Revert "Remove numpy dependency"

This reverts commit 72e0b74.

.github/workflows/build.yaml

@@ -155,7 +155,7 @@ jobs:
       with:
         update: true
         install: >-
-          autotools gcc git libcurl-devel bzip2 liblzma zip libbz2-devel xz zlib zlib-devel libtool wget liblzma-devel ncurses-devel base-devel openssl-devel mingw64/mingw-w64-x86_64-python3 mingw64/mingw-w64-x86_64-python3-pip mingw-w64-x86_64-python3-setuptools mingw-w64-x86_64-python3-wheel mingw-w64-x86_64-python-biopython
+          autotools gcc git libcurl-devel bzip2 liblzma zip libbz2-devel xz zlib zlib-devel libtool wget liblzma-devel ncurses-devel base-devel openssl-devel mingw64/mingw-w64-x86_64-python3 mingw64/mingw-w64-x86_64-python3-pip mingw-w64-x86_64-python3-numpy mingw-w64-x86_64-python3-setuptools mingw-w64-x86_64-python3-wheel mingw-w64-x86_64-python-biopython
 
     - name: Check out code for the build
       uses: actions/checkout@v3

requirements.txt

@@ -4,6 +4,7 @@ PyVCF3>=1.0.3
 requests>=2.27.1
 mongoengine>=0.24.1
 cython>=0.29.28
+numpy>=1.22.0
 ## Testing
 pytest
 pytest-cov

setup.py

@@ -131,6 +131,7 @@ def read(*names, **kwargs):
         "PyVCF3",
         "mongoengine",
         "requests",
+        "numpy",
     ],
     entry_points={
         "console_scripts": [

src/mykrobe/cmds/amr.py

@@ -3,12 +3,13 @@
 import copy
 import json
 import logging
-import math
 import random
 import sys
 import tempfile
 import time
 
+import numpy as np
+
 from mykrobe import ONT_E_RATE, ONT_PLOIDY
 from mykrobe.metagenomics import AMRSpeciesPredictor
 from mykrobe.mformat import json_to_csv
@@ -22,8 +23,6 @@
 from mykrobe.typing.typer.variant import VariantTyper
 from mykrobe.utils import fix_amino_acid_X_variants_keys
 from mykrobe.utils import load_json
-from mykrobe.utils import binomial_random_sample
-from mykrobe.utils import poisson_random_sample
 from mykrobe.version import __version__ as atlas_version
 from mykrobe.version import __version__ as predictor_version
 
@@ -76,9 +75,9 @@ def _get_incorrect_kmer_percent_cov(cls, k_count, incorrect_kmer_to_pc_cov):
         return 0
 
     def _simulate_snps(self):
-        correct_covg = poisson_random_sample(self.mean_depth, self.iterations)
-        incorrect_covg = binomial_random_sample(
-            self.mean_depth, self.error_rate, self.iterations
+        correct_covg = np.random.poisson(lam=self.mean_depth, size=self.iterations)
+        incorrect_covg = np.random.binomial(
+            self.mean_depth, self.error_rate, size=self.iterations
         )
         probe_coverage_list = []
         vtyper = VariantTyper(
@@ -123,7 +122,7 @@ def _simulate_snps(self):
             )
             call = vtyper.type(vpc)
 
-            cov = math.log10(correct_covg[i] + incorrect_covg[i])
+            cov = np.log10(correct_covg[i] + incorrect_covg[i])
             conf = call["info"]["conf"]
             self.log_conf_and_covg.append((conf, cov))
 

src/mykrobe/utils.py

@@ -1,9 +1,7 @@
 import gzip
 import hashlib
 import json
-import math
 import os
-import random
 import re
 from pathlib import Path
 from typing import TextIO, Union
@@ -192,37 +190,3 @@ def get_first_chrom_name(fp: TextIO) -> str:
     if not header.startswith(">"):
         raise ValueError(f"Expected fasta file, but it did not start with '>'")
     return header[1:].split()[0]
-
-
-def poisson_random(lam):
-    # see https://en.wikipedia.org/wiki/Poisson_distribution#Random_variate_generation
-    L = math.exp(-lam)
-    k = 0
-    p = 1
-    while p > L:
-        k += 1
-        p *= random.uniform(0, 1)
-    return k - 1
-
-
-def poisson_random_sample(lam, size):
-    return [poisson_random(lam) for _ in range(size)]
-
-
-def binomial_random(n, p):
-    if not 0 <= p <= 1:
-        raise ValueError(f"Must have  0<=p<=1. p={p}")
-    if n <= 0:
-        raise ValueError(f"Must have n>0. n={n}")
-
-    successes = 0
-
-    for _ in range(n):
-        if random.random() < p:
-            successes += 1
-
-    return successes
-
-
-def binomial_random_sample(n, p, size):
-    return [binomial_random(n, p) for _ in range(size)]