Merge pull request #10 from katehyerinjeon/hw6

hw6

Author: katehyerinjeon

.bumpversion.cfg

@@ -1,16 +1,12 @@
 [bumpversion]
-current_version = 0.1.0
+current_version = 0.0.1
 commit = True
 tag = True
 
-[bumpversion:file:pynrm/_version.py]
+[bumpversion:file:pynrm/__init__.py]
 search = __version__ = "{current_version}"
 replace = __version__ = "{new_version}"
 
 [bumpversion:file:pyproject.toml]
-search = version="{current_version}"
-replace = version="{new_version}"
-
-[bumpversion:file:setup.py]
-search = version="{current_version}"
-replace = version="{new_version}"
+search = version = "{current_version}"
+replace = version = "{new_version}"

CONTRIBUTING.md

@@ -1 +1,21 @@
-TODO
+# Contribution Guidelines
+Thanks for your interest in contributing! We want to make contributing to this project as easy and transparent as possible whether it is reporting a bug, discussing the current state of the code, submitting a fix, proposing new features, or more.
+
+## Getting Started
+We use [GitHub issues](https://github.com/katehyerinjeon/pynrm/issues) to track public bugs. Report a bug by opening a new issue and make sure to include the followings:
+
+- A quick summary of the issue
+- Clear and specific steps to reproduce with sample code if possible
+- What you expected would happen
+- What actually happens
+- Side notes possibly including why you think this might be happening or stuff you tried that didn't work
+
+## Making Changes
+Pull requests are the best way to propose changes to the codebase as we use [GitHub workflow](https://github.com/katehyerinjeon/pynrm/actions).
+
+1. Fork the repository and create your branch from `main` (or a topic branch from where you want to base your work)
+2. Make commits of logical units (rebase your feature branch before submitting if needed) and write commit messages in [proper format](https://github.blog/2022-06-30-write-better-commits-build-better-projects/)
+3. Add tests for any new features or code that should be tested
+4. Ensure the test suite passes by running `make tests`
+5. Make sure your code lints by running `make lint`
+6. Push your changes to the topic branch in your fork of the repository and submit a pull request

README.md

@@ -12,3 +12,31 @@ Numerous evaluation-selection systems have been devised to produce populations w
 `pynrm` provides a simple yet powerful simulation tool to forecast the stochastic impacts of these systems.
 One major bottleneck to running these simulations is that as the number of animals bred increases, the size of the matrix grows exponentially.
 `pynrm` efficiently solves for the numerator relationship matrix values by tracing up the pedigree for only the relevant ancestors, thereby minimizing computational overhead.
+
+## Basic Usage
+### Installation
+`pynrm` is available on PyPI:
+
+```shell
+$ pip install pyrnm
+```
+
+### Supported Features
+- Fine-tuned reproduction simulation with user-defined weights
+```python
+from pynrm.Pedigree import Pedigree
+from pynrm.Simulator import Simulator
+
+pedigree = Pedigree()
+simulator = Simulator(pedigree, 10, 100, 0.6, 1)
+
+simulator.reproduce()
+```
+
+- Data visualization and analysis of simulation results (COMING SOON)
+
+## Documentation
+-- LINK TO OFFICIAL DOC COMING SOON --
+
+## Contribution
+Contributions are welcome! There are many ways to contribute to this project. Get started [here](CONTRIBUTING.md).

pynrm/Pedigree.py

@@ -0,0 +1,34 @@
+import numpy as np
+import pandas as pd
+
+
+class Pedigree:
+    def __init__(self, data=None):
+        dtype = {"gen": int, "sire": pd.Int64Dtype(), "dam": pd.Int64Dtype(), "ebv": float, "sex": str}
+
+        if data is not None:
+            if isinstance(data, pd.DataFrame):
+                self.data = pd.DataFrame(data=data, columns=["gen", "sire", "dam", "ebv", "sex"]).astype(dtype=dtype)
+            else:
+                raise TypeError("'data' must be of type dataframe")
+        else:
+            male = pd.DataFrame(
+                {
+                    "gen": 0,
+                    "sire": None,
+                    "dam": None,
+                    "ebv": np.random.normal(0, 1, size=500),
+                    "sex": "M",
+                }
+            )
+            female = pd.DataFrame(
+                {
+                    "gen": 0,
+                    "sire": None,
+                    "dam": None,
+                    "ebv": np.random.normal(0, 1, size=500),
+                    "sex": "F",
+                }
+            )
+            data = male.append(female, ignore_index=True)
+            self.data = pd.DataFrame(data=data, columns=["gen", "sire", "dam", "ebv", "sex"]).astype(dtype=dtype)

pynrm/Simulator.py

@@ -1,17 +1,37 @@
 import numpy as np
 import random
 from .nrm import get_nrm
-from .pedigree import pedigree
+from .Pedigree import Pedigree
 
 
 class Simulator:
-    def __init__(self, male_k, female_k, h, w):
+    def __init__(self, pedigree, male_k, female_k, h, w):
+        if not isinstance(pedigree, Pedigree):
+            raise TypeError("'pedigree' must be of type Pedigree")
+        if not isinstance(male_k, int):
+            raise TypeError("'male_k' must be of type int")
+        if not isinstance(female_k, int):
+            raise TypeError("'female_k' must be of type int")
+        if not isinstance(h, float):
+            raise TypeError("'h' must be of type float")
+        if not isinstance(w, float):
+            raise TypeError("'w' must be of type float")
+
+        if male_k < 0:
+            raise ValueError("'male_k' cannot be negative")
+        if female_k < 0:
+            raise ValueError("'female_k' cannot be negative")
+        if h < 0:
+            raise ValueError("'h' cannot be negative")
+        if w < 0:
+            raise ValueError("'w' cannot be negative")
+
         self.pedigree = pedigree
-        self.gen = 0
         self.male_k = male_k
         self.female_k = female_k
         self.h = h
         self.w = w
+        self.gen = 0
 
     def get_ebv(self, sire, dam):
         """
@@ -37,7 +57,7 @@ def get_ebv(self, sire, dam):
         w = round(np.random.normal(0, 1), 3)
         f = 0.5 * (get_nrm(self.pedigree, sire, sire) + get_nrm(self.pedigree, dam, dam)) - 1
 
