Edward J. Schwartz commited on
Commit
08c3539
1 Parent(s): de399e6

Add bylibrarydedup

Browse files
Files changed (1) hide show
  1. oo-method-test-split.py +16 -2
oo-method-test-split.py CHANGED
@@ -37,6 +37,11 @@ class OOMethodTestDataset(datasets.ArrowBasedBuilder):
37
  name="bylibrary",
38
  version=datasets.Version("1.0.0"),
39
  description="Split so that library functions (those appearing in >1 exe) are used for training, and non-library functions are used for testing",
 
 
 
 
 
40
  )
41
 
42
  ]
@@ -118,7 +123,7 @@ class OOMethodTestDataset(datasets.ArrowBasedBuilder):
118
 
119
  ]
120
 
121
- elif self.config.name == "bylibrary":
122
  # A function (name) is a library function if it appears in more than one Exename
123
 
124
  # this is (('func', 'oo.exe'): 123)
@@ -133,13 +138,22 @@ class OOMethodTestDataset(datasets.ArrowBasedBuilder):
133
  grouped = {k: [b for _,b in g] for k, g in grouped}
134
 
135
  library_func_names = {f for f, exes in grouped.items() if len(exes) > 1}
 
136
  nonlibrary_func_names = {f for f, exes in grouped.items() if len(exes) == 1}
137
 
 
 
 
 
 
 
 
 
138
  return [
139
  datasets.SplitGenerator(
140
  name="train",
141
  gen_kwargs={
142
- "ds": ds.filter(lambda r: r['Name'] in library_func_names),
143
  },
144
  ),
145
  datasets.SplitGenerator(
 
37
  name="bylibrary",
38
  version=datasets.Version("1.0.0"),
39
  description="Split so that library functions (those appearing in >1 exe) are used for training, and non-library functions are used for testing",
40
+ ),
41
+ datasets.BuilderConfig(
42
+ name="bylibrarydedup",
43
+ version=datasets.Version("1.0.0"),
44
+ description="Split so that library functions (those appearing in >1 exe) are used for training, and non-library functions are used for testing. Only one example per function name is retained.",
45
  )
46
 
47
  ]
 
123
 
124
  ]
125
 
126
+ elif self.config.name in ["bylibrary", "bylibrarydedup"]:
127
  # A function (name) is a library function if it appears in more than one Exename
128
 
129
  # this is (('func', 'oo.exe'): 123)
 
138
  grouped = {k: [b for _,b in g] for k, g in grouped}
139
 
140
  library_func_names = {f for f, exes in grouped.items() if len(exes) > 1}
141
+ library_func_names_dedup = {(f, exes[0]) for f, exes in grouped.items() if len(exes) > 1}
142
  nonlibrary_func_names = {f for f, exes in grouped.items() if len(exes) == 1}
143
 
144
+ train_filter_fun = None
145
+ if self.config.name == "bylibrary":
146
+ train_filter_fun = lambda r: r['Name'] in library_func_names
147
+ elif self.config.name == "bylibrarydedup":
148
+ train_filter_fun = lambda r: (r['Name'], r['Exename']) in library_func_names_dedup
149
+ else:
150
+ assert False, "Invalid configuration"
151
+
152
  return [
153
  datasets.SplitGenerator(
154
  name="train",
155
  gen_kwargs={
156
+ "ds": ds.filter(train_filter_fun),
157
  },
158
  ),
159
  datasets.SplitGenerator(