Cleanup2

.github/workflows/black_n_pylint.yml

@@ -31,7 +31,7 @@ jobs:
       - name: Install dependencies
         run: |
           python -m pip install --upgrade pip
-          pip install black lyncs_setuptools[pylint]
+          pip install black lyncs_setuptools[all]
 
       - name: Applying black formatting
         run: |

README.md

@@ -7,7 +7,7 @@
 [![build & test](https://img.shields.io/github/workflow/status/Lyncs-API/lyncs.quda/build%20&%20test?logo=github&logoColor=white)](https://github.com/Lyncs-API/lyncs.quda/actions)		
 -->		
 [![license](https://img.shields.io/github/license/Lyncs-API/lyncs.quda?logo=github&logoColor=white)](https://github.com/Lyncs-API/lyncs.quda/blob/master/LICENSE)		
-[![pylint](https://img.shields.io/badge/pylint%20score-8.5%2F10-yellowgreen?logo=python&logoColor=white)](http://pylint.pycqa.org/)
+[![pylint](https://img.shields.io/badge/pylint%20score-7.9%2F10-yellow?logo=python&logoColor=white)](http://pylint.pycqa.org/)
 [![black](https://img.shields.io/badge/code%20style-black-000000.svg?logo=codefactor&logoColor=white)](https://github.com/ambv/black)		
 
 [QUDA](http://lattice.github.io/quda/) is a library for performing calculations in lattice QCD on GPUs.		

lyncs_quda/array.py

@@ -24,7 +24,7 @@ class Array:
     def __init__(self, typename, size, elems=None):
         self._qarray = lib.quda.array[typename, size]()
 
-        if elems != None:
+        if elems is not None:
             if isiterable(elems):
                 if len(elems) > size:
                     raise ValueError()

lyncs_quda/clover_field.py

@@ -6,15 +6,19 @@
     "CloverField",
 ]
 
-import numpy
-from cppyy.gbl.std import vector
 from functools import cache
+import numpy
 
 from lyncs_cppyy import make_shared, to_pointer
 from .lib import lib, cupy
 from .lattice_field import LatticeField
 from .gauge_field import GaugeField
-from .enums import QudaParity, QudaTwistFlavorType, QudaCloverFieldOrder, QudaFieldCreate
+from .enums import (
+    QudaParity,
+    QudaTwistFlavorType,
+    QudaCloverFieldOrder,
+    QudaFieldCreate,
+)
 
 # TODO list
 # We want dimension of (cu/num)py array to reflect parity and order

lyncs_quda/dirac.py

@@ -8,10 +8,9 @@
 
 from functools import wraps
 from dataclasses import dataclass, field
-from numpy import sqrt
 from typing import Union
 from lyncs_cppyy import make_shared, nullptr
-from .gauge_field import gauge, GaugeField
+from .gauge_field import GaugeField
 from .clover_field import CloverField
 from .spinor_field import spinor
 from .lib import lib
@@ -20,7 +19,7 @@
     QudaMatPCType,
     QudaDagType,
     QudaParity,
-    )
+)
 
 
 @dataclass(frozen=True)
@@ -222,13 +221,13 @@ def action(self, phi, **params):
         """
 
         if not self.full:
-            if "CLOVER" in self.type and self.symm == True:
+            if "CLOVER" in self.type and self.symm:
                 raise ValueError("Preconditioned matrix should be asymmetric")
-            if "CLOVER" not in self.type and self.symm != True:
+            if "CLOVER" not in self.type and not self.symm:
                 raise ValueError(
                     "Preconditioned matrix should be symmetric for non-clover type Dirac matrix"
                 )
-            if "CLOVER" in self.type and self.computeTrLog != True:
+            if "CLOVER" in self.type and not self.computeTrLog:
                 raise ValueError(
                     "computeTrLog should be set True in the preconditioned case"
                 )

