def err(self):
"""Return error stream. Returns empty string if empty or doesn't exist.
Returns (str) : error stream written to file
"""
if os.path.exists(self.stderr.name):
return read_file(self.stderr.name)
return ""
def out(self):
"""Return output stream. Returns empty string if empty or doesn't exist.
Returns (str) : output stream written to file
"""
if os.path.exists(self.stdout.name):
return read_file(self.stdout.name)
return ""
def _import_annotation(self,
input_file,
username,
stop_line="## Criteria"):
"""A general helper (private) function to import an annotation, meaning
we parse a repository and return additional lines for parsing.
"""
if not username or not input_file:
raise RuntimeError(
"A username and input file are required to import annotation criteria."
)
if not os.path.exists(input_file):
raise FileNotFoundError(input_file)
lines = read_file(input_file)
line = lines.pop(0)
# Find the repository name
while stop_line not in line:
match = re.search(repository_regex, line)
if match:
break
line = lines.pop(0)
# Retrieve the match
if not match:
raise RuntimeError(f"repository pattern not found in {input_file}")
reponame = match.group()
parser = get_parser(reponame, config=self.config)
repo = self.get(parser.uid)
return repo, lines
def bulk_add(self, filename):
"""Given a filename with a single list of repos, add each"""
repos = []
if os.path.exists(filename):
for name in read_file(filename):
uid = name.strip()
repos += [self.add(uid, quiet=True)] or []
return repos
def bulk_update(self, filename, rewrite=False):
"""Given a filename with a single list of repos, add each"""
repos = []
if os.path.exists(filename):
for name in read_file(filename):
uid = name.strip()
try:
repos += [self.update(uid, rewrite=rewrite)]
except RepoNotFoundError:
pass
return repos
def load_criteria(self):
"""Given a repository directory, load criteria files if they exist"""
criteria = {}
for filename in glob(f"{self.parser_dir}/criteria*.tsv"):
uid = (os.path.basename(filename).replace("criteria-",
"").replace(".tsv", ""))
content = read_file(filename)
if uid not in criteria:
criteria[uid] = {}
for row in content:
row = row.strip()
if not row:
continue
username, response = row.split("\t")
criteria[uid][username] = response
return criteria
def load_taxonomy(self):
"""Given a repository directory, load taxonomy annotations if they exist
The taxonomy.tsv file should be a tab separated file with:
username category-unique-id. This means that we keep a record of
who has categorized what, and load this information into the
taxonomy dictionary (organized by the category-unique-id which
then has a total count and list of users).
"""
taxonomy = {}
taxonomy_file = os.path.join(self.parser_dir, "taxonomy.tsv")
if os.path.exists(taxonomy_file):
content = read_file(taxonomy_file)
for row in content:
row = row.strip()
if not row:
continue
username, uids = row.split("\t")
taxonomy[username] = [x.strip() for x in uids.split(",")]
return taxonomy
