More tests of lazy expressions with Blosc2 NDArrays

Author: FrancescAlted

caterva2/client.py

@@ -556,6 +556,9 @@ def __init__(self, root, path):
         """
         super().__init__(root, path)
 
+    def __str__(self):
+        return self.path.as_posix()
+
     def __repr__(self):
         # TODO: add more info about dims, types, etc.
         return f"<Dataset: {self.path}>"

caterva2/services/sub.py

@@ -293,10 +293,9 @@ def open_b2(abspath, path):
             for key, value in operands.items():
                 if value is None:
                     raise ValueError(f'Missing operand "{key}"')
+                metaval = value.schunk.meta if hasattr(value, "schunk") else {}
                 vlmetaval = value.schunk.vlmeta
-                if "proxy-source" in value.schunk.meta or (
-                    "_ftype" in vlmetaval and vlmetaval["_ftype"] == "hdf5"
-                ):
+                if "proxy-source" in metaval or ("_ftype" in vlmetaval and vlmetaval["_ftype"] == "hdf5"):
                     # Save operand as Proxy, see blosc2.open doc for more info.
                     # Or, it can be an HDF5 dataset too (which should be handled in the next call)
                     relpath = srv_utils.get_relpath(
@@ -325,6 +324,11 @@ def open_b2(abspath, path):
                 elif isinstance(value, blosc2.LazyExpr):
                     # Properly open the operands (to e.g. find proxies)
                     for opkey, opvalue in value.operands.items():
+                        if isinstance(opvalue, blosc2.LazyExpr):
+                            continue
+                        relpath = srv_utils.get_relpath(
+                            opvalue, settings.cache, settings.personal, settings.shared, settings.public
+                        )
                         value.operands[opkey] = open_b2(opvalue.schunk.urlpath, relpath)
 
             return container

caterva2/tests/test_api.py

@@ -7,10 +7,12 @@
 # See LICENSE.txt for details about copyright and rights to use.
 ###############################################################################
 import contextlib
+import os
 import pathlib
 
 import blosc2
 import httpx
+import numexpr as ne
 import numpy as np
 import pytest
 
@@ -664,6 +666,50 @@ def test_lazyexpr(auth_client):
     np.testing.assert_array_equal(a[:], b[:])
 
 
+# More exercises for the expression evaluation with Blosc2 arrays
+@pytest.mark.parametrize(
+    "expression",
+    [
+        "a + 50",
+        "a ** 2.3 + b / 2.3",
+        "sqrt(a) ** sin(b)",
+        "where(a < 50, a + 50, b)",
+    ],
+)
+def test_lazyexpr2(expression, examples_dir, tmp_path, auth_client):
+    if not auth_client:
+        pytest.skip("authentication support needed")
+
+    root = pathlib.Path("@shared")
+    remote_root = auth_client.get(root)
+    remote_dir = "arrays"
+    ds_a = f"{remote_dir}/3d-blosc2-a.b2nd"
+    ds_b = f"{remote_dir}/3d-blosc2-b.b2nd"
+
+    with chdir_ctxt(tmp_path):
+        os.makedirs(remote_dir, exist_ok=True)
+        a = np.linspace(-1, 2, 1000).reshape(10, 10, 10)
+        blosc2.asarray(a, urlpath=ds_a, chunks=(5, 10, 10))
+        remote_a = remote_root.upload(ds_a)
+        b = np.linspace(-1, 2, 1000).reshape(10, 10, 10)
+        blosc2.asarray(b, urlpath=ds_b, chunks=(3, 5, 5))
+        remote_b = remote_root.upload(ds_b)
+        assert ds_a in remote_root
+        assert ds_b in remote_root
+
+        operands = {"a": remote_a, "b": remote_b}
+        lxpath = auth_client.lazyexpr("myexpr", expression, operands)
+        assert lxpath == pathlib.Path("@personal/myexpr.b2nd")
+
+        # Compute the expression
+        result = auth_client.get_slice(lxpath)
+        assert isinstance(result, blosc2.NDArray)
+
+        # Check the data
+        nresult = ne.evaluate(expression, {"a": a, "b": b})
+        np.testing.assert_allclose(result[:], nresult)
+
+
 def test_lazyexpr_getchunk(auth_client):
     if not auth_client:
         pytest.skip("authentication support needed")

caterva2/tests/test_hdf5_proxy.py

@@ -241,7 +241,7 @@ def test_unfold_fetch(fetch_or_slice, examples_dir, tmp_path, auth_client):
         "where(a < 50, a + 50, b)",
     ],
 )
-def test_expr_where(expression, examples_dir, tmp_path, auth_client):
+def test_expression(expression, examples_dir, tmp_path, auth_client):
     if not auth_client:
         pytest.skip("authentication support needed")