481
|
1 |
// Core Part 2 about Code Similarity
|
|
2 |
//===================================
|
211
|
3 |
|
320
|
4 |
|
401
|
5 |
object C2 {
|
211
|
6 |
|
481
|
7 |
// ADD YOUR CODE BELOW
|
|
8 |
//======================
|
|
9 |
|
|
10 |
//(1)
|
|
11 |
def clean(s: String) : List[String] = """(\w+)""".r.findAllIn(s).toList
|
|
12 |
|
211
|
13 |
|
481
|
14 |
|
|
15 |
//(2)
|
|
16 |
def occurrences(xs: List[String]): Map[String, Int] = {
|
|
17 |
val ls = xs.distinct
|
|
18 |
val occLs = for (s <- ls) yield (s, xs.count(_.equals(s)))
|
|
19 |
occLs.toMap
|
|
20 |
}
|
320
|
21 |
|
211
|
22 |
|
481
|
23 |
//(3)
|
401
|
24 |
def prod(lst1: List[String], lst2: List[String]) : Int = {
|
481
|
25 |
val occM1 = occurrences(lst1)
|
|
26 |
val occM2 = occurrences(lst2)
|
|
27 |
(for (s <- occM1) yield s._2 * occM2.getOrElse(s._1,0)).sum
|
401
|
28 |
}
|
|
29 |
|
481
|
30 |
|
|
31 |
//(4)
|
|
32 |
def overlap(lst1: List[String], lst2: List[String]) : Double = prod(lst1,lst2) / prod(lst1,lst1).max(prod(lst2,lst2))
|
|
33 |
|
|
34 |
def similarity(s1: String, s2: String) : Double = overlap(clean(s1), clean(s2))
|
211
|
35 |
|
|
36 |
|
481
|
37 |
|
|
38 |
/* Test cases
|
320
|
39 |
|
|
40 |
|
401
|
41 |
val list1 = List("a", "b", "b", "c", "d")
|
211
|
42 |
val list2 = List("d", "b", "d", "b", "d")
|
401
|
43 |
|
|
44 |
occurrences(List("a", "b", "b", "c", "d")) // Map(a -> 1, b -> 2, c -> 1, d -> 1)
|
|
45 |
occurrences(List("d", "b", "d", "b", "d")) // Map(d -> 3, b -> 2)
|
|
46 |
|
|
47 |
prod(list1,list2) // 7
|
|
48 |
|
211
|
49 |
overlap(list1, list2) // 0.5384615384615384
|
|
50 |
overlap(list2, list1) // 0.5384615384615384
|
|
51 |
overlap(list1, list1) // 1.0
|
|
52 |
overlap(list2, list2) // 1.0
|
401
|
53 |
|
|
54 |
// Plagiarism examples from
|
211
|
55 |
// https://desales.libguides.com/avoidingplagiarism/examples
|
401
|
56 |
|
211
|
57 |
val orig1 = """There is a strong market demand for eco-tourism in
|
|
58 |
Australia. Its rich and diverse natural heritage ensures Australia's
|
|
59 |
capacity to attract international ecotourists and gives Australia a
|
|
60 |
comparative advantage in the highly competitive tourism industry."""
|
401
|
61 |
|
211
|
62 |
val plag1 = """There is a high market demand for eco-tourism in
|
|
63 |
Australia. Australia has a comparative advantage in the highly
|
|
64 |
competitive tourism industry due to its rich and varied natural
|
|
65 |
heritage which ensures Australia's capacity to attract international
|
|
66 |
ecotourists."""
|
401
|
67 |
|
481
|
68 |
similarity(orig1, plag1) // 0.8679245283018868
|
401
|
69 |
|
|
70 |
|
|
71 |
// Plagiarism examples from
|
211
|
72 |
// https://www.utc.edu/library/help/tutorials/plagiarism/examples-of-plagiarism.php
|
401
|
73 |
|
211
|
74 |
val orig2 = """No oil spill is entirely benign. Depending on timing and
|
|
75 |
location, even a relatively minor spill can cause significant harm to
|
|
76 |
individual organisms and entire populations. Oil spills can cause
|
|
77 |
impacts over a range of time scales, from days to years, or even
|
|
78 |
decades for certain spills. Impacts are typically divided into acute
|
|
79 |
(short-term) and chronic (long-term) effects. Both types are part of a
|
|
80 |
complicated and often controversial equation that is addressed after
|
|
81 |
an oil spill: ecosystem recovery."""
|
401
|
82 |
|
211
|
83 |
val plag2 = """There is no such thing as a "good" oil spill. If the
|
|
84 |
time and place are just right, even a small oil spill can cause damage
|
|
85 |
to sensitive ecosystems. Further, spills can cause harm days, months,
|
|
86 |
years, or even decades after they occur. Because of this, spills are
|
|
87 |
usually broken into short-term (acute) and long-term (chronic)
|
|
88 |
effects. Both of these types of harm must be addressed in ecosystem
|
|
89 |
recovery: a controversial tactic that is often implemented immediately
|
|
90 |
following an oil spill."""
|
401
|
91 |
|
481
|
92 |
overlap(clean(orig2), clean(plag2)) // 0.728
|
|
93 |
similarity(orig2, plag2) // 0.728
|
|
94 |
|
401
|
95 |
|
481
|
96 |
|
401
|
97 |
// The punchline: everything above 0.6 looks suspicious and
|
481
|
98 |
// should be investigated by staff.
|
401
|
99 |
|
211
|
100 |
*/
|
|
101 |
|
320
|
102 |
}
|