def _generic_draw(root,
repo_ids,
df_type,
mean,
cov_pairs,
img_folder="",
image_file="",
small=True,
medium=False,
large=False,
lim_0_1=False):
group_by = SUITE_SIZE if df_type == DataFrameType.FIXED_SIZE else SUITE_COVERAGE_BIN
dfs = [read_combined_df(root, repo_id, df_type) for repo_id in repo_ids]
df = _join(dfs, group_by, mean=mean)
vda_df = _get_vda(df, group_by, cov_pairs=cov_pairs)
if df_type == DataFrameType.FIXED_SIZE:
order = sorted(df[group_by].unique())
else:
order = list(sorted(df[group_by].unique(), key=cov_bin_key))
ys = name_columns(cov_pairs)
fig, axs = _draw(group_by, vda_df, df, order, ys, mean)
_add_lines(axs[:-1], small, medium, large, lim_0_1)
if not image_file:
f = "fixed_size" if df_type == DataFrameType.FIXED_SIZE else "fixed_coverage"
if mean:
f += "_mean"
image_file = Path(img_folder) / f"{f}_vda.png"
plt.savefig(image_file, dpi=300)
logger.info("Saved graph to {p}", p=image_file)
def draw_grouped_fixed_size(root, repo_ids, img_folder="", image_file=""):
# size
size_dfs = [
read_combined_df(root, repo_id, DataFrameType.FIXED_SIZE)
for repo_id in repo_ids
]
size_dfs = [
df.groupby([SUITE_SIZE, METRIC], as_index=False).mean()
for df in size_dfs
]
size_dfs = pd.concat(size_dfs, axis=0, ignore_index=True, sort=False)
if not image_file:
image_file = Path(img_folder) / "fixed_size_grouped.png"
draw_vertically(size_dfs, SUITE_SIZE, image_file)
def _generic_draw(root, repo_ids, df_type,
cov_pairs,
img_folder="", image_file="",
small=True, medium=False, large=False, lim_0_1=False,
auto_lines=False,
plot_type="p",
for_presentation=True
):
group_by = SUITE_SIZE if df_type == DataFrameType.FIXED_SIZE else SUITE_COVERAGE_BIN
dfs = [read_combined_df(root, repo_id, df_type) for repo_id in repo_ids]
dfs = [_get_vda(df, group_by, cov_pairs=cov_pairs) for df in dfs]
# with pd.option_context('display.max_rows', None, 'display.max_columns',
# None): # more options can be specified also
# for df in dfs:
# print(df)
vda_df = _join(dfs)
if df_type == DataFrameType.FIXED_SIZE:
order = sorted(vda_df[group_by].unique())
else:
order = list(sorted(vda_df[group_by].unique(), key=cov_bin_key))
ys = name_columns(cov_pairs)
sns.set_context("paper")
if for_presentation:
sns.set_context("talk")
with sns.axes_style(rc=RC if for_presentation else {}):
fig, axs = _draw(group_by, vda_df, order, ys, plot_type=plot_type)
_add_lines(axs[:-1], small, medium, large, lim_0_1, auto_lines=auto_lines)
if not image_file:
f = ["fixed_size" if df_type == DataFrameType.FIXED_SIZE else "fixed_coverage"]
f += [_explain_plot(plot_type)]
f += ["vda"]
stem = "_".join(f)
image_file = Path(img_folder) / f"{stem}.png"
plt.tight_layout()
plt.savefig(image_file, dpi=300, transparent=for_presentation)
logger.info("Saved graph to {p}", p=image_file)
zip(
map(str, map(operator.attrgetter("path"), ms)),
map(str, map(operator.attrgetter("path"), ms)),
map(str, map(operator.attrgetter("total_cases"), ms)),
map(str, map(operator.attrgetter("statements"), ms))
)
)
)))
# exit()
repos = []
for m in ms:
repo_id = m.repo.id
repo_id = repo_id.replace("@", "_")
df = read_combined_df(graphs_path, repo_id, DataFrameType.FIXED_SIZE)
if df is not None:
metrics = df[METRIC].unique()
if len(metrics) < 3:
continue
by_metric = df.groupby(METRIC)[SUITE_COVERAGE].agg(["min", "max"])
r = (
repo_id,
df[SUITE_SIZE].max(),
# by_metric["max"][str(CoverageMetric.STATEMENT)],
# by_metric["max"][str(CoverageMetric.BRANCH)],
# by_metric["max"][str(CoverageMetric.ALL_PAIRS)],
# by_metric["min"][str(CoverageMetric.STATEMENT)],
# by_metric["min"][str(CoverageMetric.BRANCH)],
# by_metric["min"][str(CoverageMetric.ALL_PAIRS)],