MAINT: Fix Cython/C type problems

- Single-source Cython/C type definitions: `core._ext.types`
- Introduce semantic type aliases: ADJ, NODE, DEGREE, WEIGHT, FIELD, etc.
- Remove deprecated type aliases: `np.{int|float}`
- Replace C type aliases by Numpy type aliases
- Add missing Numpy type casts before Python/Cython boundary
- Disambiguate `timeseries._ext.numerics._twin_surrogates_{s|r}`
  (blame: 098489b & 2f6a5b9)
- Fix some broken output pointers
- Reduce redundant copying
- Vectorize some simple Python loops
- Eliminate some dead code
- Potentially related issues: #128, #141, #142, #145

pyunicorn/climate/_ext/numerics.pyx

@@ -15,57 +15,37 @@
 # for complex systems science: The pyunicorn package"
 
 cimport cython
-from cpython cimport bool
-
 
 import numpy as np
 cimport numpy as np
 
-
-BOOLTYPE = np.uint8
-INTTYPE = np.int
-INT8TYPE = np.int8
-INT16TYPE = np.int16
-INT32TYPE = np.int32
-INT64TYPE = np.int64
-FLOATTYPE = np.float
-FLOAT32TYPE = np.float32
-FLOAT64TYPE = np.float64
-ctypedef np.uint8_t BOOLTYPE_t
-ctypedef np.int_t INTTYPE_t
-ctypedef np.int8_t INT8TYPE_t
-ctypedef np.int16_t INT16TYPE_t
-ctypedef np.int32_t INT32TYPE_t
-ctypedef np.int64_t INT64TYPE_t
-ctypedef np.float_t FLOATTYPE_t
-ctypedef np.float32_t FLOAT32TYPE_t
-ctypedef np.float64_t FLOAT64TYPE_t
-
+from ...core._ext.types import FIELD, DFIELD
+from ...core._ext.types cimport MASK_t, FIELD_t, DFIELD_t
 
 cdef extern from "src_numerics.c":
     void _cython_calculate_mutual_information(
             float *anomaly, int n_samples, int N, int n_bins, double scaling,
             double range_min, long *symbolic, long *hist, long *hist2d,
             float *mi)
-    void _calculate_corr_fast(int m, int tmax, int *final_mask,
+    void _calculate_corr_fast(int m, int tmax, bint *final_mask,
             float *time_series_ranked, float *spearman_rho)
 
 
 # mutual_info =================================================================
 
 def _calculate_mutual_information_cython(
-    np.ndarray[float, ndim=2, mode='c'] anomaly not None,
+    np.ndarray[FIELD_t, ndim=2, mode='c'] anomaly not None,
     int n_samples, int N, int n_bins, double scaling, double range_min):
 
     cdef:
-        np.ndarray[long, ndim=2, mode='c'] symbolic = \
-            np.zeros((N, n_samples), dtype='int64')
-        np.ndarray[long, ndim=2, mode='c'] hist = \
-            np.zeros((N, n_bins), dtype='int64')
-        np.ndarray[long, ndim=2, mode='c'] hist2d = \
-            np.zeros((n_bins, n_bins), dtype='int64')
-        np.ndarray[float, ndim=2, mode='c'] mi = \
-            np.zeros((N, N), dtype='float32')
+        np.ndarray[DFIELD_t, ndim=2, mode='c'] symbolic = np.zeros(
+            (N, n_samples), dtype=DFIELD)
+        np.ndarray[DFIELD_t, ndim=2, mode='c'] hist = np.zeros(
+            (N, n_bins), dtype=DFIELD)
+        np.ndarray[DFIELD_t, ndim=2, mode='c'] hist2d = np.zeros(
+            (n_bins, n_bins), dtype=DFIELD)
+        np.ndarray[FIELD_t, ndim=2, mode='c'] mi = np.zeros(
+            (N, N), dtype=FIELD)
 
