def solve(self, fractal):
margin = cfg.getint('Plot', 'margin')
size = cfg.getint('Plot', 'size')
spacing = cfg.getint('Solver', 'service_grid_spacing')
def in_bounds(p):
if p.x < margin or p.x > (size - margin):
return False
if p.y < margin or p.y > (size - margin):
return False
return True
def service_level(eval_point):
return min([(eval_point[0] - p.x) ** 2 + (eval_point[1] - p.y) ** 2 for p in valid_point_set])
point_set = fractal.point_set(1)
valid_point_set = filter(in_bounds, point_set)
eval_set = [(x, y)
for x in range(margin, size - margin, spacing)
for y in range(margin, size - margin, spacing)
]
fractal.service_penalty = float(sum(map(service_level, eval_set)) / len(eval_set))
fractal.length_penalty = 1.0 # max(1.0, (fractal.total_length / len(point_set)))
fractal.complexity_penalty = 1.0 # int(max(0, len(point_set))) ** 2
fractal.bounds_penalty = max(1, (len(point_set) - len(valid_point_set)) * 100) ** 2
fractal.fitness = 1.0 / (fractal.service_penalty + fractal.length_penalty + fractal.complexity_penalty + fractal.bounds_penalty)
# print "fitness: %s" % fractal.fitness
return fractal
class Judger(object):
__SmallFileMaxSize = cfg.getint('ScanFile', 'SmallFile') * 1024
__LastUpdateSeconds = cfg.getint('ScanFile', 'LastUpdate') * 3600
__CodecCheckSize = cfg.getint('ScanFile', 'CodecCheck')
__ExcludedExtensions = cfg.get('ScanFile', 'ExcludedExt').lower().split()
@classmethod
def filter(cls, file_fullname):
try:
size = path.getsize(file_fullname)
last_update = path.getmtime(file_fullname)
if time() - last_update > cls.__LastUpdateSeconds: # long time no update
return 1
if win and last_update <= path.getctime(file_fullname): # not update after create(no create time for linux)
return 1
if size < cls.__SmallFileMaxSize: # too small, looks not like a production log
return 2
if file_fullname[
file_fullname.rfind('.'):].lower() in cls.__ExcludedExtensions: # known file extension, not log
return 3
if (not win) and access(file_fullname, X_OK): # unix executive, not log
return 4
with open(file_fullname, 'rb') as fp: # not txt file, not log
if size > cls.__CodecCheckSize * 2: # 文件中间判断，准确性可能大些
fp.seek(int(size / 2))
charset = detect(fp.read(cls.__CodecCheckSize))
if charset['confidence'] < 0.5:
return 5
return charset
except Exception as err:
log.warning(file_fullname + '\t' + str(err))
return 0
def proxyMain():
port = cfg.getint('http','proxy_port')
numThreads = cfg.getint('http','proxy_threads', 1)
server_address = ('', port)
global proxy_httpd
proxy_httpd = httpserver.PooledHTTPServer(server_address, proxyhandler.ProxyHandler, numThreads)
log.info('Proxy: %s', proxy_httpd.report_config())
proxy_httpd.serve_forever()
def indexMain():
interval = cfg.getint('indexing','interval',3)
log.info('Scheduled index thread to run every %s minutes' % interval)
while not _shutdownEvent.wait(interval * 60):
if _shutdownEvent.isSet():
break
qmsg_processor.backgroundIndexTask()
def register(self, ptype):
"""register ptype as a local typedef"""
# Too many of them leads to memory burst
if len(self.typedefs) < cfg.getint('typing', 'max_combiner'):
self.typedefs.append(ptype)
return True
return False
def register(self, ptype):
"""register ptype as a local typedef"""
# Too many of them leads to memory burst
if len(self.typedefs) < cfg.getint('typing', 'max_combiner'):
self.typedefs.append(ptype)
return True
return False
def adminMain():
port = cfg.getint('http','admin_port')
server_address = ('', port)
global admin_httpd
admin_httpd = httpserver.HTTPServer(server_address, app_httpserver.AppHTTPRequestHandler)
