def get_modules_from_db(repo_ids):
repo_ids = [r.replace("_", "@") for r in repo_ids]
ms: List[Module] = Module.select() \
.join(Repo) \
.group_by(Repo.name) \
.where(Repo.id << repo_ids)
return ms
def draw_count_test_cases():
ms: List[Module] = Module.select() \
.join(Repo) \
.group_by(Repo.name) \
.order_by(Module.total_cases.desc())
df = pd.DataFrame(list(ms.dicts()))
f, ax = plt.subplots(figsize=(10, 5))
bin_column_name = "bin_tc"
column_to_bin = "total_cases"
bins = [1, 2, 6, 10, 20, 50, 100, 2000]
bin_labels = [f"[{l}-{r}]" for l, r in zip(bins[:-1], bins[1:])]
kw = {
bin_column_name:
lambda x: pd.cut(x[column_to_bin], bins=bins, labels=bin_labels)
}
df: pd.DataFrame = df.assign(**kw)
color = sns.color_palette()[0]
sns.countplot(data=df, x=bin_column_name, color=color)
x_label = "Suite size"
y_label = "Repositories count"
ax.set(xlabel=x_label, ylabel=y_label)
total = len(df)
# plt.title(f"Repositories by test suite size (total: {total})")
plt.tight_layout()
plt.savefig("good_repos_by_suite_size_bin.png",
dpi=300,
transparent=transparent_bg)
def get_modules_from_db(repo_ids):
to_id = lambda s: s[:-41] + "@" + s[-40:]
repo_ids = [to_id(r) for r in repo_ids]
print(repo_ids)
ms: List[Module] = Module.select() \
.join(Repo) \
.where(Repo.id << repo_ids)
return ms
def draw_coverage_goals():
ms: List[Module] = Module.select() \
.join(Repo) \
.group_by(Repo.name) \
.order_by(Module.total_cases.desc())
df = pd.DataFrame(list(ms.dicts()))
humanize = lambda l: HUMAN_READABLE_LABELS[l]
df["all_pairs"] = df[["m_pairs", "im_pairs", "ic_pairs"]].sum(axis=1)
for col, readable_col in HUMAN_READABLE_LABELS.items():
df[readable_col] = df[col]
df[col] = None
coverage_goals_cols = ["statements", "branches", "all_pairs"]
pairs_cols = ["m_pairs", "im_pairs", "ic_pairs"]
coverage_goals_cols = list(map(humanize, coverage_goals_cols))
# coverage goals
f, ax = plt.subplots(figsize=(10, 5))
sns.boxplot(data=df.loc[:, coverage_goals_cols])
if for_presentation:
make_box_plot_white(ax)
x_label = "Goal types"
y_label = "Goals count"
ax.set(xlabel=x_label, ylabel=y_label)
plt.tight_layout()
plt.savefig("coverage_goals_all_unique.png",
dpi=300,
transparent=transparent_bg)
# pair detailed
pairs_cols = list(map(humanize, pairs_cols))
f, ax = plt.subplots(figsize=(10, 5))
sns.boxplot(data=df.loc[:, pairs_cols])
if for_presentation:
make_box_plot_white(ax)
x_label = "Definition-use pair types"
y_label = "Pairs count"
ax.set(xlabel=x_label, ylabel=y_label)
plt.tight_layout()
plt.savefig("pairs_detailed_all_unique.png",
dpi=300,
transparent=transparent_bg)
pragmas={
"journal_mode": "wal",
"cache_size": -64 * 1000,
"foreign_key": 1,
"ignore_check_constraints": 9,
"synchronous": 0,
}
)
DATABASE_PROXY.initialize(db)
ms: List[Module] = Module.select() \
.where(Module.st_cov > 70) \
.where(Module.du_cov > 70) \
.where(Module.br_cov > 70) \
.where(Module.bugs < 8) \
.where(Module.statements > 50) \
.where(Module.total_cases > 20) \
.where(Module.total_cases < 1000) \
.where(Module.is_full_cfg == 1) \
.group_by(Module.path) \
.order_by(Module.total_cases.desc()) \
# .limit(10)
print(len(list(ms)))
# exit()
print("\n".join(list(
map(lambda t: " ".join(t),
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))
database = "../selection.db"
db = SqliteExtDatabase(database,
pragmas={
"journal_mode": "wal",
"cache_size": -64 * 1000,
"foreign_key": 1,
"ignore_check_constraints": 9,
"synchronous": 0,
})
DATABASE_PROXY.initialize(db)
# .where(Module.total_cases < 100) \
# .where(Module.bugs < 8) \
ms: List[Module] = Module.select() \
.join(Repo) \
.group_by(Repo.name) \
.order_by(Module.total_cases.desc()) \
.offset(off) \
.limit(lim)
ms = list(ms)
repos = []
for m in ms:
repo_id = m.repo.id
repo_id = repo_id.replace("@", "_")
df = read_combined_df(data_path, repo_id, DataFrameType.FIXED_SIZE)
if df is not None:
# print(df)
metrics = df[METRIC].unique()
if len(metrics) < 3:
continue
by_metric = df.groupby(METRIC)[SUITE_COVERAGE].agg(["min", "max"])