Updated the schema of the dataset to contain the indices of the words that are keywords
Browse files- keyword_pubmed.py +18 -8
keyword_pubmed.py
CHANGED
@@ -10,7 +10,7 @@ import datasets
|
|
10 |
|
11 |
class KeywordPubmedDataset(datasets.GeneratorBasedBuilder):
|
12 |
|
13 |
-
VERSION = datasets.Version("1.
|
14 |
|
15 |
|
16 |
|
@@ -28,6 +28,7 @@ class KeywordPubmedDataset(datasets.GeneratorBasedBuilder):
|
|
28 |
"sentence": datasets.Value("string"),
|
29 |
"pmcid": datasets.Value("string"),
|
30 |
"keyword_rank": datasets.Value("int32"),
|
|
|
31 |
}
|
32 |
)
|
33 |
else:
|
@@ -36,6 +37,7 @@ class KeywordPubmedDataset(datasets.GeneratorBasedBuilder):
|
|
36 |
"sentence": datasets.Value("string"),
|
37 |
"pmcid": datasets.Value("string"),
|
38 |
"keyword_rank": datasets.Value("int32"),
|
|
|
39 |
}
|
40 |
)
|
41 |
return datasets.DatasetInfo(
|
@@ -101,14 +103,16 @@ class KeywordPubmedDataset(datasets.GeneratorBasedBuilder):
|
|
101 |
sentence = sentence.strip()
|
102 |
if sentence: # Ignore blanks
|
103 |
sentence = re.sub("\s+", " ", sentence)
|
104 |
-
|
|
|
105 |
if self.config.name == "sentence":
|
106 |
# Yields examples as (key, example) tuples
|
107 |
if has_keyword:
|
108 |
yield item_ix, {
|
109 |
"sentence": sentence,
|
110 |
"keyword_rank": rank,
|
111 |
-
"pmcid": pmcid
|
|
|
112 |
}
|
113 |
item_ix += 1
|
114 |
|
@@ -116,18 +120,24 @@ class KeywordPubmedDataset(datasets.GeneratorBasedBuilder):
|
|
116 |
yield item_ix, {
|
117 |
"sentence": sentence,
|
118 |
"keyword_rank": rank,
|
119 |
-
"pmcid": pmcid
|
|
|
120 |
}
|
121 |
item_ix += 1
|
122 |
|
123 |
-
def
|
124 |
# Lowercase and split the sentence
|
125 |
words = sentence.lower().split()
|
|
|
|
|
126 |
# Check every word until it finds a keyword
|
127 |
-
for word in words:
|
128 |
if word in keywords:
|
129 |
-
|
130 |
-
|
|
|
|
|
|
|
131 |
|
132 |
|
133 |
if __name__ == "__main__":
|
|
|
10 |
|
11 |
class KeywordPubmedDataset(datasets.GeneratorBasedBuilder):
|
12 |
|
13 |
+
VERSION = datasets.Version("1.1.0")
|
14 |
|
15 |
|
16 |
|
|
|
28 |
"sentence": datasets.Value("string"),
|
29 |
"pmcid": datasets.Value("string"),
|
30 |
"keyword_rank": datasets.Value("int32"),
|
31 |
+
"keyword_indices": datasets.Sequence(datasets.Value("int32"))
|
32 |
}
|
33 |
)
|
34 |
else:
|
|
|
37 |
"sentence": datasets.Value("string"),
|
38 |
"pmcid": datasets.Value("string"),
|
39 |
"keyword_rank": datasets.Value("int32"),
|
40 |
+
"keyword_indices": datasets.Sequence(datasets.Value("int32"))
|
41 |
}
|
42 |
)
|
43 |
return datasets.DatasetInfo(
|
|
|
103 |
sentence = sentence.strip()
|
104 |
if sentence: # Ignore blanks
|
105 |
sentence = re.sub("\s+", " ", sentence)
|
106 |
+
kw_indices, rank = self._keyword_indices(sentence, keywords, ranks)
|
107 |
+
has_keyword = rank > -1
|
108 |
if self.config.name == "sentence":
|
109 |
# Yields examples as (key, example) tuples
|
110 |
if has_keyword:
|
111 |
yield item_ix, {
|
112 |
"sentence": sentence,
|
113 |
"keyword_rank": rank,
|
114 |
+
"pmcid": pmcid,
|
115 |
+
"keyword_indices": kw_indices
|
116 |
}
|
117 |
item_ix += 1
|
118 |
|
|
|
120 |
yield item_ix, {
|
121 |
"sentence": sentence,
|
122 |
"keyword_rank": rank,
|
123 |
+
"pmcid": pmcid,
|
124 |
+
"keyword_indices": kw_indices
|
125 |
}
|
126 |
item_ix += 1
|
127 |
|
128 |
+
def _keyword_indices(self, sentence, keywords, ranks):
|
129 |
# Lowercase and split the sentence
|
130 |
words = sentence.lower().split()
|
131 |
+
indices = list()
|
132 |
+
top_rank = -1
|
133 |
# Check every word until it finds a keyword
|
134 |
+
for w_ix, word in enumerate(words):
|
135 |
if word in keywords:
|
136 |
+
indices.append(w_ix)
|
137 |
+
rank = ranks[word]
|
138 |
+
if rank < top_rank or top_rank == -1:
|
139 |
+
top_rank = rank
|
140 |
+
return indices, top_rank
|
141 |
|
142 |
|
143 |
if __name__ == "__main__":
|