![\"\"](\"https:\/\/www.fagao.me\/images\/uploads\/2019_004.jpg\")
<\/p>\n<\/p>\n
\u641c\u7d22\u662f\u5927\u6570\u636e\u9886\u57df\u91cc\u5e38\u89c1\u7684\u9700\u6c42\u3002Splunk\u548cELK\u5206\u522b\u662f\u8be5\u9886\u57df\u5728\u975e\u5f00\u6e90\u548c\u5f00\u6e90\u9886\u57df\u91cc\u7684\u9886\u5bfc\u8005\u3002\u672c\u6587\u5229\u7528\u5f88\u5c11\u7684Python\u4ee3\u7801\u5b9e\u73b0\u4e86\u4e00\u4e2a\u57fa\u672c\u7684\u6570\u636e\u641c\u7d22\u529f\u80fd\uff0c\u8bd5\u56fe\u8ba9\u5927\u5bb6\u7406\u89e3\u5927\u6570\u636e\u641c\u7d22\u7684\u57fa\u672c\u539f\u7406\u3002<\/p>\n
\u5e03\u9686\u8fc7\u6ee4\u5668 \uff08Bloom Filter\uff09<\/p>\n
\u7b2c\u4e00\u6b65\u6211\u4eec\u5148\u8981\u5b9e\u73b0\u4e00\u4e2a\u5e03\u9686\u8fc7\u6ee4\u5668\u3002<\/p>\n
\u5e03\u9686\u8fc7\u6ee4\u5668\u662f\u5927\u6570\u636e\u9886\u57df\u7684\u4e00\u4e2a\u5e38\u89c1\u7b97\u6cd5\uff0c\u5b83\u7684\u76ee\u7684\u662f\u8fc7\u6ee4\u6389\u90a3\u4e9b\u4e0d\u662f\u76ee\u6807\u7684\u5143\u7d20\u3002\u4e5f\u5c31\u662f\u8bf4\u5982\u679c\u4e00\u4e2a\u8981\u641c\u7d22\u7684\u8bcd\u5e76\u4e0d\u5b58\u5728\u4e0e\u6211\u7684\u6570\u636e\u4e2d\uff0c\u90a3\u4e48\u5b83\u53ef\u4ee5\u4ee5\u5f88\u5feb\u7684\u901f\u5ea6\u8fd4\u56de\u76ee\u6807\u4e0d\u5b58\u5728\u3002<\/p>\n
\u8ba9\u6211\u4eec\u770b\u770b\u4ee5\u4e0b\u5e03\u9686\u8fc7\u6ee4\u5668\u7684\u4ee3\u7801\uff1a<\/p>\n
\nclassBloomfilter(object):<\/p>\n
“””<\/p>\n
A Bloom filter is a probabilistic data-structure that trades space for accuracy<\/p>\n
when determining if a value is in a set. It can tell you if a value was possibly<\/p>\n
added, or if it was definitely not added, but it can’t tell you for certain that<\/p>\n
it was added.<\/p>\n
“””<\/p>\n
def __init__(self,size):<\/p>\n
“””Setup the BF with the appropriate size”””<\/p>\n
self.values= [False]* size<\/p>\n
self.size= size<\/p>\n
def hash_value(self,value):<\/p>\n
“””Hash the value provided and scale it to fit the BF size”””<\/p>\n
returnhash(value)% self.size<\/p>\n
def add_value(self,value):<\/p>\n
“””Add a value to the BF”””<\/p>\n
h= self.hash_value(value)<\/p>\n
self.values[h]= True<\/p>\n
def might_contain(self,value):<\/p>\n
“””Check if the value might be in the BF”””<\/p>\n
h= self.hash_value(value)<\/p>\n
returnself.values[h]<\/p>\n
def print_contents(self):<\/p>\n
“””Dump the contents of the BF for debugging purposes”””<\/p>\n
print self.values<\/p>\n<\/blockquote>\n