lyncs_quda/gauge_field.py

@@ -18,12 +18,11 @@
 from collections import defaultdict
 from functools import cache
 import numpy
-from lyncs_cppyy import make_shared, lib as tmp, to_pointer, array_to_pointers
+from lyncs_cppyy import make_shared, to_pointer, array_to_pointers
 from lyncs_utils import prod, isiterable
 from .lib import lib, cupy
 from .array import Array
-from .lattice_field import LatticeField, backend
-from .spinor_field import spinor
+from .lattice_field import LatticeField
 from .time_profile import default_profiler
 from .enums import (
     QudaReconstructType,
@@ -121,8 +120,8 @@ def new(self, reconstruct=None, geometry=None, **kwargs):
                     kwargs["dofs"][0],
                     val // 2 if self.iscomplex else val,
                 )
-            except ValueError:
-                raise ValueError(f"Invalid reconstruct {reconstruct}")
+            except ValueError as VE:
+                raise VE(f"Invalid reconstruct {reconstruct}")
         out = super().new(**kwargs)
         is_momentum = kwargs.get("is_momentum", self.is_momentum)
         out.is_momentum = is_momentum
@@ -197,9 +196,8 @@ def ncol(self):
     def order(self):
         "Data order of the field"
         dofs = self.dofs_per_link
-        if self.precision != "double" and (
-            dofs == 8 or dofs == 12
-        ):  # if FLOAT8 defined, if prec=half/quarter and recon=8, FLOAT8
+        if self.precision != "double" and dofs in (8, 12):
+            # if FLOAT8 defined, if prec=half/quarter and recon=8, FLOAT8
             return "FLOAT4"
         return "FLOAT2"
 
@@ -346,7 +344,7 @@ def extended_field(self, sites=1):
                     self.ptr, self.quda_params, numpy.array(sites, dtype="int32")
                 )
             )
-        elif self.location == "CUDA":
+        if self.location == "CUDA":
             "Returns cudaGaugeField"
             return make_shared(
                 lib.createExtendedGauge(
@@ -531,7 +529,7 @@ def gaussian(self, epsilon=1, seed=None):
         seed = seed or int(time() * 1e9)
         lib.gaugeGauss(self.quda_field, seed, epsilon)
 
-    def uniform(self, epsilon=1, seed=None):
+    def uniform(self, seed=None):
         """
         Generates Uniform distributed SU(N) field.
         """
@@ -756,7 +754,7 @@ def compute_paths(
             if not len(paths) == len(coeffs):
                 raise ValueError("Paths and coeffs must have the same length")
         else:
-            assert coeffs == None, "coeffs not used in case of not sum_paths"
+            assert coeffs is None, "coeffs not used in case of not sum_paths"
 
         # Preparing fnc
         if insertion is not None:

lyncs_quda/lattice_field.py

@@ -6,7 +6,6 @@
     "LatticeField",
 ]
 
-from array import array
 from contextlib import contextmanager
 from functools import cache
 import numpy
@@ -16,10 +15,6 @@
 from .lib import lib, cupy
 from .array import lat_dims
 
-from lyncs_cppyy import to_pointer
-import ctypes
-import traceback
-
 
 def get_precision(dtype):
     if dtype in ["float64", "complex128"]:

lyncs_quda/lib.py

@@ -13,11 +13,10 @@
 from os import environ
 from pathlib import Path
 from array import array
-from appdirs import user_data_dir
 from math import prod
-from lyncs_cppyy import Lib, nullptr, cppdef
-from lyncs_cppyy.ll import addressof, to_pointer
-from lyncs_utils import static_property, lazy_import
+from appdirs import user_data_dir
+from lyncs_cppyy import Lib, cppdef
+from lyncs_utils import lazy_import
 from . import __path__
 from .config import QUDA_MPI, QUDA_GITVERSION, QUDA_PRECISION, QUDA_RECONSTRUCT
 
@@ -86,7 +85,7 @@ def device_id(self, value):
                 f"device_id cannot be changed: current={self.device_id}, given={value}"
             )
         if not isinstance(value, int):
-            raise TypeError(f"Unsupported type for device: {type(device)}")
+            raise TypeError(f"Unsupported type for device ID: {type(value)}")
         self._device_id = value
 
     def get_current_device(self):
@@ -262,6 +261,7 @@ def __del__(self):
 PATHS = list(__path__)
 
 headers = [
+    "comm_quda.h",
     "quda.h",
     "gauge_field.h",
     "gauge_tools.h",

lyncs_quda/solver.py

@@ -7,7 +7,7 @@
     "Solver",
 ]
 
-from functools import wraps, cache
+from functools import wraps
 from warnings import warn
 from lyncs_cppyy import nullptr, make_shared
 from .dirac import Dirac, DiracMatrix

lyncs_quda/spinor_field.py

@@ -1,5 +1,5 @@
 """
-Interface to gauge_field.h
+Interface to color_spinor_field.h
 """
 
 __all__ = [
@@ -8,7 +8,7 @@
     "SpinorField",
 ]
 
-from functools import reduce, cache
+from functools import cache
 from time import time
 from lyncs_cppyy import make_shared
 from lyncs_cppyy.ll import to_pointer
@@ -22,7 +22,7 @@
     QudaPCType,
     QudaFieldCreate,
     QudaNoiseType,
-    )
+)
 
