Update plant-genomic-benchmark.py
Browse files- plant-genomic-benchmark.py +31 -8
plant-genomic-benchmark.py
CHANGED
@@ -4,7 +4,6 @@
|
|
4 |
|
5 |
import os
|
6 |
from typing import List
|
7 |
-
from Bio import SeqIO
|
8 |
import datasets
|
9 |
|
10 |
|
@@ -69,6 +68,25 @@ _TASK_INFO = {'poly_a':{'type': 'binary', 'val_set':False},
|
|
69 |
'pro_seq':{'type':'binary','val_set':True}
|
70 |
}
|
71 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
72 |
class AgroNtTasksConfig(datasets.BuilderConfig):
|
73 |
"""BuilderConfig for the Agro NT supervised learning tasks dataset."""
|
74 |
|
@@ -167,24 +185,29 @@ class AgroNtTasks(datasets.GeneratorBasedBuilder):
|
|
167 |
|
168 |
# method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
|
169 |
def _generate_examples(self, filepath):
|
170 |
-
|
171 |
-
|
172 |
-
for record in SeqIO.parse(f,'fasta'):
|
173 |
-
# Yields examples as (key, example) tuples
|
174 |
|
175 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
176 |
name = split_name[0]
|
177 |
labels = split_name[1:]
|
178 |
|
179 |
if 'multi' in self.config.task_type:
|
180 |
yield key, {
|
181 |
-
"sequence":
|
182 |
"name": name,
|
183 |
"labels": labels
|
184 |
}
|
185 |
else:
|
186 |
yield key, {
|
187 |
-
"sequence":
|
188 |
"name": name,
|
189 |
"label": labels[0],
|
190 |
}
|
|
|
4 |
|
5 |
import os
|
6 |
from typing import List
|
|
|
7 |
import datasets
|
8 |
|
9 |
|
|
|
68 |
'pro_seq':{'type':'binary','val_set':True}
|
69 |
}
|
70 |
|
71 |
+
|
72 |
+
# This function is a basic reimplementation of SeqIO's parse method. This allows the
|
73 |
+
# dataset viewer to work as it does not require an external package.
|
74 |
+
def parse_fasta(fp):
|
75 |
+
name, seq = None, []
|
76 |
+
for line in fp:
|
77 |
+
line = line.rstrip()
|
78 |
+
if line.startswith(">"):
|
79 |
+
if name:
|
80 |
+
# Slice to remove '>'
|
81 |
+
yield (name[1:], "".join(seq))
|
82 |
+
name, seq = line, []
|
83 |
+
else:
|
84 |
+
seq.append(line)
|
85 |
+
if name:
|
86 |
+
# Slice to remove '>'
|
87 |
+
yield (name[1:], "".join(seq))
|
88 |
+
|
89 |
+
|
90 |
class AgroNtTasksConfig(datasets.BuilderConfig):
|
91 |
"""BuilderConfig for the Agro NT supervised learning tasks dataset."""
|
92 |
|
|
|
185 |
|
186 |
# method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
|
187 |
def _generate_examples(self, filepath):
|
188 |
+
|
189 |
+
key = 0
|
|
|
|
|
190 |
|
191 |
+
with open(file, "rt") as f:
|
192 |
+
fasta_sequences = parse_fasta(f)
|
193 |
+
|
194 |
+
for name, seq in fasta_sequences:
|
195 |
+
# Yields examples as (key, example) tuples
|
196 |
+
sequence, name = str(seq), str(name)
|
197 |
+
|
198 |
+
split_name = name.split("|")
|
199 |
name = split_name[0]
|
200 |
labels = split_name[1:]
|
201 |
|
202 |
if 'multi' in self.config.task_type:
|
203 |
yield key, {
|
204 |
+
"sequence": sequence,
|
205 |
"name": name,
|
206 |
"labels": labels
|
207 |
}
|
208 |
else:
|
209 |
yield key, {
|
210 |
+
"sequence": sequence,
|
211 |
"name": name,
|
212 |
"label": labels[0],
|
213 |
}
|