\u57fa\u672c\u7684\u6570\u636e\u7ed3\u6784\u662f\u4e2a\u6570\u7ec4\uff08\u5b9e\u9645\u4e0a\u662f\u4e2a\u4f4d\u56fe\uff0c\u75281\/0\u6765\u8bb0\u5f55\u6570\u636e\u662f\u5426\u5b58\u5728\uff09\uff0c\u521d\u59cb\u5316\u662f\u6ca1\u6709\u4efb\u4f55\u5185\u5bb9\uff0c\u6240\u4ee5\u5168\u90e8\u7f6eFalse\u3002\u5b9e\u9645\u7684\u4f7f\u7528\u5f53\u4e2d\uff0c\u8be5\u6570\u7ec4\u7684\u957f\u5ea6\u662f\u975e\u5e38\u5927\u7684\uff0c\u4ee5\u4fdd\u8bc1\u6548\u7387\u3002<\/p>\n
\u5229\u7528\u54c8\u5e0c\u7b97\u6cd5\u6765\u51b3\u5b9a\u6570\u636e\u5e94\u8be5\u5b58\u5728\u54ea\u4e00\u4f4d\uff0c\u4e5f\u5c31\u662f\u6570\u7ec4\u7684\u7d22\u5f15<\/p>\n
\u5f53\u4e00\u4e2a\u6570\u636e\u88ab\u52a0\u5165\u5230\u5e03\u9686\u8fc7\u6ee4\u5668\u7684\u65f6\u5019\uff0c\u8ba1\u7b97\u5b83\u7684\u54c8\u5e0c\u503c\u7136\u540e\u628a\u76f8\u5e94\u7684\u4f4d\u7f6e\u4e3aTrue<\/p>\n
\u5f53\u68c0\u67e5\u4e00\u4e2a\u6570\u636e\u662f\u5426\u5df2\u7ecf\u5b58\u5728\u6216\u8005\u8bf4\u88ab\u7d22\u5f15\u8fc7\u7684\u65f6\u5019\uff0c\u53ea\u8981\u68c0\u67e5\u5bf9\u5e94\u7684\u54c8\u5e0c\u503c\u6240\u5728\u7684\u4f4d\u7684True\uff0fFasle<\/p>\n
\u770b\u5230\u8fd9\u91cc\uff0c\u5927\u5bb6\u5e94\u8be5\u53ef\u4ee5\u770b\u51fa\uff0c\u5982\u679c\u5e03\u9686\u8fc7\u6ee4\u5668\u8fd4\u56deFalse\uff0c\u90a3\u4e48\u6570\u636e\u4e00\u5b9a\u662f\u6ca1\u6709\u7d22\u5f15\u8fc7\u7684\uff0c\u7136\u800c\u5982\u679c\u8fd4\u56deTrue\uff0c\u90a3\u4e5f\u4e0d\u80fd\u8bf4\u6570\u636e\u4e00\u5b9a\u5c31\u5df2\u7ecf\u88ab\u7d22\u5f15\u8fc7\u3002\u5728\u641c\u7d22\u8fc7\u7a0b\u4e2d\u4f7f\u7528\u5e03\u9686\u8fc7\u6ee4\u5668\u53ef\u4ee5\u4f7f\u5f97\u5f88\u591a\u6ca1\u6709\u547d\u4e2d\u7684\u641c\u7d22\u63d0\u524d\u8fd4\u56de\u6765\u63d0\u9ad8\u6548\u7387\u3002<\/p>\n
\u6211\u4eec\u770b\u770b\u8fd9\u6bb5 code\u662f\u5982\u4f55\u8fd0\u884c\u7684\uff1a<\/p>\n
\nbf= Bloomfilter(10)<\/p>\n
bf.add_value(‘dog’)<\/p>\n
bf.add_value(‘fish’)<\/p>\n
bf.add_value(‘cat’)<\/p>\n
bf.print_contents()<\/p>\n
bf.add_value(‘bird’)<\/p>\n
bf.print_contents()<\/p>\n
# Note: contents are unchanged after adding bird – it collides<\/p>\n
forterm in[‘dog’,’fish’,’cat’,’bird’,’duck’,’emu’]:<\/p>\n
print'{}: {} {}’.format(term,bf.hash_value(term),bf.might_contain(term))<\/p>\n<\/blockquote>\n
\u7ed3\u679c\uff1a<\/p>\n
\n[False,False,False,False,True,True,False,False,False,True]<\/p>\n
[False,False,False,False,True,True,False,False,False,True]<\/p>\n
