Merge pull request #13 from pbashyal-nmdp/cython_fix

Cython fix to build the package

Author: mmaiers-nmdp

.gitignore

@@ -134,3 +134,6 @@ dmypy.json
 # behave
 pretty.output
 allure_report/
+
+# cython temp files
+grim/**/*.c

MANIFEST.in

@@ -0,0 +1,9 @@
+include requirements.txt
+include README.md
+include LICENSE
+include *.txt
+recursive-include grim *.py
+recursive-include grim *.txt
+recursive-include grim *.json
+recursive-include grim *.pyx
+recursive-include grim *.pyd

grim/__init__.py

@@ -26,4 +26,4 @@
 """Top-level package for py-grim."""
 
 __organization__ = "NMDP/CIBMTR Bioinformatics"
-__version__ = "0.0.6"
+__version__ = "0.0.7"

grim/conf/__init__.py

@@ -23,6 +23,6 @@
 #
 
 
-"""Top-level package for py-grim."""
-
-__organization__ = "NMDP/CIBMTR Bioinformatics"
+__author__ = """Martin Maiers"""
+__email__ = "[email protected]"
+__version__ = "0.0.7"

grim/imputation/__init__.py

@@ -1,16 +1,13 @@
 # -*- coding: utf-8 -*-
 
-#
-#    grim Graph Imputation
-#    Copyright (c) 2021 Be The Match operated by National Marrow Donor Program. All Rights Reserved.
 #
 #    This library is free software; you can redistribute it and/or modify it
 #    under the terms of the GNU Lesser General Public License as published
 #    by the Free Software Foundation; either version 3 of the License, or (at
 #    your option) any later version.
 #
 #    This library is distributed in the hope that it will be useful, but WITHOUT
-#    ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+#    ANY WARRANTY; with out even the implied warranty of MERCHANTABILITY or
 #    FITNESS FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public
 #    License for more details.
 #
@@ -23,6 +20,6 @@
 #
 
 
-"""Top-level package for py-grim."""
-
-__organization__ = "NMDP/CIBMTR Bioinformatics"
+__author__ = """Martin Maiers"""
+__email__ = "[email protected]"
+__version__ = "0.0.7"

grim/imputation/graph_generation/__init__.py

@@ -1,16 +1,13 @@
 # -*- coding: utf-8 -*-
 
-#
-#    grim Graph Imputation
-#    Copyright (c) 2021 Be The Match operated by National Marrow Donor Program. All Rights Reserved.
 #
 #    This library is free software; you can redistribute it and/or modify it
 #    under the terms of the GNU Lesser General Public License as published
 #    by the Free Software Foundation; either version 3 of the License, or (at
 #    your option) any later version.
 #
 #    This library is distributed in the hope that it will be useful, but WITHOUT
-#    ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+#    ANY WARRANTY; with out even the implied warranty of MERCHANTABILITY or
 #    FITNESS FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public
 #    License for more details.
 #
@@ -23,6 +20,6 @@
 #
 
 
-"""Top-level package for py-grim."""
-
-__organization__ = "NMDP/CIBMTR Bioinformatics"
+__author__ = """Martin Maiers"""
+__email__ = "[email protected]"
+__version__ = "0.0.7"

grim/imputation/imputegl/__init__.py

@@ -24,4 +24,4 @@
 
 __author__ = """Martin Maiers"""
 __email__ = "[email protected]"
-__version__ = "0.0.4"
+__version__ = "0.0.7"

grim/imputation/imputegl/cutils.pyx

@@ -0,0 +1,65 @@
+import cython
+
+
+@cython.boundscheck(False)
+@cython.wraparound(False)
+cpdef open_ambiguities(list hap, unsigned char loc, tuple split_loc):
+    cdef unsigned int k, i, p, j #hap_len, haps_len, splits_len
+    cdef Py_ssize_t hap_len, haps_len, splits_len
+    cdef list hap_new, hap1
+    # cdef np.ndarray[STR, ndim=1] hap_new, hap1
+    p = 0
+    if len(split_loc) > 1:
+        # This opens all allele ambiguities
+        hap_len = len(hap[0])
+        haps_len = len(hap)
+        splits_len = len(split_loc)
+        hap_new = [None] * (haps_len * splits_len)
+        # hap_new = np.empty(haps_len * splits_len, dtype=np.object)  # produces an empty list of haplotypes
+        hap1 = [None] * hap_len
+        # hap1 = np.empty(haps_len, dtype=np.object)
+        for k in range(haps_len):  # split a given locus in all haps.
+
+            for j in range(hap_len):
+                hap1[j] = hap[k][j]
+
+            for i in range(splits_len):
+                hap1[loc] = split_loc[i]
+                hap_new[p] = hap1[:]
+                p += 1
+        return hap_new
+    return hap
+
+@cython.boundscheck(False)
+@cython.wraparound(False)
+cpdef create_hap_list(list all_haps, dict optionDict, unsigned int N_Loc):
+    cdef unsigned int i, j, count
+    cdef list hap_list = []
+    cdef list all_hap_split
+
+    for i in range(len(all_haps)):
+        all_hap_split = all_haps[i].split('~')
+        count = 0
+        for j in range(len(all_hap_split)):
+            if all_hap_split[j] not in optionDict:
+                break
+            else:
+                count += 1
+
+        if count == N_Loc:
+            hap_list.append(all_hap_split)
+    return hap_list
+
+@cython.boundscheck(False)
+@cython.wraparound(False)
+cpdef deepcopy_list(list l):
+    cdef list copy_l
+    cdef unsigned int i, length
+    length = len(l)
+    copy_l = [None] * length
+    for i in range(length):
+        if isinstance(l[i], list):
+            copy_l[i] = deepcopy_list(l[i])
+        else:
+            copy_l[i] = l[i]
+    return copy_l