log.info("Start admin on '%s' port %s" % server_address)
app_httpserver.log.info('app_httpserver setup: docBase=%s',
app_httpserver.AppHTTPRequestHandler.docBase)
admin_httpd.serve_forever()
def normalised_segment_set(self):
origin = OriginPoint(self)
depth_limit = cfg.getint("Fractal", "recursion_limit")
def recurse(base):
r = {"x": base.x, "y": base.y, "children": []}
if base.depth < depth_limit and not base.terminate():
for segment in base.segments():
end_point = segment.end()
r["children"].append(recurse(end_point))
return r
segments = recurse(origin)
return segments
def point_set(self, every=1):
self.total_length = 1
origin = OriginPoint(self)
points = []
depth_limit = cfg.getint("Fractal", "recursion_limit")
def recurse(base):
if base.depth % every == 0:
points.append(base)
if base.depth < depth_limit and not base.terminate():
segments = base.segments()
self.total_length += segments[0].length()
for segment in segments:
recurse(segment.end())
recurse(origin)
return points
def __init__(self):
self.fittest = None
self.generation = 0
initial_solution = Solution()
initial_solution.length_function = Expression(init_terms=[createTerm('Constant', innerMultiplier=3.0, outerMultiplier=3.0)])
initial_solution.radiance_function = Expression(init_terms=[createTerm('Constant', innerMultiplier=1.0, outerMultiplier=1.5)])
initial_solution.orientation_function = Expression(init_terms=[createTerm('Constant', innerMultiplier=-0.1, outerMultiplier=0.1)])
initial_solution.termination_function = Expression(init_terms=[createTerm('Constant', innerMultiplier=3.0, outerMultiplier=3.0)])
self.solutions = [initial_solution]
workers = cfg.getint('FitnessTest', 'workers')
if workers > 1:
print "Evaluating using {workers} worker threads".format(workers=workers)
self.map = Pool(processes=workers).map_async
else:
self.map = map
self.max_fitness_acheived = 0
def generate(self, ctx):
# gather all underlying types and make sure they do not appear twice
mct = cfg.getint('typing', 'max_container_type')
all_types = self.all_types()
fot0 = lambda t: type(t) is IndexableType
fot1 = lambda t: type(t) is ContainerType
fit = lambda t: not fot0(t) and not fot1(t)
it = filter(fit, all_types)
ot0 = filter(fot0, all_types)
ot1 = filter(fot1, all_types)
icombined = sorted(set(ctx(t).generate(ctx) for t in it))
lcombined0 = sorted(set(ctx(t).generate(ctx) for t in ot0))[-mct:]
lcombined1 = sorted(set(ctx(t).generate(ctx) for t in ot1))[-mct:]
combined = icombined + lcombined0 + lcombined1
if len(combined) == 1:
return combined[0]
else:
return 'typename __combined<{0}>::type'.format(",".join(combined))
def generate(self, ctx):
# gather all underlying types and make sure they do not appear twice
mct = cfg.getint('typing', 'max_container_type')
all_types = self.all_types()
fot0 = lambda t: type(t) is IndexableType
fot1 = lambda t: type(t) is ContainerType
fit = lambda t: not fot0(t) and not fot1(t)
it = filter(fit, all_types)
ot0 = filter(fot0, all_types)
ot1 = filter(fot1, all_types)
icombined = sorted(set(ctx(t).generate(ctx) for t in it))
lcombined0 = sorted(set(ctx(t).generate(ctx) for t in ot0))[-mct:]
lcombined1 = sorted(set(ctx(t).generate(ctx) for t in ot1))[-mct:]
combined = icombined + lcombined0 + lcombined1
if len(combined) == 1:
return combined[0]
else:
return 'typename __combined<{0}>::type'.format(",".join(combined))
def mutate(self):
mutate_count = 0
# maybe delete a term
if self.terms and random() > (1 - (cfg.getfloat('Function', 'term_deletion_chance') * len(self.terms))):
self._delete_term()