dog: 5True<\/p>\n
fish: 4True<\/p>\n
cat: 9True<\/p>\n
bird: 9True<\/p>\n
duck: 5True<\/p>\n
emu: 8False<\/p>\n<\/blockquote>\n
\u9996\u5148\u521b\u5efa\u4e86\u4e00\u4e2a\u5bb9\u91cf\u4e3a10\u7684\u7684\u5e03\u9686\u8fc7\u6ee4\u5668<\/p>\n<\/p>\n
\u7136\u540e\u5206\u522b\u52a0\u5165 \u2018dog\u2019\uff0c\u2018fish\u2019\uff0c\u2018cat\u2019\u4e09\u4e2a\u5bf9\u8c61\uff0c\u8fd9\u65f6\u7684\u5e03\u9686\u8fc7\u6ee4\u5668\u7684\u5185\u5bb9\u5982\u4e0b\uff1a<\/p>\n<\/p>\n
\u7136\u540e\u52a0\u5165\u2018bird\u2019\u5bf9\u8c61\uff0c\u5e03\u9686\u8fc7\u6ee4\u5668\u7684\u5185\u5bb9\u5e76\u6ca1\u6709\u6539\u53d8\uff0c\u56e0\u4e3a\u2018bird\u2019\u548c\u2018fish\u2019\u6070\u597d\u62e5\u6709\u76f8\u540c\u7684\u54c8\u5e0c\u3002<\/p>\n<\/p>\n
\u6700\u540e\u6211\u4eec\u68c0\u67e5\u4e00\u5806\u5bf9\u8c61\uff08\u2019dog\u2019, \u2018fish\u2019, \u2018cat\u2019, \u2018bird\u2019, \u2018duck\u2019, \u2019emu\u2019\uff09\u662f\u4e0d\u662f\u5df2\u7ecf\u88ab\u7d22\u5f15\u4e86\u3002\u7ed3\u679c\u53d1\u73b0\u2018duck\u2019\u8fd4\u56deTrue\uff0c2\u800c\u2018emu\u2019\u8fd4\u56deFalse\u3002\u56e0\u4e3a\u2018duck\u2019\u7684\u54c8\u5e0c\u6070\u597d\u548c\u2018dog\u2019\u662f\u4e00\u6837\u7684\u3002<\/p>\n<\/p>\n
\u5206\u8bcd<\/p>\n
\u4e0b\u9762\u4e00\u6b65\u6211\u4eec\u8981\u5b9e\u73b0\u5206\u8bcd\u3002 \u5206\u8bcd\u7684\u76ee\u7684\u662f\u8981\u628a\u6211\u4eec\u7684\u6587\u672c\u6570\u636e\u5206\u5272\u6210\u53ef\u641c\u7d22\u7684\u6700\u5c0f\u5355\u5143\uff0c\u4e5f\u5c31\u662f\u8bcd\u3002\u8fd9\u91cc\u6211\u4eec\u4e3b\u8981\u9488\u5bf9\u82f1\u8bed\uff0c\u56e0\u4e3a\u4e2d\u6587\u7684\u5206\u8bcd\u6d89\u53ca\u5230\u81ea\u7136\u8bed\u8a00\u5904\u7406\uff0c\u6bd4\u8f83\u590d\u6742\uff0c\u800c\u82f1\u6587\u57fa\u672c\u53ea\u8981\u7528\u6807\u70b9\u7b26\u53f7\u5c31\u597d\u4e86\u3002<\/p>\n
\u4e0b\u9762\u6211\u4eec\u770b\u770b\u5206\u8bcd\u7684\u4ee3\u7801\uff1a<\/p>\n
\ndef major_segments(s):<\/p>\n
“””<\/p>\n
Perform major segmenting on a string. Split the string by all of the major<\/p>\n
breaks, and return the set of everything found. The breaks in this implementation<\/p>\n
are single characters, but in Splunk proper they can be multiple characters.<\/p>\n
A set is used because ordering doesn’t matter, and duplicates are bad.<\/p>\n
“””<\/p>\n
major_breaks= ‘ ‘<\/p>\n
last= -1<\/p>\n
results= set()<\/p>\n
