Made a few tweaks
[cipher-training.git] / slides / keyword-break.html
1 <!DOCTYPE html>
2 <html>
3 <head>
4 <title>Breaking keyword ciphers</title>
5 <meta http-equiv="Content-Type" content="text/html; charset=UTF-8"/>
6 <style type="text/css">
7 /* Slideshow styles */
8 body {
9 font-size: 20px;
10 }
11 h1, h2, h3 {
12 font-weight: 400;
13 margin-bottom: 0;
14 }
15 h1 { font-size: 3em; }
16 h2 { font-size: 2em; }
17 h3 { font-size: 1.6em; }
18 a, a > code {
19 text-decoration: none;
20 }
21 code {
22 -moz-border-radius: 5px;
23 -web-border-radius: 5px;
24 background: #e7e8e2;
25 border-radius: 5px;
26 font-size: 16px;
27 }
28 .plaintext {
29 background: #272822;
30 color: #80ff80;
31 text-shadow: 0 0 20px #333;
32 padding: 2px 5px;
33 }
34 .ciphertext {
35 background: #272822;
36 color: #ff6666;
37 text-shadow: 0 0 20px #333;
38 padding: 2px 5px;
39 }
40 .indexlink {
41 position: absolute;
42 bottom: 1em;
43 left: 1em;
44 }
45 .float-right {
46 float: right;
47 }
48 </style>
49 </head>
50 <body>
51 <textarea id="source">
52
53 # Breaking keyword ciphers
54
55 a | b | c | d | e | f | g | h | i | j | k | l | m | n | o | p | q | r | s | t | u | v | w | x | y | z
56 --|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|--
57 k | e | y | w | o | r | d | a | b | c | f | g | h | i | j | l | m | n | p | q | s | t | u | v | x | z
58
59 ---
60
61 layout: true
62
63 .indexlink[[Index](index.html)]
64
65 ---
66
67 # Duplicate and extend your `affine_break()` function
68
69 * How to cycle through all the keys? What _are_ all the keys?
70
71 * Look at `words.txt`
72
73 ---
74
75 # Test it.
76
77 * `2013/4a.ciphertext`
78 * `2013/4b.ciphertext`
79
80 This will take a while. Fire up a system monitor. What's wrong?
81
82 ---
83
84 # Python, threads, and the GIL
85
86 Thread-safe shared-memory code is hard.
87
88 Python's Global Interpreter Lock prevents shooting yourself in the foot.
89
90 Where you want true parallelism, need different threads (Python processes).
91
92 * Thread-safe shared-memory code is hard.
93
94 The `multiprocessing` library makes this easier.
95
96 But before we get there, a couple of diversions...
97
98 ---
99
100 # DRYing code
101
102 Three cipher breaking tasks so far.
103
104 All working on the same principle:
105
106 ```
107 find a way to enumerate all the possible keys
108 initialise 'best so far'
109 for each key:
110 decipher message with this key
111 score it
112 if it's better than the best so far:
113 update best so far
114 ```
115
116 Repetition of code is a bad smell.
117
118 Separate out
119
120 * enumerate the keys
121 * score a key
122 * find the key with the best score
123
124 ---
125
126 # map()
127
128 A common task is to apply a function to each item in a sequence, returning a sequence of the results.
129
130 ```python
131 def double(x):
132 return x * 2
133
134 >>> map(double, [1,2,3])
135 [2,4,6]
136 ```
137
138 * `map()` is a higher-order function: its first argument is the function that's applied.
139
140 How can we use this for keyword cipher breaking?
141
142 ---
143
144 # Mapping keyword decipherings
145
146 Define a function that takes a possible key (keyword and cipher type) and returns the key and its fitness.
147
148 * (Also pass in the message and the fitness function)
149
150 Use `map()` and `max()` to find the best key
151
152 ---
153
154 # Arity of print()
155
156 How many arguments does this take?
157
158 How do you write a function that takes this many arguments?
159
160 ---
161
162 # Function arguments
163
164 ## Positional, keyword
165
166 * Common or garden parameters, as you're used to.
167 * `def keyword_encipher(message, keyword, Keyword_wrap_alphabet.from_a):`
168
169 ## Excess positional
170 * `def mean(x, *xs):`
171
172 First number goes in `x`, remaining go in the tuple `xs`
173
174 ## Excess keyword
175
176 * `def myfunc(arg1=0, **kwargs):`
177
178 `kwargs` will be a Dict of the remaining keywords (not `arg1`)
179
180 ---
181
182 # Back to `multiprocessing`
183
184 What does `Pool.starmap()` do?
185
186 ---
187
188 ```python
189 from multiprocessing import Pool
190
191 def keyword_break_mp(message, wordlist=keywords, fitness=Pletters, chunksize=500):
192 helper_args = [??? for word in wordlist] # One tuple for each possible key
193 with Pool() as pool:
194 breaks = pool.starmap(keyword_break_worker, helper_args, chunksize)
195 return max(breaks, key=lambda k: k[1])
196
197 def keyword_break_worker(???):
198 ???
199 return (key, fitness)
200 ```
201
202 * Gotcha: the function in `Pool.starmap()` must be defined at the top level
203 * This is definitely a "feature"
204
205 ---
206
207 # Performance and chunksize
208
209 Try the multiprocessing keyword break. Is it using all the resources?
210
211 Setting `chunksize` is an art.
212
213 ## Map-reduce as a general pattern for multiprocessing
214
215 </textarea>
216 <script src="http://gnab.github.io/remark/downloads/remark-0.6.0.min.js" type="text/javascript">
217 </script>
218
219 <script type="text/javascript"
220 src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML&delayStartupUntil=configured"></script>
221
222 <script type="text/javascript">
223 var slideshow = remark.create({ ratio: "16:9" });
224
225 // Setup MathJax
226 MathJax.Hub.Config({
227 tex2jax: {
228 skipTags: ['script', 'noscript', 'style', 'textarea', 'pre']
229 }
230 });
231 MathJax.Hub.Queue(function() {
232 $(MathJax.Hub.getAllJax()).map(function(index, elem) {
233 return(elem.SourceElement());
234 }).parent().addClass('has-jax');
235 });
236 MathJax.Hub.Configured();
237 </script>
238 </body>
239 </html>