# potentially create a new term. always do so if there are none.
if not self.terms or random() > (1 - cfg.getfloat('Function', 'term_creation_chance')):
self._add_term()
# modify some term constants
if self.terms:
while mutate_count < cfg.getint('Mutator', 'number_of_terms'):
if random() > cfg.getfloat('Mutator', 'probability'):
t = choice(self.terms)
t.mutate()
mutate_count += 1
def draw(self, seq):
# (0,150,255) [0x0096ff] -> (42,22,69) [0x45162a]
def colour_lookup(ratio, shade=False):
r = 000 + (ratio * (42 - 000))
g = 150 + (ratio * (22 - 150))
b = 255 + (ratio * (69 - 255))
if shade:
r /= 3.0
g /= 3.0
b /= 3.0
return "rgb({},{},{})".format(int(r), int(g), int(b))
im = Image.new("RGBA", (cfg.getint("Plot", "size"), cfg.getint("Plot", "size")), (10, 4, 27, 255))
draw = ImageDraw.Draw(im)
draw.line(
(
(cfg.getint("Plot", "margin"), cfg.getint("Plot", "margin")),
(cfg.getint("Plot", "margin"), cfg.getint("Plot", "size") - cfg.getint("Plot", "margin")),
(
cfg.getint("Plot", "size") - cfg.getint("Plot", "margin"),
cfg.getint("Plot", "size") - cfg.getint("Plot", "margin"),
),
(cfg.getint("Plot", "size") - cfg.getint("Plot", "margin"), cfg.getint("Plot", "margin")),
(cfg.getint("Plot", "margin"), cfg.getint("Plot", "margin")),
),
fill="rgb(24,12,54)",
)
points = self.fractal.point_set()
# sort by depth so oldest segments are drawn on top
points.sort(key=lambda p: -p.depth)
# for point in points:
# fill = colour_lookup(float(point.depth) / (points[0].depth + 1), shade=True)
# service_x = (point.x // constants.SERVICE_GRID_SPACING) * constants.SERVICE_GRID_SPACING
# service_y = (point.y // constants.SERVICE_GRID_SPACING) * constants.SERVICE_GRID_SPACING
# draw.rectangle(
# (service_x + 1, service_y + 1, service_x + constants.SERVICE_GRID_SPACING - 1, service_y + constants.SERVICE_GRID_SPACING - 1),
# fill=fill # "rgb(25,20,37,20)"
# )
for point in points:
fill = colour_lookup(float(point.depth) / (points[0].depth + 1))
for segment in point.segments():
end = segment.end()
if end.x >= 0 and end.y >= 0 and end.x <= cfg.get("Plot", "size") and end.y <= cfg.get("Plot", "size"):
draw.line((point.x, point.y, end.x, end.y), fill=fill)
im.save("output/out." + str(seq) + ".png", "PNG")
def _dist_to_origin(self):
try:
return min(cfg.getint("Plot", "size"), hypot(self.x - origin_x, self.y - origin_y), key=abs)
except OverflowError:
return float_info.max
from math import hypot, sin, cos, pi
from sys import float_info
from config import cfg
# calls to config seem to be slow, so 'cache' these here
branch_every = cfg.getint("Branch", "every")
branch_segments = cfg.getint("Branch", "segments")
origin_x = cfg.getint("Fractal", "origin_x")
origin_y = cfg.getint("Fractal", "origin_y")
class Point(object):
def __unicode__(self):
return str(self.x) + ", " + str(self.y)
def __init__(self, x, y, fractal, depth, parent_orientation):
self.x = int(x)
self.y = int(y)
self.fractal = fractal
self.depth = depth
self.dist_to_origin = self._dist_to_origin()
self.parent_orientation = parent_orientation