# enumerate() will give us (0, s[0]), (1, s[1]), …<\/p>\n
foridx,ch inenumerate(s):<\/p>\n
ifch inmajor_breaks:<\/p>\n
segment= s[last+1:idx]<\/p>\n
results.add(segment)<\/p>\n
last= idx<\/p>\n
# The last character may not be a break so always capture<\/p>\n
# the last segment (which may end up being “”, but yolo)<\/p>\n
segment= s[last+1:]<\/p>\n
results.add(segment)<\/p>\n
returnresults<\/p>\n<\/blockquote>\n
\u4e3b\u8981\u5206\u5272<\/p>\n
\u4e3b\u8981\u5206\u5272\u4f7f\u7528\u7a7a\u683c\u6765\u5206\u8bcd\uff0c\u5b9e\u9645\u7684\u5206\u8bcd\u903b\u8f91\u4e2d\uff0c\u8fd8\u4f1a\u6709\u5176\u5b83\u7684\u5206\u9694\u7b26\u3002\u4f8b\u5982Splunk\u7684\u7f3a\u7701\u5206\u5272\u7b26\u5305\u62ec\u4ee5\u4e0b\u8fd9\u4e9b\uff0c\u7528\u6237\u4e5f\u53ef\u4ee5\u5b9a\u4e49\u81ea\u5df1\u7684\u5206\u5272\u7b26\u3002<\/p>\n
] < >( ) { } | ! ; , \u2018 \u201d * s & ? + %21 %26 %2526 %3B %7C %20 %2B %3D \u2014 %2520 %5D %5B %3A %0A %2C %28 %29<\/p>\n
\ndef minor_segments(s):<\/p>\n
“””<\/p>\n
Perform minor segmenting on a string. This is like major<\/p>\n
segmenting, except it also captures from the start of the<\/p>\n
input to each break.<\/p>\n
“””<\/p>\n
minor_breaks= ‘_.’<\/p>\n
last= -1<\/p>\n
results= set()<\/p>\n
foridx,ch inenumerate(s):<\/p>\n
ifch inminor_breaks:<\/p>\n
segment= s[last+1:idx]<\/p>\n
results.add(segment)<\/p>\n
segment= s[:idx]<\/p>\n
results.add(segment)<\/p>\n
last= idx<\/p>\n
segment= s[last+1:]<\/p>\n
results.add(segment)<\/p>\n
results.add(s)<\/p>\n
returnresults<\/p>\n<\/blockquote>\n
\u6b21\u8981\u5206\u5272<\/p>\n
\u6b21\u8981\u5206\u5272\u548c\u4e3b\u8981\u5206\u5272\u7684\u903b\u8f91\u7c7b\u4f3c\uff0c\u53ea\u662f\u8fd8\u4f1a\u628a\u4ece\u5f00\u59cb\u90e8\u5206\u5230\u5f53\u524d\u5206\u5272\u7684\u7ed3\u679c\u52a0\u5165\u3002\u4f8b\u5982\u201c1.2.3.4\u201d\u7684\u6b21\u8981\u5206\u5272\u4f1a\u67091\uff0c2\uff0c3\uff0c4\uff0c1.2\uff0c1.2.3<\/p>\n
\ndef segments(event):<\/p>\n
“””Simple wrapper around major_segments \/ minor_segments”””<\/p>\n
results= set()<\/p>\n
formajor inmajor_segments(event):<\/p>\n
forminor inminor_segments(major):<\/p>\n
results.add(minor)<\/p>\n
returnresults<\/p>\n<\/blockquote>\n
\u5206\u8bcd\u7684\u903b\u8f91\u5c31\u662f\u5bf9\u6587\u672c\u5148\u8fdb\u884c\u4e3b\u8981\u5206\u5272\uff0c\u5bf9\u6bcf\u4e00\u4e2a\u4e3b\u8981\u5206\u5272\u5728\u8fdb\u884c\u6b21\u8981\u5206\u5272\u3002\u7136\u540e\u628a\u6240\u6709\u5206\u51fa\u6765\u7684\u8bcd\u8fd4\u56de\u3002<\/p>\n
