lint

.flake8

@@ -1,5 +1,6 @@
 [flake8]
 ignore = E203, E231, E501, E722, W503, B950
-select = C,E,F,W,T,B,B9,I
+select = C,E,F,W,T,B9,I
 per-file-ignores =
-    tests/*: T
+    __init__.py: D104,
+    setup.py: D100

.github/workflows/ci.yml

@@ -34,13 +34,13 @@ jobs:
       run: python -m pytest -ra --cov-report=xml
 
     - name: Upload coverage to codecov
-      if: ${{ matrix.python-version == '3.7' && matrix.runs-on == 'ubuntu-latest'}}$
-        uses: codecov/codecov-action@v1
-        with:
-          #token: ${{ secrets.CODECOV_TOKEN }}
-          file: ./coverage.xml
-          name: codecov-umbrella
-          fail_ci_if_error: true
+      if: ${{ matrix.python-version == '3.7' && matrix.runs-on == 'ubuntu-latest'}}
+      uses: codecov/codecov-action@v1
+      with:
+        #token: ${{ secrets.CODECOV_TOKEN }}
+        file: ./coverage.xml
+        name: codecov-umbrella
+        fail_ci_if_error: true
 
   dist:
     name: Distribution build

Makefile

@@ -26,4 +26,4 @@ dist: clean
 	python setup.py sdist bdist_wheel
 
 clean:
-	rm -rf dist
+	rm -rf dist

README.md

@@ -37,7 +37,7 @@
 [sk-badge]:                 https://scikit-hep.org/assets/images/Scikit--HEP-Project-blue.svg
 
 
-[![DOI](https://zenodo.org/badge/235776682.svg)](https://zenodo.org/badge/latestdoi/235776682)  
+[![DOI](https://zenodo.org/badge/235776682.svg)](https://zenodo.org/badge/latestdoi/235776682)
 [![Binder](https://mybinder.org/badge_logo.svg)](https://mybinder.org/v2/gh/gradhep/neos/main?filepath=demo_kde_pyhf.ipynb)
 