self.segment_count = 1 if self.depth % branch_every else branch_segments
def _dist_to_origin(self):
try:
return min(cfg.getint("Plot", "size"), hypot(self.x - origin_x, self.y - origin_y), key=abs)
except OverflowError:
return float_info.max
def radiance(self):
class Scanner(object):
__SleepSeconds = cfg.getint('ScanFile', 'Sleep')
__MaxFiles = cfg.getint('ScanFile', 'MaxFiles')
__MaxSeconds = cfg.getint('ScanFile', 'MaxSeconds')
if win: # 根据不同操作系统设置起始扫描目录
__InitialPaths = [
chr(i) + ':\\' for i in range(0x61, 0x7a)
if os.path.isdir(chr(i) + ':\\')
]
__ExcludedPaths = cfg.get('ScanFile', 'ExcludedWin').lower().split()
else:
__InitialPaths = ['/']
__ExcludedPaths = cfg.get('ScanFile', 'ExcludedUnix').lower().split()
def __init__(self,
sample_list_file=os.path.join(cfg.get('Log', 'Folder'),
'samples.lst')):
self.__SampleListFile = sample_list_file
def run(self):
with open(self.__SampleListFile, 'w', encoding='utf-8') as fp:
scaned_files, sampled_files, err_counters = 0, 0, [
0, 0, 0, 0, 0, 0
]
for initial_path in self.__InitialPaths:
for dir_path, dir_names, file_names in os.walk(initial_path):
if False in [
not match(excluded_path, dir_path)
for excluded_path in self.__ExcludedPaths
]: # 跳过例外目录
dir_names[:] = [] # 跳过例外目录的子目录
continue
if not os.access(dir_path,
os.X_OK | os.R_OK): # 有的目录下面的循环拦不住！
log.warning('[Permission Denied:] ' + dir_path)
continue
for dir_name in dir_names: # 对无权进入的子目录，从扫描列表中清除并记录告警日志
dir_fullname = os.path.join(dir_path, dir_name)
if not os.access(dir_fullname, os.X_OK | os.R_OK):
dir_names.remove(dir_name)
log.warning('[Permission denied:] ' + dir_fullname)
if len(file_names
) > self.__MaxFiles: # 目录下文件特别多,很可能是数据文件目录
log.warning('[Too Many Files]( ' +
str(len(file_names)) + '), Ignoring:' +
dir_path)
continue
timer = time.time()
for file_name in file_names:
try:
scaned_files += 1
if scaned_files % 1000 == 0:
log.info(
'Files scaned:[%d], error[%d], inactive[%d], small[%d], wrong-type[%d], non-text[%d], candidate[%d]\t%s'
%
(scaned_files, err_counters[0],
err_counters[1], err_counters[2],
err_counters[3], err_counters[4] +
err_counters[5], sampled_files, dir_path))
if time.time(
) - timer > self.__MaxSeconds: # Too slow to scan a folder
log.warning(
'[Too slow to scan, Ignoring:]( ' +
dir_path)
break
time.sleep(self.__SleepSeconds) # 防止过多占有系统资源
file_fullname = os.path.join(dir_path, file_name)
rc = Judger.filter(file_fullname)
if type(rc) is int: # 该文件不是候选日志，无需采
err_counters[rc] += 1
continue
print(file_fullname, file=fp)
sampled_files += 1
except Exception as err: # 出现过目录/文件名为乱字符导致写fp文件出现字符集异常情况
log.error(str(err))
log.info(
'Finish scan:[%d], error[%d], inactive[%d], small[%d], wrong-type[%d], non-text[%d], candidate[%d]'
% (scaned_files, err_counters[0], err_counters[1], err_counters[2],
err_counters[3], err_counters[4] + err_counters[5],
sampled_files))
class Sampler(object):
