def unix(value):
if not _convert:
_late_import()
if isinstance(value, (date, builtin_datetime)):
pass
elif value < 10000000000:
value = _convert.unix2datetime(value)
else:
value = _convert.milli2datetime(value)
return str(_convert.datetime2unix(value))
def unix(value):
if not _convert:
_late_import()
if isinstance(value, (date, builtin_datetime)):
pass
elif value < 10000000000:
value = _convert.unix2datetime(value)
else:
value = _convert.milli2datetime(value)
return str(_convert.datetime2unix(value))
def _scrub(value, is_done):
type = value.__class__
if type in (NoneType, NullType):
return None
elif type in (date, datetime):
return float(datetime2unix(value))
elif type is timedelta:
return value.total_seconds()
elif type is Date:
return float(value.unix)
elif type is Duration:
return value.seconds
elif type is str:
return utf82unicode(value)
elif type is Decimal:
return float(value)
elif isinstance(value, Mapping):
_id = id(value)
if _id in is_done:
Log.error("possible loop in structure detected")
is_done.add(_id)
output = {}
for k, v in value.iteritems():
if not isinstance(k, basestring):
Log.error("keys must be strings")
v = _scrub(v, is_done)
if v != None:
output[k] = v
is_done.discard(_id)
return output
elif type in (list, DictList):
output = []
for v in value:
v = _scrub(v, is_done)
output.append(v)
return output
elif type.__name__ == "bool_": # DEAR ME! Numpy has it's own booleans (value==False could be used, but 0==False in Python. DOH!)
if value == False:
return False
else:
return True
elif hasattr(value, '__json__'):
try:
output = json._default_decoder.decode(value.__json__())
return output
except Exception, e:
Log.error("problem with calling __json__()", e)
def get_changesets(date_range=None, revision_range=None, repo=None):
# GET ALL CHANGESET INFO
args = [
"hg",
"log",
"--cwd",
File(repo.directory).filename,
"-v",
# "-p", # TO GET PATCH CONTENTS
"--style",
TEMPLATE_FILE.filename
]
if date_range is not None:
if date_range.max == None:
if date_range.min == None:
drange = ">0 0"
else:
drange = ">" + unicode(convert.datetime2unix(date_range.min)) + " 0"
else:
if date_range.min == None:
drange = "<" + unicode(convert.datetime2unix(date_range.max) - 1) + " 0"
else:
drange = unicode(convert.datetime2unix(date_range.min)) + " 0 to " + unicode(convert.datetime2unix(date_range.max) - 1) + " 0"
args.extend(["--date", drange])
if revision_range is not None:
args.extend(["-r", str(revision_range.min) + ":" + str(revision_range.max)])
proc = subprocess.Popen(
args,
stdin=subprocess.PIPE,
stdout=subprocess.PIPE,
stderr=subprocess.STDOUT,
bufsize=-1
)
def iterator():
try:
while True:
try:
line = proc.stdout.readline()
if line == '':
proc.wait()
if proc.returncode:
Log.error("Unable to pull hg log: return code {{return_code}}", {
"return_code": proc.returncode
})
return
except Exception, e:
Log.error("Problem getting another line", e)
if line.strip() == "":
continue
Log.note(line)
# changeset = "{date|hgdate|urlescape}\t{node}\t{rev}\t{author|urlescape}\t{branches}\t\t\t\t{p1rev}\t{p1node}\t{parents}\t{children}\t{tags}\t{desc|urlescape}\n"
# branch = "{branch}%0A"
# parent = "{parent}%0A"
# tag = "{tag}%0A"
# child = "{child}%0A"
(
date,
node,
rev,
author,
branches,
files,
file_adds,
file_dels,
p1rev,
p1node,
parents,
children,
tags,
desc
) = (urllib.unquote(c) for c in line.split("\t"))
file_adds = set(file_adds.split("\n")) - {""}
file_dels = set(file_dels.split("\n")) - {""}
files = set(files.split("\n")) - set()
doc = {
"repos": repo.name,
"date": convert.unix2datetime(convert.value2number(date.split(" ")[0])),
"node": node,
"revision": rev,
"author": author,
"branches": set(branches.split("\n")) - {""},
"file_changes": files - file_adds - file_dels - {""},
"file_adds": file_adds,
"file_dels": file_dels,
"parents": set(parents.split("\n")) - {""} | {p1rev+":"+p1node},
"children": set(children.split("\n")) - {""},
"tags": set(tags.split("\n")) - {""},
"description": desc
}
doc = elasticsearch.scrub(doc)
yield doc
except Exception, e:
if isinstance(e, ValueError) and e.message.startswith("need more than "):
Log.error("Problem iterating through log ({{message}})", {
"message": line
}, e)
Log.error("Problem iterating through log", e)
def _scrub(value, is_done):
type_ = value.__class__
if type_ in (NoneType, NullType):
return None
elif type_ is unicode:
value_ = value.strip()
if value_:
return value_
else:
return None
elif type_ is float:
if math.isnan(value) or math.isinf(value):
return None
return value
elif type_ in (int, long, bool):
return value
elif type_ in (date, datetime):
return float(datetime2unix(value))
elif type_ is timedelta:
return value.total_seconds()
elif type_ is Date:
return float(value.unix)
elif type_ is Duration:
return float(value.seconds)
elif type_ is str:
return utf82unicode(value)
elif type_ is Decimal:
return float(value)
elif type_ is Dict:
return _scrub(unwrap(value), is_done)
elif isinstance(value, Mapping):
_id = id(value)
if _id in is_done:
_Log.warning("possible loop in structure detected")
return '"<LOOP IN STRUCTURE>"'
is_done.add(_id)
output = {}
for k, v in value.iteritems():
if isinstance(k, basestring):
pass
elif hasattr(k, "__unicode__"):
k = unicode(k)
else:
_Log.error("keys must be strings")
v = _scrub(v, is_done)
if v != None or isinstance(v, Mapping):
output[k] = v
is_done.discard(_id)
return output
elif type_ in (tuple, list, DictList):
output = []
for v in value:
v = _scrub(v, is_done)
output.append(v)
return output
elif type_ is type:
return value.__name__
elif type_.__name__ == "bool_": # DEAR ME! Numpy has it's own booleans (value==False could be used, but 0==False in Python. DOH!)