 <img src="nbs/assets/neos_logo.png" alt="neos logo" width="250">
@@ -102,13 +102,13 @@ init_random_params, predict = stax.serial(
 )
 ```
 
-Now, let's compose a workflow that can make use of this network in a typical high-energy physics statistical analysis. 
+Now, let's compose a workflow that can make use of this network in a typical high-energy physics statistical analysis.
 
 Our workflow is as follows:
 - From a set of normal distributions with different means, we'll generate four blobs of `(x,y)` points, corresponding to a signal process, a nominal background process, and two variations of the background from varying the background distribution's mean up and down.
 - We'll then feed these points into the previously defined neural network for each blob, and construct a histogram of the output using kernel density estimation. The difference between the two background variations is used as a systematic uncertainty on the nominal background.
 - We can then leverage the magic of `pyhf` to construct an [event-counting statistical model](https://scikit-hep.org/pyhf/intro.html#histfactory) from the histogram yields.
-- Finally, we calculate the p-value of a test between the nominal signal and background-only hypotheses. This uses a [profile likelihood-based test statistic](https://arxiv.org/abs/1007.1727). 
+- Finally, we calculate the p-value of a test between the nominal signal and background-only hypotheses. This uses a [profile likelihood-based test statistic](https://arxiv.org/abs/1007.1727).
 
 In code, `neos` can specify this workflow through function composition:
 

nbs/02_makers.ipynb

@@ -658,7 +658,8 @@
   "kernelspec": {
    "display_name": "Python 3.9.1 64-bit ('3.9.1': pyenv)",
    "name": "python3"
-  }
+  },
+  "language_info": {}
  },
  "nbformat": 4,
  "nbformat_minor": 4

nbs/demo_kde_pyhf.ipynb

@@ -18,7 +18,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 14,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -44,7 +44,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 15,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -68,7 +68,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 16,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -105,14 +105,14 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 17,
    "metadata": {},
    "outputs": [
     {
      "data": {
       "text/plain": "(DeviceArray(0.05989214, dtype=float64),\n [(DeviceArray([[-2.6561793e-05, -1.3431838e-04,  3.7267394e-04, ...,\n                 -9.5937648e-05, -3.1907926e-04, -4.7556667e-05],\n                [ 3.2410408e-05,  1.6153457e-04, -2.3639399e-04, ...,\n                 -7.6833248e-06,  3.2346524e-04,  2.8453616e-05]],            dtype=float32),\n   DeviceArray([-1.0775060e-06,  1.6448357e-05,  2.5383127e-04, ...,\n                 9.4325784e-05, -2.0307447e-04, -1.4521212e-05],            dtype=float32)),\n  (),\n  (DeviceArray([[ 2.0301204e-07, -2.2220850e-06,  6.1780779e-06, ...,\n                 -1.0296096e-07,  6.6260043e-07,  4.6643160e-07],\n                [ 4.1513945e-07,  9.9888112e-08,  1.8419865e-05, ...,\n                  4.3947058e-08,  6.2674635e-06, -1.1988274e-06],\n                [-2.0467695e-08, -1.2231366e-05,  5.6882186e-08, ...,\n                 -6.0378164e-07, -1.1139725e-05,  3.0422871e-06],\n                ...,\n                [ 4.0854286e-07, -6.6176659e-07,  1.1947506e-05, ...,\n                 -2.9116174e-08,  3.6848719e-06,  7.2041793e-08],\n                [-5.8339246e-08, -1.7854960e-05,  5.4534485e-07, ...,\n                 -9.0966802e-07, -1.6797023e-05,  4.3758305e-06],\n                [ 4.5675144e-08, -3.7403470e-06,  5.2949540e-06, ...,\n                 -2.3437008e-07, -4.0581904e-06,  8.9615037e-07]],            dtype=float32),\n   DeviceArray([ 1.49559655e-05, -8.68027928e-05,  5.29851706e-04, ...,\n                -3.46148317e-06,  5.94751364e-05, -9.33103820e-06],            dtype=float32)),\n  (),\n  (DeviceArray([[ 3.7021862e-05],\n                [-2.0765483e-04],\n                [ 3.0189482e-04],\n                ...,\n                [-4.2103562e-05],\n                [ 2.0076217e-04],\n                [ 6.4033600e-05]], dtype=float32),\n   DeviceArray([-5.677808e-06], dtype=float32)),\n  ()])"
      },
-     "execution_count": null,
+     "execution_count": 17,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -133,7 +133,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 18,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -173,7 +173,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 20,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -257,7 +257,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 23,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -329,9 +329,16 @@