-        ebv = 0.5 * (self.pedigree.iloc[sire].ebv + self.pedigree.iloc[dam].ebv)
+        ebv = 0.5 * (self.pedigree.data.iloc[sire].ebv + self.pedigree.data.iloc[dam].ebv)
         +v * h_sqrt * np.sqrt((1 - f) / 2)
         +w * np.sqrt(1 - self.h)
 
@@ -56,7 +76,7 @@ def get_adjusted_ebv(self, top_k, candidate):
         :rtype: float
         """
 
-        ebv = self.pedigree.iloc[candidate].ebv
+        ebv = self.pedigree.data.iloc[candidate].ebv
 
         # if none selected as top k, then original ebv is used for scoring
         if len(top_k) == 0:
@@ -118,15 +138,15 @@ def reproduce(self):
         :rtype: dataframe
         """
 
-        all_males = self.pedigree[self.pedigree["sex"] == "M"].index.tolist()
-        all_females = self.pedigree[self.pedigree["sex"] == "F"].index.tolist()
+        all_males = self.pedigree.data[self.pedigree.data["sex"] == "M"].index.tolist()
+        all_females = self.pedigree.data[self.pedigree.data["sex"] == "F"].index.tolist()
 
         top_males = self.get_top_k([], all_males, self.male_k)
         top_females = self.get_top_k([], all_females, self.female_k)
         random.shuffle(top_females)
 
         self.gen += 1
-        new_pedigree = pedigree
+        new_pedigree_data = self.pedigree.data
 
         for i in range(len(top_males)):
             sire = top_males[i]
@@ -145,8 +165,8 @@ def reproduce(self):
                         "ebv": ebv,
                         "sex": random.choices(["M", "F"])[0],
                     }
-                    new_pedigree = new_pedigree.append(new_animal, ignore_index=True)
+                    new_pedigree_data = new_pedigree_data.append(new_animal, ignore_index=True)
 
-        self.pedigree = new_pedigree
+        self.pedigree = Pedigree(new_pedigree_data)
 
-        return new_pedigree
+        return new_pedigree_data

pynrm/__init__.py

@@ -1,2 +1 @@
-# from ._version import __version__
-# from .nrm import get_nrm
+__version__ = "0.0.1"

pynrm/__main__.py

@@ -1,5 +1,2 @@
-# from .nrm import get_nrm
-
 if __name__ == "__main__":
-    # TODO
     pass

pynrm/_version.py

@@ -1,3 +1,7 @@
+import pandas as pd
+from .Pedigree import Pedigree
+
+
 def get_nrm(pedigree, i, j):
     """
     Calculate nrm value of i-th row and j-th column
@@ -10,36 +14,43 @@ def get_nrm(pedigree, i, j):
     :raises ValueError: if i or j is not non-negative
     """
 
-    if i < 0:
+    if not isinstance(pedigree, Pedigree):
+        raise TypeError("'pedigree' must be of type Pedigree")
+    if i is not pd.NA and not isinstance(i, int):
+        raise TypeError("'i' must be of type int")
+    if j is not pd.NA and not isinstance(j, int):
+        raise TypeError("'j' must be of type int")
+
+    if i is not pd.NA and i < 0:
         raise ValueError("'i' cannot be negative")
-    if j < 0:
+    if j is not pd.NA and j < 0:
         raise ValueError("'j' cannot be negative")
 
     # swap i and j if i is larger than j
-    if i > j:
+    if i is not pd.NA and j is not pd.NA and i > j:
         tmp = i
         i = j
         j = tmp
 
     # get sire and dam of j from the pedigree
-    sire = pedigree.iloc[j].sire
-    dam = pedigree.iloc[j].dam
+    sire = pedigree.data.iloc[j].sire
+    dam = pedigree.data.iloc[j].dam
 
     # diagonal - i and j is the same
-    if i is not None and i == j:
-        if sire is not None and dam is not None:
-            res = 1 + 0.5 * get_nrm(sire, dam)
+    if i is not pd.NA and i == j:
+        if sire is not pd.NA and dam is not pd.NA:
+            res = 1 + 0.5 * get_nrm(pedigree, sire, dam)
         else:
             res = 1
 
     # off-diagonal - i and j is not the same
     else:
-        if sire is not None and dam is not None:
-            res = 0.5 * (get_nrm(sire, i) + get_nrm(dam, i))
-        elif sire is not None:
-            res = 0.5 * get_nrm(sire, i)
-        elif dam is not None:
-            res = 0.5 * get_nrm(dam, i)
+        if sire is not pd.NA and dam is not pd.NA:
+            res = 0.5 * (get_nrm(pedigree, sire, i) + get_nrm(pedigree, dam, i))
+        elif sire is not pd.NA:
+            res = 0.5 * get_nrm(pedigree, sire, i)
+        elif dam is not pd.NA:
+            res = 0.5 * get_nrm(pedigree, dam, i)
         else:
             res = 0