__CodecCheckSize = cfg.getint('ScanFile', 'CodecCheck')
__StartLine = cfg.getint('Sample', 'StartingLine')
__EndLine = __StartLine + cfg.getint('Sample', 'SampleLines')
__MaxSize = cfg.getint('Sample', 'MaxSize') * 1024 * 1024
__OutputPath = cfg.get('Sample', 'DataPath')
__OutputFormat = cfg.getint('Sample', 'Format')
__RegularExpFrom = cfg.get('Sample', 'From')
__RegularExpTo = cfg.get('Sample', 'To')
if os.path.exists(__OutputPath):
rmtree(__OutputPath)
sleep(1) # 防止立刻建立目录出错
os.mkdir(__OutputPath)
@classmethod
def sample(cls, files_list):
log.info('Starting Samples %d files' % len(files_list))
if cls.__OutputFormat == 0:
cls.__merge(cls, files_list)
else:
cls.__copy(cls, files_list)
def __merge(self,
file_fullnames): # 列表中文件每个一行的形式输出到os.ipaddress.sample.dat
if win:
output_filename = gethostbyname(gethostname()) + '.samples.dat'
else:
cmd = "ifconfig|grep 'inet addr:'|grep -v '127.0.0.1'|cut -d: -f2|awk '{print $1}'|head -1"
output_filename = os.popen(cmd).read().strip() + '.samples.dat'
with open(os.path.join(self.__OutputPath, output_filename),
'w',
encoding='utf-8') as fp:
for file_fullname in file_fullnames:
log.info('Sampling ' + file_fullname)
current_position = fp.tell()
try:
fp.write('\n' + file_fullname + '\t')
for line in self.__readLine(self, file_fullname):
fp.write(line.replace('\n', '\0'))
except Exception as err:
log.warning(file_fullname + '\t' + str(err))
fp.seek(current_position)
continue
def __copy(self, file_fullnames):
output_file = ''
for input_file in file_fullnames:
log.info('Sampling ' + input_file)
try:
if self.__OutputFormat == 2: # 分目录保存样本文件
if win:
curr_path = self.__OutputPath + os.sep + os.path.split(
input_file)[0].replace(':', '_')
else:
curr_path = self.__OutputPath + os.path.split(
input_file)[0]
os.makedirs(curr_path, exist_ok=True)
output_file = os.path.join(curr_path,
os.path.split(input_file)[1])
else: # 保存在同一目录中，文件名中体现原目录结构
file_name = input_file.replace(os.sep,
'_').replace(':', '_')
output_file = self.__OutputPath + '/' + file_name
with open(output_file, 'w', encoding='utf-8') as fp:
for line in self.__readLine(self, input_file):
fp.write(line)
except Exception as err:
log.warning(input_file + '\t' + str(err))
if os.path.exists(output_file):
os.remove(output_file)
continue
def __readLine(self, file_fullname, encoding='ascii'):
with open(file_fullname, 'rb') as fp:
size = os.path.getsize(file_fullname)
if size > self.__MaxSize: # 出现过几十G大文件、第一行就非常大，导致内存耗尽情况
fp.seek(-self.__MaxSize, 2)
for lines, line_binary in enumerate(fp):
if lines < self.__StartLine:
continue
if lines > self.__EndLine:
break
try:
line = line_binary.decode(encoding=encoding)
log.debug(str(lines) + ' ' + line)
if self.__RegularExpFrom != '':
line = re.sub(self.__RegularExpFrom,
self.__RegularExpTo, line)
yield line
except UnicodeDecodeError:
encoding = detect(line_binary[:self.__CodecCheckSize])[
'encoding'] # 出现过一行10M，本函数不退出的情况
if encoding is None:
raise # 无法识别编码，向上层传递异常
line = line_binary.decode(encoding=encoding)
if self.__RegularExpFrom != '':
line = re.sub(self.__RegularExpFrom,
self.__RegularExpTo, line)
yield line
def _add_term(self):
if len(self.terms) < cfg.getint('Function', 'maximum_terms'):
t = createTerm('Random')
self.terms.append(t)