grim/imputation/imputegl/impute.py

@@ -7,8 +7,9 @@
 import os.path
 import json
 
-import numpy as np
 
+import numpy as np
+from .cutils import open_ambiguities, create_hap_list, deepcopy_list
 from .cypher_plan_b import CypherQueryPlanB
 from .cypher_query import CypherQuery
 
@@ -118,6 +119,31 @@ def clean_up_gl(gl):
 
 
 class Imputation(object):
+    __slots__ = (
+        "logger",
+        "verbose",
+        "populations",
+        "netGraph",
+        "priorMatrix",
+        "full_hapl",
+        "index_dict",
+        "full_loci",
+        "factor",
+        "_factor_missing_data",
+        "cypher",
+        "cypher_plan_b",
+        "matrix_planb",
+        "count_by_prob",
+        "number_of_options_threshold",
+        "plan",
+        "option_1",
+        "option_2",
+        "haplotypes_number_in_phase",
+        "save_space_mode",
+        "nodes_for_plan_A",
+        "unk_priors",
+    )
+
     def __init__(self, net=None, config=None, count_by_prob=None, verbose=False):
         """Constructor
         Intialize an instance of `Imputation` with a py2neo graph
@@ -914,8 +940,8 @@ def open_phases(self, haps, N_Loc, gl_string):
             fq = []
 
             for k in range(2):
-                hap_list = []
-                hap_list.append(haps[j][k])
+                hap_list = [haps[j][k]]
+                hap_list_splits = [tuple(allele.split("/")) for allele in hap_list[0]]
 
                 # compute the number of options:
                 options = 1
@@ -1590,12 +1616,12 @@ def comp_cand(
         # probabilties and accumulate cartesian productEpsilon=0.0001
         chr = self.gl2haps(gl_string)
         if chr == []:
-            return
+            return None, None
         # if we in 9-loci, check if the type input in valid format
         if self.nodes_for_plan_A:
             geno_type = self.input_type(chr["Genotype"][0])
             if not geno_type in self.nodes_for_plan_A:
-                return
+                return None, None
 
         n_loci = chr["N_Loc"]
 
@@ -1604,7 +1630,7 @@ def comp_cand(
 
         # return if the result is empty (why would that be?)
         if pmags == []:
-            return
+            return None, None
 
         # res_muugs = {'Haps': 'NaN', 'Probs': 0}
         res_muugs = {"MaxProb": 0, "Haps": {}, "Pops": {}}
@@ -1713,7 +1739,7 @@ def call_comp_phase_prob(self, epsilon, n, phases, chr, MUUG_output, planb):
                     epsilon /= 10
                     if epsilon < min_epsilon:
                         epsilon = 0.0
-                    phases_planb = copy.deepcopy(phases)
+                    phases_planb = deepcopy_list(phases)
                     # Find the option according to plan b
                     if MUUG_output:
                         res = self.comp_phase_prob_plan_b(

grim/imputation/imputegl/networkx_graph.py

@@ -9,10 +9,19 @@ def missing(labelA, labelB):
 
 
 class Graph(object):
+    __slots__ = (
+        "graph",
+        "labelDict",
+        "whole_graph",
+        "full_loci",
+        "nodes_plan_a",
+        "nodes_plan_b",
+    )
+
     def __init__(self, config):
-        self.graph = nx.Graph()
+        self.graph = nx.DiGraph()
         self.labelDict = {}
-        self.whole_graph = nx.Graph()
+        self.whole_graph = nx.DiGraph()
         self.full_loci = config["full_loci"]
         self.nodes_plan_a, self.nodes_plan_b = [], []
         if config["nodes_for_plan_A"]:
@@ -63,8 +72,11 @@ def build_graph(self, nodesFile, edgesFile, allEdgesFile):
                 if len(row) > 0:
                     node1 = nodesDict[row[0]]
                     node2 = nodesDict[row[1]]
-                    if node1 in self.graph.nodes() and node2 in self.graph.nodes():
-                        self.graph.add_edge(node1, node2)
+                    if node1 in self.graph and node2 in self.graph:
+                        if self.graph.nodes[node1]["label"] == self.full_loci:
+                            self.graph.add_edge(node2, node1)
+                        else:
+                            self.graph.add_edge(node1, node2)
 
         edgesfile.close()
 
@@ -88,7 +100,6 @@ def build_graph(self, nodesFile, edgesFile, allEdgesFile):
                     self.whole_graph.add_edge(node1, node2, color=kind)
 
         allEdgesfile.close()
-
         nodesDict.clear()
 
     # return all haplotype by specific label