   }
  ],
  "metadata": {
+  "interpreter": {
+   "hash": "653c4ff722aa20b955619ff1fd61a2acdddc645229f21af3c6e92ae1ff77e485"
+  },
   "kernelspec": {
    "display_name": "Python 3.8.3  ('venv': venv)",
    "name": "python3"
+  },
+  "language_info": {
+   "name": "python",
+   "version": ""
   }
  },
  "nbformat": 4,

pyproject.toml

@@ -19,8 +19,8 @@ warn_unused_configs = true
 
 disallow_any_generics = true
 disallow_subclassing_any = true
-disallow_untyped_calls = true
-disallow_untyped_defs = true
+disallow_untyped_calls = false
+disallow_untyped_defs = false
 disallow_incomplete_defs = true
 check_untyped_defs = true
 disallow_untyped_decorators = true

setup.cfg

@@ -8,21 +8,23 @@ author = Nathan Simpson
 author_email = [email protected]
 maintainer = Nathan Simpson
 maintainer_email = [email protected]
-license = BSD-3-Clause
+license = Apache-2.0
 license_file = LICENSE
 platforms =
     Any
 classifiers =
     Development Status :: 1 - Planning
     Intended Audience :: Developers
     Intended Audience :: Science/Research
+    License :: OSI Approved :: Apache Software License
     License :: OSI Approved :: BSD License
     Operating System :: OS Independent
     Programming Language :: Python
     Programming Language :: Python :: 3
     Programming Language :: Python :: 3 :: Only
     Programming Language :: Python :: 3.7
     Programming Language :: Python :: 3.8
+    Programming Language :: Python :: 3.9
     Topic :: Scientific/Engineering
 project_urls =
     Documentation = https://neos.readthedocs.io/

src/neos/data.py

@@ -1,6 +1,6 @@
 # AUTOGENERATED! DO NOT EDIT! File to edit: nbs/00_data.ipynb (unless otherwise specified).
 
-__all__ = ['generate_blobs']
+__all__ = ["generate_blobs"]
 
 # Cell
 import jax

src/neos/examples/__init__.py

@@ -1,3 +1,3 @@
-__all__ = ('neos_pyhf_example')
+__all__ = ["neos_pyhf_example"]
 
-from .nn_pyhf_cls import neos_pyhf_example
+from .nn_pyhf_cls import neos_pyhf_example

src/neos/examples/nn_pyhf_cls.py

@@ -2,7 +2,7 @@
 # To add a new markdown cell, type '# %% [markdown]'
 # %% [markdown]
 # # KDE demo, with histosys!
-# 
+#
 # > It works :)
 # %% [markdown]
 # ![](assets/kde_pyhf_animation.gif)
@@ -14,19 +14,18 @@
 import jax.experimental.optimizers as optimizers
 import jax.experimental.stax as stax
 import jax.random
-from jax.random import PRNGKey
 import numpy as np
-
 import pyhf
-pyhf.set_backend('jax')
-pyhf.default_backend = pyhf.tensor.jax_backend(precision='64b')
+from jax.random import PRNGKey
+from relaxed import infer
 
 from .. import data, makers
-from relaxed import infer
 
+pyhf.set_backend("jax")
+pyhf.default_backend = pyhf.tensor.jax_backend(precision="64b")
 
-def neos_pyhf_example(rng=PRNGKey(1)):
 
+def neos_pyhf_example(rng=PRNGKey(1)):
 
     # regression net
     init_random_params, predict = stax.serial(
@@ -42,15 +41,20 @@ def neos_pyhf_example(rng=PRNGKey(1)):
     # ## Compose differentiable workflow
 
     # %%
-    dgen = data.generate_blobs(rng,blobs=4) 
+    dgen = data.generate_blobs(rng, blobs=4)
 
     # Specify our hyperparameters ahead of time for the kde histograms
-    bins = np.linspace(0,1,4)
-    bandwidth=0.27
+    bins = np.linspace(0, 1, 4)
+    bandwidth = 0.27
     reflect_infinite_bins = True
 
-    hmaker = makers.hists_from_nn(dgen, predict, hpar_dict = dict(bins=bins,bandwidth=bandwidth),method='kde', reflect_infinities=reflect_infinite_bins)
-
+    hmaker = makers.hists_from_nn(
+        dgen,
+        predict,
+        hpar_dict=dict(bins=bins, bandwidth=bandwidth),
+        method="kde",
+        reflect_infinities=reflect_infinite_bins,
+    )
 
     # %%
     nnm = makers.histosys_model_from_hists(hmaker)
@@ -76,7 +80,6 @@ def loss(params, test_mu):
     opt_init, opt_update, opt_params = optimizers.adam(1e-3)
 
     def train_network(N):
-        cls_vals = []
         _, network = init_random_params(jax.random.PRNGKey(1), (-1, 2))
         state = opt_init(network)
         losses = []
@@ -95,16 +98,13 @@ def update_and_value(i, opt_state, mu):
             losses.append(value)
             metrics = {"loss": losses}
 
-
-
             yield network, metrics, epoch_time
 
     maxN = 50  # make me bigger for better results!
 
     # Training
     for i, (network, metrics, epoch_time) in enumerate(train_network(maxN)):
-        print(f"epoch {i}:", f'CLs = {metrics["loss"][-1]}, took {epoch_time}s')
+        pass  # print(f"epoch {i}:", f'CLs = {metrics["loss"][-1]}, took {epoch_time}s')
     # %%
 
-    return metrics["loss"][-1] 
-
+    return metrics["loss"][-1]