\u6211\u4eec\u770b\u770b\u8fd9\u6bb5 code\u662f\u5982\u4f55\u8fd0\u884c\u7684\uff1a<\/p>\n
\nforterm insegments(‘src_ip = 1.2.3.4’):<\/p>\n
print term<\/p>\n<\/blockquote>\n
\nsrc<\/p>\n
1.2<\/p>\n
1.2.3.4<\/p>\n
src_ip<\/p>\n
3<\/p>\n
1<\/p>\n
1.2.3<\/p>\n
ip<\/p>\n
2<\/p>\n
=<\/p>\n
4<\/p>\n<\/blockquote>\n
\u641c\u7d22<\/p>\n
\u597d\u4e86\uff0c\u6709\u4e2a\u5206\u8bcd\u548c\u5e03\u9686\u8fc7\u6ee4\u5668\u8fd9\u4e24\u4e2a\u5229\u5668\u7684\u652f\u6491\u540e\uff0c\u6211\u4eec\u5c31\u53ef\u4ee5\u6765\u5b9e\u73b0\u641c\u7d22\u7684\u529f\u80fd\u4e86\u3002<\/p>\n
\u4e0a\u4ee3\u7801\uff1a<\/p>\n
\nclassSplunk(object):<\/p>\n
def __init__(self):<\/p>\n
self.bf= Bloomfilter(64)<\/p>\n
self.terms= {}# Dictionary of term to set of events<\/p>\n
self.events= []<\/p>\n
def add_event(self,event):<\/p>\n
“””Adds an event to this object”””<\/p>\n
# Generate a unique ID for the event, and save it<\/p>\n
event_id= len(self.events)<\/p>\n
self.events.append(event)<\/p>\n
# Add each term to the bloomfilter, and track the event by each term<\/p>\n
forterm insegments(event):<\/p>\n
self.bf.add_value(term)<\/p>\n
ifterm notinself.terms:<\/p>\n
self.terms[term]= set()<\/p>\n
self.terms[term].add(event_id)<\/p>\n
def search(self,term):<\/p>\n
“””Search for a single term, and yield all the events that contain it”””<\/p>\n
# In Splunk this runs in O(1), and is likely to be in filesystem cache (memory)<\/p>\n
ifnotself.bf.might_contain(term):<\/p>\n
return<\/p>\n
# In Splunk this probably runs in O(log N) where N is the number of terms in the tsidx<\/p>\n
ifterm notinself.terms:<\/p>\n
return<\/p>\n
forevent_id insorted(self.terms[term]):<\/p>\n
yield self.events[event_id]<\/p>\n<\/blockquote>\n
Splunk\u4ee3\u8868\u4e00\u4e2a\u62e5\u6709\u641c\u7d22\u529f\u80fd\u7684\u7d22\u5f15\u96c6\u5408<\/p>\n
\u6bcf\u4e00\u4e2a\u96c6\u5408\u4e2d\u5305\u542b\u4e00\u4e2a\u5e03\u9686\u8fc7\u6ee4\u5668\uff0c\u4e00\u4e2a\u5012\u6392\u8bcd\u8868\uff08\u5b57\u5178\uff09\uff0c\u548c\u4e00\u4e2a\u5b58\u50a8\u6240\u6709\u4e8b\u4ef6\u7684\u6570\u7ec4<\/p>\n
\u5f53\u4e00\u4e2a\u4e8b\u4ef6\u88ab\u52a0\u5165\u5230\u7d22\u5f15\u7684\u65f6\u5019\uff0c\u4f1a\u505a\u4ee5\u4e0b\u7684\u903b\u8f91<\/p>\n
\u4e3a\u6bcf\u4e00\u4e2a\u4e8b\u4ef6\u751f\u6210\u4e00\u4e2aunqie id\uff0c\u8fd9\u91cc\u5c31\u662f\u5e8f\u53f7<\/p>\n