if value == False:
return False
else:
return True
elif hasattr(value, '__json__'):
try:
output = json._default_decoder.decode(value.__json__())
return output
except Exception, e:
_Log.error("problem with calling __json__()", e)
def _scrub(value, is_done):
type_ = value.__class__
if type_ in (NoneType, NullType):
return None
elif type_ is unicode:
value_ = value.strip()
if value_:
return value_
else:
return None
elif type_ is float:
if math.isnan(value) or math.isinf(value):
return None
return value
elif type_ in (int, long, bool):
return value
elif type_ in (date, datetime):
return float(datetime2unix(value))
elif type_ is timedelta:
return value.total_seconds()
elif type_ is Date:
return float(value.unix)
elif type_ is Duration:
return float(value.seconds)
elif type_ is str:
return utf82unicode(value)
elif type_ is Decimal:
return float(value)
elif type_ is Dict:
return _scrub(unwrap(value), is_done)
elif isinstance(value, Mapping):
_id = id(value)
if _id in is_done:
_Log.warning("possible loop in structure detected")
return '"<LOOP IN STRUCTURE>"'
is_done.add(_id)
output = {}
for k, v in value.iteritems():
if isinstance(k, basestring):
pass
elif hasattr(k, "__unicode__"):
k = unicode(k)
else:
_Log.error("keys must be strings")
v = _scrub(v, is_done)
if v != None or isinstance(v, Mapping):
output[k] = v
is_done.discard(_id)
return output
elif type_ in (tuple, list, DictList):
output = []
for v in value:
v = _scrub(v, is_done)
output.append(v)
return output
elif type_ is type:
return value.__name__
elif type_.__name__ == "bool_": # DEAR ME! Numpy has it's own booleans (value==False could be used, but 0==False in Python. DOH!)
if value == False:
return False
else:
return True
elif hasattr(value, '__json__'):
try:
output = json._default_decoder.decode(value.__json__())
return output
except Exception, e:
_Log.error("problem with calling __json__()", e)
def get_changesets(date_range=None, revision_range=None, repo=None):
# GET ALL CHANGESET INFO
args = [
"hg",
"log",
"--cwd",
File(repo.directory).filename,
"-v",
# "-p", # TO GET PATCH CONTENTS
"--style",
TEMPLATE_FILE.filename
]
if date_range is not None:
if date_range.max == None:
if date_range.min == None:
drange = ">0 0"
else:
drange = ">" + unicode(convert.datetime2unix(
date_range.min)) + " 0"
else:
if date_range.min == None:
drange = "<" + unicode(
convert.datetime2unix(date_range.max) - 1) + " 0"
else:
drange = unicode(convert.datetime2unix(
date_range.min)) + " 0 to " + unicode(
convert.datetime2unix(date_range.max) - 1) + " 0"
args.extend(["--date", drange])
if revision_range is not None:
args.extend(
["-r",
str(revision_range.min) + ":" + str(revision_range.max)])
proc = subprocess.Popen(args,
stdin=subprocess.PIPE,
stdout=subprocess.PIPE,
stderr=subprocess.STDOUT,
bufsize=-1)
def iterator():
try:
while True:
try:
line = proc.stdout.readline()
if line == '':
proc.wait()
if proc.returncode:
Log.error(
"Unable to pull hg log: return code {{return_code}}",
{"return_code": proc.returncode})
return
except Exception, e:
Log.error("Problem getting another line", e)
if line.strip() == "":
continue
Log.note(line)
# changeset = "{date|hgdate|urlescape}\t{node}\t{rev}\t{author|urlescape}\t{branches}\t\t\t\t{p1rev}\t{p1node}\t{parents}\t{children}\t{tags}\t{desc|urlescape}\n"
# branch = "{branch}%0A"
# parent = "{parent}%0A"
# tag = "{tag}%0A"
# child = "{child}%0A"
(date, node, rev, author, branches, files, file_adds,
file_dels, p1rev, p1node, parents, children, tags,
desc) = (urllib.unquote(c) for c in line.split("\t"))
file_adds = set(file_adds.split("\n")) - {""}
file_dels = set(file_dels.split("\n")) - {""}
files = set(files.split("\n")) - set()
doc = {
"repos":
repo.name,
"date":
convert.unix2datetime(
convert.value2number(date.split(" ")[0])),
"node":
node,
"revision":
rev,
"author":
author,
"branches":
set(branches.split("\n")) - {""},
"file_changes":
files - file_adds - file_dels - {""},
"file_adds":
file_adds,
"file_dels":
file_dels,
"parents":
set(parents.split("\n")) - {""} | {p1rev + ":" + p1node},
"children":
set(children.split("\n")) - {""},
"tags":
set(tags.split("\n")) - {""},
"description":
desc
}
doc = elasticsearch.scrub(doc)
yield doc
except Exception, e:
if isinstance(
e, ValueError) and e.message.startswith("need more than "):
Log.error("Problem iterating through log ({{message}})",
{"message": line}, e)
Log.error("Problem iterating through log", e)
