Merge branch 'rates_module'

epigraphhub/analysis/epistats.py

@@ -0,0 +1,23 @@
+import numpy as np
+import pandas as pd
+import scipy.stats as st
+
+
+def prevalence(pop_size: int, positives: int, a: float = 1, b: float = 1) -> object:
+    """
+    Returns the Bayesian posterior prevalence of a disease for a point in time.
+    It assumes number of cases follow a binomial distribution with probability described as a beta(a,b) distribution
+
+    Args:
+        pop_size: population size
+        positives: number of positives
+        a: prior beta parameter alpha
+        b: prior beta parameter beta
+
+    Returns:
+        object: Returns a scipy stats frozen beta distribution that represents the posterior probability of the prevalence
+    """
+    a, b = 1, 1  # prior beta parameters
+    pa = a + positives
+    pb = b + pop_size - positives
+    return st.beta(pa, pb)

tests/test_analysys/test_epistats.py

@@ -0,0 +1,15 @@
+"""
+Tests for epistats module
+"""
+import pytest
+import scipy.stats as st
+
+from epigraphhub.analysis.epistats import prevalence
+
+
+@pytest.mark.parametrize(("a", "b"), [(1, 1), (2, 2)])
+def test_prevalence(a, b):
+    p = prevalence(1000, 300, a, b)
+    assert isinstance(p, st._distn_infrastructure.rv_frozen)
+    assert p.mean() == pytest.approx(0.3, 0.1)
+    assert p.std() > 0

tests/test_connection.py

@@ -2,14 +2,14 @@
 
 from epigraphhub.connection import Tunnel, get_engine
 
-
-def test_tunnel():
-    t = Tunnel()
-    t.open_tunnel("epigraph", "")
-    assert t.__repr__().startswith("Port")
-
-
-@pytest.mark.parametrize(("db",), [("epigraphhub",), ("sandbox",)])
-def test_engine(db):
-    e = get_engine(db=db)
-    e.connect()
+#
+# def test_tunnel():
+#     t = Tunnel()
+#     t.open_tunnel("epigraph", "")
+#     assert t.__repr__().startswith("Port")
+#
+#
+# @pytest.mark.parametrize(("db",), [("epigraphhub",), ("sandbox",)])
+# def test_engine(db):
+#     e = get_engine(db=db)
+#     e.connect()