 """
 NOTE:
@@ -124,7 +124,7 @@ def site_order(self):
     def site_order(self, value):
         if value is None:
             value = "NONE"
-        values = f"Possible values are NONE, EVEN_ODD, ODD_EVEN"
+        values = "Possible values are NONE, EVEN_ODD, ODD_EVEN"
         if not isinstance(value, str):
             raise TypeError("Expected a string. " + values)
         value = value.upper()
@@ -215,15 +215,15 @@ def norm1(self, parity=None):
         "L1 norm of the field"
         if parity == "EVEN":
             return lib.blas.norm1(self.quda_field.Even())
-        elif parity == "ODD":
+        if parity == "ODD":
             return lib.blas.norm1(self.quda_field.Odd())
         return lib.blas.norm1(self.quda_field)
 
     def norm2(self, parity=None):
         "L2 norm of the field"
         if parity == "EVEN":
             return lib.blas.norm2(self.quda_field.Even())
-        elif parity == "ODD":
+        if parity == "ODD":
             return lib.blas.norm2(self.quda_field.Odd())
         return lib.blas.norm2(self.quda_field)
 

patches/error.patch

@@ -0,0 +1,63 @@
+diff --git a/include/comm_quda.h b/include/comm_quda.h
+index e99d412..f221eab 100644
+--- a/include/comm_quda.h
++++ b/include/comm_quda.h
+@@ -4,6 +4,7 @@
+ #include <quda_constants.h>
+ #include <quda_api.h>
+ #include <array.h>
++#include <stdexcept>
+
+ #ifdef __cplusplus
+ extern "C" {
+@@ -420,4 +421,7 @@ namespace quda
+   bool commAsyncReduction();
+   void commAsyncReductionSet(bool global_reduce);
+
++  class QudaException : public std::runtime_error {
++    using std::runtime_error::runtime_error;
++  };
+ } // namespace quda
+diff --git a/lib/comm_common.cpp b/lib/comm_common.cpp
+index 2c58188..330b0e0 100644
+--- a/lib/comm_common.cpp
++++ b/lib/comm_common.cpp
+@@ -1,6 +1,9 @@
+ #include <unistd.h> // for gethostname()
+ #include <assert.h>
+ #include <limits>
++#include <cstdlib>
++#include <exception>
++#include <string>
+
+ #include <quda_internal.h>
+ #include <communicator_quda.h>
+@@ -130,7 +133,12 @@ namespace quda
+
+     return topo;
+   }
+-
++  
++  static int current_status;
++  void abort(){
++    comm_abort_(current_status);
++  }
++  
+   void comm_abort(int status)
+   {
+ #ifdef HOST_DEBUG
+@@ -142,7 +150,13 @@ namespace quda
+   backward::Printer p;
+   p.print(st, getOutputFile());
+ #endif
+-  comm_abort_(status);
++  static bool called = false;
++  if (not called) {
++    std::set_terminate(abort);
++    called = true;
++  }
++  current_status = status;
++  throw QudaException("QUDA Error: exit code " + std::to_string(status));
+ }
+
+ } // namespace quda

test/test_error.py

@@ -0,0 +1,19 @@
+from lyncs_quda import gauge, gauge_scalar
+
+import pytest
+from lyncs_quda.testing import (
+    fixlib as lib,
+    lattice_loop,
+)
+
+
+@lattice_loop
+def test_error(lib, lattice):
+    gf = gauge(lattice)
+    gs = gauge_scalar(lattice)
+
+    gf.quda_field.copy(gs.quda_field)
+    with pytest.raises(lib.std.runtime_error):
+        gf.quda_field.copy(gs.quda_field)
+    gf.zero()
+    assert gf == 0

test/test_evenodd.py

@@ -1,7 +1,7 @@
 import pytest
 from lyncs_quda import evenodd, continous, to_quda, from_quda
 
-# from lyncs_quda.lib import fixlib as lib
+from lyncs_quda.lib import fixlib as lib
 import numpy as np
 
 
@@ -20,17 +20,15 @@ def outer(request):
     return request.param
 
 
-def test_evenodd(shape, inner, outer):
+def test_evenodd(lib, shape, inner, outer):
     tile = np.array([1, -1])
     for i in range(1, len(shape)):
         tile = np.array([tile, tile * -1])
 
     arr = np.tile(tile, shape)
     shape = arr.shape
     out = evenodd(arr)
-
     assert (continous(out) == arr).all()
-
     out = out.flatten()
     n = out.shape[0] // 2