analysis_java.py

import numpy as np
import pandas as pd
import dask.dataframe as dd
import dask.array as da
import matplotlib.pyplot as plt
import seaborn as sns
from dask.diagnostics import ProgressBar
ProgressBar().register()

dists = np.load('saved_tensors/java-huge-bpe-2000/test_proj_dist_cache.npy')
ranks = np.load('saved_tensors/java-huge-bpe-2000/test_proj_rank_cache.npy')
pkg_locality = np.load('saved_tensors/java-huge-bpe-2000/test_pkg_locality_cache.npy')
proj_locality = np.load('saved_tensors/java-huge-bpe-2000/test_proj_locality_cache.npy')
correctness = np.load('saved_tensors/java-huge-bpe-2000/test_proj_correctness_cache.npy')

dists = da.from_array(dists)
ranks = da.from_array(ranks)
pkg_locality = da.from_array(pkg_locality)
proj_locality = da.from_array(proj_locality)
correctness = da.from_array(correctness)

project_local_only = (proj_locality == 1) & (pkg_locality == 0).astype('int8')
locality = project_local_only + 2*pkg_locality
arr_all = da.stack([dists, ranks, locality, correctness], axis=1)

ddf = dd.from_array(arr_all, columns=['dist', 'rank', 'locality', 'correctness'])
ddf = ddf[ddf['dist'] >= -400]

# print(ddf.groupby('locality').count().compute())
# exit()
print('df build complete')

ddf = ddf.sort_values(['dist']).reset_index(drop=True)

ddf['overall_rank'] = ddf.groupby('locality').cumcount()

# dist - acc
# bins = [-10000] + list(range(-500, 0, 10)) + [0]
bins = list(np.arange(0, 2727431522, 100000))

ddf['rank_range'] = ddf['overall_rank'].map_partitions(pd.cut, bins)

dist_grouped = ddf.groupby(['locality', 'rank_range']).mean().reset_index().compute()

dist_grouped.to_csv('figures/java_dist_correctness.csv')

fig, ax = plt.subplots(figsize=(8, 4))
sns.scatterplot(x='dist_right', y='correctness', hue='locality', data=dist_grouped, s=5)

plt.savefig('figures/java_avg_correctness_by_dist_1024.pdf')

exit()
# rank - acc
grouped = ddf.groupby(['locality', 'rank']).mean().reset_index().compute()

grouped.to_csv('figures/java_rank.csv')

fig, ax = plt.subplots(figsize=(8, 4))
sns.scatterplot(x='rank', y='correctness', hue='locality', data=grouped, s=5)

plt.savefig('figures/java_avg_correctness_by_rank_1024.pdf')

# rank - dist

fig, ax = plt.subplots(figsize=(8, 4))
sns.scatterplot(x='rank', y='dist', hue='locality', data=grouped, s=5)

plt.savefig('figures/java_avg_dist_by_rank_1024.pdf')