Merge pull request #16 from NREL/compactInstall

fix plot and compact install

.github/workflows/ci.yml

@@ -32,11 +32,12 @@ jobs:
         run: |
           python -m pip install --upgrade pip
           python -m pip install -r requirements.txt
-      - name: Formatting and sorting import
+      - name: Formatting
         run: |
           source .github/linters/formatting.sh
           format *.py true
           format utils true
+          codespell
       - name: Normalizing flow test
         run: |
           python main_iterative.py -i tests/input_test

README.md

@@ -10,8 +10,7 @@
 
 1. `conda create --name uips python=3.10`
 2. `conda activate uips`
-3. `conda install openmpi`
-4. `pip install -r requirements.txt`
+3. `pip install -r requirements.txt`
 
 ### Option 2: From `poetry`
 
@@ -20,7 +19,7 @@ This requires [poetry](https://python-poetry.org/docs/#installation)
 
 ## Purpose
 
-The purpose of the tool is to perform a smart downselection of a large number of datapoints. Typically, large numerical simulations generate billions, or even trillions of datapoints. However, there may be redundancy in the dataset which unecessarily constrains memory and computing requirements. Here, redundancy is defined as closeness in feature space. The method is called phase-space sampling.
+The purpose of the tool is to perform a smart downselection of a large number of datapoints. Typically, large numerical simulations generate billions, or even trillions of datapoints. However, there may be redundancy in the dataset which unnecessarily constrains memory and computing requirements. Here, redundancy is defined as closeness in feature space. The method is called phase-space sampling.
 
 ## Running the example without poetry
 

data-efficientML/README.md

@@ -4,7 +4,7 @@ This folder is NOT necessary for using the phase-space sampling package. It only
 Uniform-in-phase-space data-selection with iterative normalizing flows, Under review.
 
 # Data
-The folder `data/` contains the phase space sampled data for five independent runs of each case described in the paper aforementionned. To run all the codes and reproduce the results, the file `fullData.npy` is necessary. Due to space constraints it is not in the repository but is available upon request.
+The folder `data/` contains the phase space sampled data for five independent runs of each case described in the paper aforementioned. To run all the codes and reproduce the results, the file `fullData.npy` is necessary. Due to space constraints it is not in the repository but is available upon request.
 
 
 # Usage

fixFormat.sh

@@ -2,3 +2,4 @@ source .github/linters/formatting.sh
 
 format *.py
 format utils
+codespell

main_iterative.py

@@ -7,6 +7,7 @@
 # import matplotlib.pyplot as plt
 
 sys.path.append("utils")
+import argparse
 import time
 
 import parallel as par
@@ -17,7 +18,6 @@
 import sampler
 import utils
 
-import argparse
 parser = argparse.ArgumentParser(description="Downsampler")
 parser.add_argument(
     "-i",

requirements.txt

@@ -1,7 +1,9 @@
+openmpi
 torch
 numpy
 mpi4py
 scikit-learn
 prettyPlot>=0.0.10
 black
 isort
+codespell

visualizeDownSampled_subplots.py

@@ -48,6 +48,8 @@ def plotScatterProjection(data, fullData, fieldNames, lims):
             tick.set_rotation(33)
         ax.set_ylim(lims[1])
 
+    plt.tight_layout() 
+
 import argparse
 parser = argparse.ArgumentParser(description="Visualize downsampled data")
 parser.add_argument(
@@ -118,12 +120,7 @@ def plotScatterProjection(data, fullData, fieldNames, lims):
         downSampledData = fullData[np.load(dataFile)["indices"], :]
         plotScatterProjection(downSampledData, fullData, fieldNames, lims)
         plt.savefig(
-            figureFolder
-            + "/"
-            + inpt["prefixDownsampledData"]
-            + "_"
-            + str(nSample)
-            + ".png"
+            os.path.join(figureFolder, f"{inpt['prefixDownsampledData']}_{nSample}_it{pdf_iter}.png")
         )
         plt.close()
         print("DONE!")