     _cython_calculate_mutual_information(
         <float*> np.PyArray_DATA(anomaly), n_samples, N, n_bins, scaling,
@@ -79,14 +59,14 @@ def _calculate_mutual_information_cython(
 # rainfall ====================================================================
 
 def _calculate_corr(int m, int tmax,
-    np.ndarray[int, ndim=2, mode='c'] final_mask not None,
-    np.ndarray[float, ndim=2, mode='c'] time_series_ranked not None):
+    np.ndarray[MASK_t, ndim=2, mode='c'] final_mask not None,
+    np.ndarray[FIELD_t, ndim=2, mode='c'] time_series_ranked not None):
 
-    cdef np.ndarray[float, ndim=2, mode='c'] spearman_rho = \
-            np.zeros((m, m), dtype='float32')
+    cdef np.ndarray[FIELD_t, ndim=2, mode='c'] spearman_rho = np.zeros(
+        (m, m), dtype=FIELD)
 
     _calculate_corr_fast(m, tmax,
-            <int*> np.PyArray_DATA(final_mask),
+            <bint*> np.PyArray_DATA(final_mask),
             <float*> np.PyArray_DATA(time_series_ranked),
             <float*> np.PyArray_DATA(spearman_rho))
 

pyunicorn/climate/_ext/src_numerics.c

@@ -8,6 +8,8 @@
 * License: BSD (3-clause)
 */
 
+#include <math.h>
+
 // mutual_info ================================================================
 
 void _cython_calculate_mutual_information(float *anomaly, int n_samples,
@@ -18,7 +20,7 @@ void _cython_calculate_mutual_information(float *anomaly, int n_samples,
          in_nodes;
     double norm, rescaled, hpl, hpm, plm;
 
-    double *p_anomaly;
+    float *p_anomaly;
     float *p_mi, *p_mi2;
     long *p_symbolic, *p_symbolic1, *p_symbolic2, *p_hist, *p_hist1,
          *p_hist2, *p_hist2d;

pyunicorn/climate/mutual_info.py

@@ -26,6 +26,7 @@
 # array object and fast numerics
 import numpy as np
 
+from ..core._ext.types import to_cy, FIELD
 from ._ext.numerics import _calculate_mutual_information_cython
 
 #  Import progress bar for easy progress bar handling
@@ -155,10 +156,8 @@ def _cython_calculate_mutual_information(self, anomaly, n_bins=32):
         #  using the maximum range of the whole dataset.
         scaling = 1./(range_max - range_min)
 
-        anomaly = anomaly.astype(np.float32).copy(order='c')
-        mi = _calculate_mutual_information_cython(anomaly, n_samples, N,
-                                                  n_bins, scaling,
-                                                  range_min)
+        mi = _calculate_mutual_information_cython(
+            to_cy(anomaly, FIELD), n_samples, N, n_bins, scaling, range_min)
 
         if self.silence_level <= 1:
             print("Done!")

pyunicorn/climate/rainfall.py

@@ -23,6 +23,9 @@
 #  Import essential packages
 #
 
+import numpy as np
+
+from ..core._ext.types import to_cy, MASK, FIELD
 from ._ext.numerics import _calculate_corr
 
 #  Import cnTsonisClimateNetwork for TsonisClimateNetwork class
@@ -276,10 +279,7 @@ def calculate_corr(self, final_mask, anomaly):
         :return: the Spearman correlation matrix.
         """
         # Get rank time series
-        time_series_ranked = self.rank_time_series(anomaly).astype('float32')
-
+        time_series_ranked = self.rank_time_series(anomaly)
         m, tmax = anomaly.shape
-
-        return _calculate_corr(m, tmax,
-                               final_mask.astype('int32').copy(order='c'),
-                               time_series_ranked.copy(order='c'))
+        return _calculate_corr(
+            m, tmax, to_cy(final_mask, MASK), to_cy(time_series_ranked, FIELD))