bed-reader 1.0.0b3__tar.gz → 1.0.0b4__tar.gz

{bed_reader-1.0.0b3 → bed_reader-1.0.0b4}/Cargo.lock

@@ -86,7 +86,7 @@ checksum = "904dfeac50f3cdaba28fc6f57fdcddb75f49ed61346676a78c4ffe55877802fd"
 
 [[package]]
 name = "bed-reader"
-version = "1.0.0-beta.3"
+version = "1.0.0-beta.4"
 dependencies = [
  "anyinput",
  "approx 0.5.1",

{bed_reader-1.0.0b3 → bed_reader-1.0.0b4}/Cargo.toml

@@ -1,7 +1,7 @@
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [package]
-version = "1.0.0-beta.3"
+version = "1.0.0-beta.4"
 name = "bed-reader"
 description = "Read and write the PLINK BED format, simply and efficiently."
 repository = "https://github.com/fastlmm/bed-reader"

{bed_reader-1.0.0b3 → bed_reader-1.0.0b4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: bed-reader
-Version: 1.0.0b3
+Version: 1.0.0b4
 Classifier: Programming Language :: Python :: 3.8
 Classifier: Programming Language :: Python :: 3.9
 Classifier: Programming Language :: Python :: 3.10

{bed_reader-1.0.0b3 → bed_reader-1.0.0b4}/bed_reader/_open_bed.py

@@ -1511,11 +1511,36 @@ def _read_csv(filepath, delimiter=None, dtype=None, usecols=None):
         # Find the dtype for this column
         col_dtype = dtype.get(input_index, np.str_)
         # Convert the column list to a numpy array with the specified dtype
-        columns.append(np.array(col, dtype=col_dtype))
+        columns.append(_convert_to_dtype(col, col_dtype))
 
     return columns, row_count
 
 
+def _convert_to_dtype(str_arr, dtype):
+    assert dtype in [np.str_, np.float32, np.int32]  # real assert
+
+    if dtype == np.str_:
+        return str_arr
+
+    try:
+        new_arr = str_arr.astype(dtype)
+    except ValueError as e:
+        if dtype == np.float32:
+            raise e
+        # for backwards compatibility, see if intermediate float helps int conversion
+        try:
+            assert dtype == np.int32  # real assert
+            float_arr = str_arr.astype(np.float32)
+        except ValueError:
+            raise e
+        new_arr = float_arr.astype(np.int32)
+        if not np.array_equal(new_arr, float_arr):
+            raise ValueError(
+                f"invalid literal for int: '{str_arr[np.where(new_arr != float_arr)][:1]}')"
+            )
+    return new_arr
+
+
 if __name__ == "__main__":
     import pytest
 

{bed_reader-1.0.0b3 → bed_reader-1.0.0b4}/bed_reader/tests/test_open_bed.py

@@ -898,6 +898,39 @@ def test_sparse():
         print(val_sparse.shape)
 
 
+def test_convert_to_dtype():
+    from bed_reader._open_bed import _convert_to_dtype
+
+    input = [
+        [["a", "b", "c"], ["a", "b", "c"], None, None],
+        [["1.0", "2.0", "3.0"], ["1.0", "2.0", "3.0"], [1, 2, 3], [1.0, 2.0, 3.0]],
+        [["1.0", "2.0", "3.5"], ["1.0", "2.0", "3.5"], None, [1.0, 2.0, 3.5]],
+        [["1", "2", "3"], ["1", "2", "3"], [1, 2, 3], [1.0, 2.0, 3.0]],
+        [["1", "A", "3"], ["1", "A", "3"], None, None],
+    ]
+    # convert all to np.array
+    input = [
+        [np.array(inner) if inner is not None else None for inner in outer]
+        for outer in input
+    ]
+
+    for ori, exp_str, exp_int, exp_float in input:
+        for dtype, exp in (
+            [np.str_, exp_str],
+            [np.int32, exp_int],
+            [
+                np.float32,
+                exp_float,
+            ],
+        ):
+            try:
+                actual = _convert_to_dtype(ori, dtype)
+                assert np.array_equal(actual, exp)
+            except ValueError as e:
+                print(e)
+                assert exp is None
+
+
 if __name__ == "__main__":
     logging.basicConfig(level=logging.INFO)