars_bio as pb
+from _expected import DF_OVERLAP, DF1, DF2
+
+
+
+
+
+
+class TestOverlapPandas:
+    def test_overlap_count(self):
+        assert len(pb.overlap(DF1, DF2, output_type="pandas.DataFrame")) == 16
+
+    def test_overlap_schema_rows(self):
+        result = pb.overlap(DF1, DF2, output_type="pandas.DataFrame")
+        result = result.sort_values(by=list(result.columns)).reset_index(drop=True)
+        expected = DF_OVERLAP
+        pd.testing.assert_frame_equal(result, expected)
+
+
+
+

tests/test_polars_overlap.py

@@ -0,0 +1,4 @@
+import pandas as pd
+import polars_bio.overlap as overlap
+
+