def run():
if not git_is_available():
msg = "contribcompl requires git to be installed and accessible on path"
print(msg)
sys.exit(1)
arguments = docopt(__doc__, version=__version__)
if arguments["-v"] == True or arguments["--verbose"] == True:
toggle_verbose_output()
path_to_repo = arguments["<repository>"]
if not (is_git_url(path_to_repo) or is_git_dir(path_to_repo)):
print(__doc__)
sys.exit(1)
if is_git_url(path_to_repo):
path_to_repo = clone_to_tmp(path_to_repo)
if arguments["commits"]:
commit_shas = arguments["<commit_sha>"]
elif arguments["issue"]:
issue_re = arguments["<issue_regex>"]
commit_shas = find_commits_for_issue(path_to_repo, issue_re)
# print(commit_shas)
contribcompl = compute_contrib_compl(path_to_repo, commit_shas)
print(contribcompl)
def test_compute_contrib_compl_low():
path_to_repo = "/tmp/cassandra/"
commit_shas = [
"021df085074b761f2b3539355ecfc4c237a54a76",
"2f1d6c7254342af98c2919bd74d37b9944c41a6b",
]
result = compute_contrib_compl(path_to_repo, commit_sha)
assert result == ContributionComplexity.LOW
def main(sys_name):
df = pd.read_csv(f"data/{sys_name}_issues.csv")
if sys_name == "cassandra":
closed_col = "resolved"
issue_key_col = "key"
elif sys_name == "gaffer":
closed_col = "closed_at"
issue_key_col = "number"
# Filter for only resolved issues: 2300 closed issues for Gaffer and
# 14158 closed issues for Cassandra
df = df[~df[closed_col].isnull()]
df.reset_index(drop=True, inplace=True)
# df = df.iloc[:10]
path_to_repo = f"/tmp/{sys_name}"
contribcompls = []
commit_shas_per_contrib = []
for issue_key in df[issue_key_col].values:
if sys_name == "cassandra":
issue_re = f"{issue_key}( |$)"
elif sys_name == "gaffer":
issue_re = f"(Gh |gh-){issue_key}( |$)"
commit_shas = find_commits_for_issue(path_to_repo, issue_re)
commit_shas_per_contrib.append(commit_shas)
print(issue_re, commit_shas, flush=True)
if commit_shas:
try:
contribcompl = compute_contrib_compl(path_to_repo, commit_shas)
except:
print(
f"Skipping {issue_key}",
issue_re,
commit_shas,
type(commit_shas),
flush=True,
)
contribcompl = None
contribcompl = contribcompl.value
else:
contribcompl = None
contribcompls.append(contribcompl)
df["commit_shas"] = commit_shas_per_contrib
df["contrib_complexity"] = contribcompls
# Prevent sha lists from being truncated
np.set_printoptions(threshold=sys.maxsize)
df.to_csv(f"data/{sys_name}_contrib_compl.csv", index=False)
def get_complete_issue_df(sys_name):
procfname = f"data/processing/{sys_name[:3]}_issues.csv"
if os.path.isfile(procfname):
date_cols = ["created", "resolved", "updated"]
df_iss = pd.read_csv(procfname, parse_dates=date_cols)
# parse commits again in
df_iss.commit_shas = [
eval(h.replace("\n", ",")) if type(h) == str else ""
for h in df_iss.commit_shas
]
df_iss.t_lead = pd.to_timedelta(df_iss.t_lead)
return df_iss
# For first time preprocessing
fname_issues = f"data/input/{sys_name[:3]}_issues.csv"
if sys_name == "cassandra":
date_cols = ["created", "resolved", "updated"]
elif sys_name == "gaffer":
date_cols = ["created_at", "closed_at", "updated_at"]
df_iss = pd.read_csv(fname_issues, parse_dates=date_cols)
if sys_name == "cassandra":
# Adjust priority values to same format as for Gaffer
df_iss["priority"] = df_iss.priority.str.lower()
df_iss.rename(columns={"description": "body"}, inplace=True)
df_iss.status = df_iss.status.str.lower()
# Rename resolved to closed as for Gaffer
df_iss[df_iss.status == "resolved"].status = "closed"
burl = "https://issues.apache.org/jira/browse/"
df_iss["url"] = df_iss.key.apply(lambda k: burl + k)
if sys_name == "gaffer":
# Rename the columns to allow for unique code in the following
df_iss.rename(columns={"created_at": "created"}, inplace=True)
df_iss.rename(columns={"closed_at": "resolved"}, inplace=True)
df_iss.rename(columns={"updated_at": "updated"}, inplace=True)
df_iss.rename(columns={"number": "key"}, inplace=True)
# df_iss.rename(columns={"labels": "issue_type"}, inplace=True)
df_iss.rename(columns={"state": "status"}, inplace=True)
# clean the data
df_iss["labels"] = df_iss.labels.apply(labels_to_list)
df_iss["issue_type"] = df_iss.labels.apply(remove_priority)
df_iss["priority"] = df_iss.labels.apply(extract_priority)
# Does not exist in Github Isssue tracker
df_iss["resolution"] = df_iss.issue_type.apply(lambda _: "")
# Compute lead time
df_iss["t_lead"] = df_iss.resolved - df_iss.created
df_iss["t_lead_s"] = df_iss.t_lead.dt.total_seconds()
# df_iss = df_iss.loc[:200]
# Find related commits
path_to_repo = f"{os.getenv('HOME')}/case_systems/{sys_name}"
contribcompls = []
commit_shas_per_contrib = []
for issue_key in tqdm(df_iss["key"].values):
if sys_name == "cassandra":
issue_re = f"{issue_key}( |$)"
elif sys_name == "gaffer":
issue_re = f"(Gh |gh-){issue_key}( |$)"
commit_shas = find_commits_for_issue(path_to_repo, issue_re)
commit_shas_per_contrib.append(commit_shas)
if commit_shas:
try:
contribcompl = compute_contrib_compl(path_to_repo, commit_shas)
except:
print(
f"Skipping {issue_key}",
issue_re,
commit_shas,
type(commit_shas),
flush=True,
)
contribcompl = None
contribcompl = contribcompl.value
else:
contribcompl = None
contribcompls.append(contribcompl)
df_iss["commit_shas"] = commit_shas_per_contrib
# Compute Contribution Complexities ... This takes multiple hours
df_iss["contrib_complexity"] = contribcompls
cols = [
"key",
"title",
"body",
"created",
"updated",
"resolved",
"status",
"issue_type",
"labels",
"priority",
"resolution",
"t_lead",
"t_lead_s",
"url",
"commit_shas",
"contrib_complexity",
]
np.set_printoptions(threshold=sys.maxsize)
df_iss[cols].to_csv(procfname, index=False)
return df_iss[cols]
def test_compute_contrib_compl_high():
path_to_repo = "/tmp/cassandra/"
commit_shas = ["a991b64811f4d6adb6c7b31c0df52288eb06cf19"]
result = compute_contrib_compl(path_to_repo, commit_sha)
assert result == ContributionComplexity.HIGH