Updated all astype calls to use np types

Author: seanlaw

setup.sh

@@ -1,5 +1,5 @@
 #!/bin/sh
 
 echo "y" |  python -m pip uninstall stumpy
-python -m pip install .
+python -m pip install --use-feature=in-tree-build .
 rm -rf stumpy.egg-info build dist __pycache__

stumpy/aamp_motifs.py

@@ -224,7 +224,7 @@ def aamp_motifs(
 
     T, T_subseq_isfinite = core.preprocess_non_normalized(T[np.newaxis, :], m)
     T_squared = np.sum(core.rolling_window(T * T, m), axis=-1)
-    P = P[np.newaxis, :].astype("float64")
+    P = P[np.newaxis, :].astype(np.float64)
 
     motif_distances, motif_indices = _aamp_motifs(
         T,

stumpy/aampdist_snippets.py

@@ -231,7 +231,7 @@ def aampdist_snippets(
     for i in range(k):
         mask = snippets_profiles[i] <= total_min
         snippets_fractions[i] = np.sum(mask) / total_min.shape[0]
-        total_min = total_min - mask.astype(float)
+        total_min = total_min - mask.astype(np.float64)
 
     return (
         snippets,

stumpy/aampi.py

@@ -97,9 +97,9 @@ def __init__(self, T, m, excl_zone=None, egress=True):
         self._egress = egress
 
         mp = aamp(self._T, self._m)
-        self._P = mp[:, 0]
-        self._I = mp[:, 1]
-        self._left_I = mp[:, 2]
+        self._P = mp[:, 0].astype(np.float64)
+        self._I = mp[:, 1].astype(np.int64)
+        self._left_I = mp[:, 2].astype(np.int64)
         self._left_P = np.empty(self._P.shape)
         self._left_P[:] = np.inf
 

stumpy/floss.py

@@ -33,7 +33,7 @@ def _nnmark(I):
 
     This is a fast and vectorized implementation of the nnmark algorithm.
     """
-    I = I.astype(int)
+    I = I.astype(np.int64)
 
     # Replace index values that are less than zero with its own positional index
     idx = np.argwhere(I < 0).flatten()
@@ -569,25 +569,25 @@ def cac_1d_(self):
         """
         Get the updated 1-dimensional corrected arc curve (CAC_1D)
         """
-        return self._cac.astype(np.float)
+        return self._cac.astype(np.float64)
 
     @property
     def P_(self):
         """
         Get the updated matrix profile
         """
-        return self._mp[:, 0].astype(np.float)
+        return self._mp[:, 0].astype(np.float64)
 
     @property
     def I_(self):
         """
         Get the updated (right) matrix profile indices
         """
-        return self._mp[:, 3].astype(np.int)
+        return self._mp[:, 3].astype(np.int64)
 
     @property
     def T_(self):
         """
         Get the updated time series, `T`
         """
-        return self._T.astype(np.float)
+        return self._T.astype(np.float64)

stumpy/motifs.py

@@ -233,7 +233,7 @@ def motifs(
         cutoff = max(np.mean(P) - 2 * np.std(P), np.min(P))
 
     T, M_T, Σ_T = core.preprocess(T[np.newaxis, :], m)
-    P = P[np.newaxis, :].astype("float64")
+    P = P[np.newaxis, :].astype(np.float64)
 
     motif_distances, motif_indices = _motifs(
         T,

stumpy/scraamp.py

@@ -489,25 +489,25 @@ def P_(self):
         """
         Get the updated matrix profile
         """
-        return self._P[:, 0].astype(np.float)
+        return self._P[:, 0].astype(np.float64)
 
     @property
     def I_(self):
         """
         Get the updated matrix profile indices
         """
-        return self._I[:, 0].astype(np.int)
+        return self._I[:, 0].astype(np.int64)
 
     @property
     def left_I_(self):
         """
         Get the updated left matrix profile indices
         """
-        return self._I[:, 1].astype(np.int)
+        return self._I[:, 1].astype(np.int64)
 
     @property
     def right_I_(self):
         """
         Get the updated right matrix profile indices
         """
-        return self._I[:, 2].astype(np.int)
+        return self._I[:, 2].astype(np.int64)

stumpy/scrump.py

@@ -519,25 +519,25 @@ def P_(self):
         """
         Get the updated matrix profile
         """
-        return self._P[:, 0].astype(np.float)
+        return self._P[:, 0].astype(np.float64)
 
     @property
     def I_(self):
         """
         Get the updated matrix profile indices
         """
-        return self._I[:, 0].astype(np.int)
+        return self._I[:, 0].astype(np.int64)
 
     @property
     def left_I_(self):
         """
         Get the updated left matrix profile indices
         """
-        return self._I[:, 1].astype(np.int)
+        return self._I[:, 1].astype(np.int64)
 
     @property
     def right_I_(self):
         """
         Get the updated right matrix profile indices
         """
-        return self._I[:, 2].astype(np.int)
+        return self._I[:, 2].astype(np.int64)

stumpy/snippets.py

@@ -240,7 +240,7 @@ def snippets(
     for i in range(k):
         mask = snippets_profiles[i] <= total_min
         snippets_fractions[i] = np.sum(mask) / total_min.shape[0]
-        total_min = total_min - mask.astype(float)
+        total_min = total_min - mask.astype(np.float64)
 
     return (
         snippets,

stumpy/stimp.py

@@ -22,19 +22,19 @@ def _bfs_indices(n):
         The breadth first search (level order) indices
     """
     if n == 1:  # pragma: no cover
-        return np.array([0], dtype=np.int32)
+        return np.array([0], dtype=np.int64)
 
-    nlevel = np.floor(np.log2(n) + 1).astype(int)
+    nlevel = np.floor(np.log2(n) + 1).astype(np.int64)
     nindices = np.power(2, np.arange(nlevel))
     cumsum_nindices = np.cumsum(nindices)
     nindices[-1] = n - cumsum_nindices[np.searchsorted(cumsum_nindices, n) - 1]
 
-    indices = np.empty((2, nindices.max()), dtype=np.int32)
+    indices = np.empty((2, nindices.max()), dtype=np.int64)
     indices[0, 0] = 0
     indices[1, 0] = n
-    tmp_indices = np.empty((2, 2 * nindices.max()), dtype=np.int32)
+    tmp_indices = np.empty((2, 2 * nindices.max()), dtype=np.int64)
 
-    out = np.empty(n, dtype=np.int32)
+    out = np.empty(n, dtype=np.int64)
     out_idx = 0
 
     for nidx in nindices:

stumpy/stumpi.py

@@ -110,9 +110,9 @@ def __init__(self, T, m, excl_zone=None, egress=True, normalize=True):
         self._egress = egress
 
         mp = stump(self._T, self._m)
-        self._P = mp[:, 0]
-        self._I = mp[:, 1]
-        self._left_I = mp[:, 2]
+        self._P = mp[:, 0].astype(np.float64)
+        self._I = mp[:, 1].astype(np.int64)
+        self._left_I = mp[:, 2].astype(np.int64)
         self._left_P = np.empty(self._P.shape)
         self._left_P[:] = np.inf