Update jolma.py
Browse filesremove deepbindweight
jolma.py
CHANGED
@@ -3,7 +3,7 @@ import json
|
|
3 |
import re
|
4 |
import pandas as pd
|
5 |
import datasets
|
6 |
-
|
7 |
|
8 |
logger = datasets.logging.get_logger(__name__)
|
9 |
|
@@ -54,34 +54,26 @@ _URL = "ftp://ftp.sra.ebi.ac.uk/vol1/run/"
|
|
54 |
# _REVERSE_PRIMER = "CCTATGCGTGCTAGTGTGA"
|
55 |
# _DESIGN_LENGTH = 30
|
56 |
|
|
|
|
|
|
|
57 |
|
58 |
-
config = datasets.load_dataset(path="thewall/deepbindweight", split="all")
|
59 |
-
info = pd.read_excel(config['selex'][0], index_col=0)
|
60 |
-
|
61 |
-
_URLS = {}
|
62 |
-
_DESIGN_LENGTH = {}
|
63 |
pattern = re.compile("(\d+)")
|
64 |
-
for idx, row in info.iterrows():
|
65 |
-
sra_id = idx
|
66 |
-
file = row["file"]
|
67 |
-
_URLS[sra_id] = "/".join([_URL, sra_id[:6], sra_id, file])
|
68 |
-
_DESIGN_LENGTH[sra_id] = int(pattern.search(row["Ligand"]).group(0))
|
69 |
|
70 |
|
71 |
class JolmaConfig(datasets.BuilderConfig):
|
72 |
-
def __init__(self,
|
73 |
super(JolmaConfig, self).__init__(**kwargs)
|
74 |
-
self.url = url
|
75 |
-
self.sra_id = sra_id
|
76 |
self.length_match = length_match
|
77 |
-
self.design_length = design_length
|
78 |
self.filter_N = filter_N
|
79 |
-
self.file = file
|
80 |
|
81 |
|
82 |
class Jolma(datasets.GeneratorBasedBuilder):
|
|
|
|
|
|
|
83 |
BUILDER_CONFIGS = [
|
84 |
-
JolmaConfig(name=
|
85 |
]
|
86 |
|
87 |
DEFAULT_CONFIG_NAME = "ERR173157"
|
@@ -101,15 +93,29 @@ class Jolma(datasets.GeneratorBasedBuilder):
|
|
101 |
citation=_CITATION,
|
102 |
)
|
103 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
104 |
def _split_generators(self, dl_manager):
|
105 |
downloaded_files = None
|
|
|
|
|
|
|
|
|
106 |
if getattr(self.config, "data_dir") is not None:
|
107 |
-
|
108 |
-
downloaded_files = dl_manager.extract(
|
109 |
logger.info(f"Load from {downloaded_files}")
|
110 |
if downloaded_files is None or not os.path.exists(downloaded_files):
|
111 |
-
|
112 |
-
|
|
|
113 |
return [
|
114 |
datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloaded_files}),
|
115 |
]
|
@@ -134,7 +140,7 @@ class Jolma(datasets.GeneratorBasedBuilder):
|
|
134 |
|
135 |
def filter_fn(self, example):
|
136 |
seq = example["seq"]
|
137 |
-
if self.config.length_match and len(seq)!=self.
|
138 |
return False
|
139 |
if self.config.filter_N and "N" in seq:
|
140 |
return False
|
@@ -145,5 +151,3 @@ if __name__=="__main__":
|
|
145 |
from datasets import load_dataset
|
146 |
dataset = load_dataset("jolma.py", name="ERR173157", split="all")
|
147 |
# dataset = load_dataset("jolma.py", name="ERR173157", split="all", data_dir="/root/autodl-fs/ERP001824-461")
|
148 |
-
|
149 |
-
|
|
|
3 |
import re
|
4 |
import pandas as pd
|
5 |
import datasets
|
6 |
+
from functools import cached_property
|
7 |
|
8 |
logger = datasets.logging.get_logger(__name__)
|
9 |
|
|
|
54 |
# _REVERSE_PRIMER = "CCTATGCGTGCTAGTGTGA"
|
55 |
# _DESIGN_LENGTH = 30
|
56 |
|
57 |
+
_DOWNLODE_MANAGER = datasets.DownloadManager()
|
58 |
+
_RESOURCE_URL = "https://huggingface.co/datasets/thewall/DeepBindWeight/resolve/main"
|
59 |
+
SELEX_INFO_FILE = _DOWNLODE_MANAGER.download(f"{_RESOURCE_URL}/ERP001824-deepbind.xlsx")
|
60 |
|
|
|
|
|
|
|
|
|
|
|
61 |
pattern = re.compile("(\d+)")
|
|
|
|
|
|
|
|
|
|
|
62 |
|
63 |
|
64 |
class JolmaConfig(datasets.BuilderConfig):
|
65 |
+
def __init__(self, length_match=True, filter_N=True, **kwargs):
|
66 |
super(JolmaConfig, self).__init__(**kwargs)
|
|
|
|
|
67 |
self.length_match = length_match
|
|
|
68 |
self.filter_N = filter_N
|
|
|
69 |
|
70 |
|
71 |
class Jolma(datasets.GeneratorBasedBuilder):
|
72 |
+
|
73 |
+
SELEX_INFO = pd.read_excel(SELEX_INFO_FILE, index_col=0)
|
74 |
+
|
75 |
BUILDER_CONFIGS = [
|
76 |
+
JolmaConfig(name=index) for index in SELEX_INFO.index
|
77 |
]
|
78 |
|
79 |
DEFAULT_CONFIG_NAME = "ERR173157"
|
|
|
93 |
citation=_CITATION,
|
94 |
)
|
95 |
|
96 |
+
@cached_property
|
97 |
+
def selex_info(self):
|
98 |
+
return self.SELEX_INFO.loc[self.config.name]
|
99 |
+
|
100 |
+
@cached_property
|
101 |
+
def design_length(self):
|
102 |
+
selex_info = self.selex_info
|
103 |
+
return int(pattern.search(selex_info["Ligand"]).group(0))
|
104 |
+
|
105 |
def _split_generators(self, dl_manager):
|
106 |
downloaded_files = None
|
107 |
+
sra_id = self.config.name
|
108 |
+
selex_info = self.selex_info
|
109 |
+
file = selex_info['file']
|
110 |
+
|
111 |
if getattr(self.config, "data_dir") is not None:
|
112 |
+
path = os.path.join(self.config.data_dir, sra_id, file)
|
113 |
+
downloaded_files = dl_manager.extract(path)
|
114 |
logger.info(f"Load from {downloaded_files}")
|
115 |
if downloaded_files is None or not os.path.exists(downloaded_files):
|
116 |
+
downloaded_url = "/".join([_URL, sra_id[:6], sra_id, file])
|
117 |
+
logger.info(f"Download from {downloaded_url}")
|
118 |
+
downloaded_files = dl_manager.download_and_extract(downloaded_url)
|
119 |
return [
|
120 |
datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloaded_files}),
|
121 |
]
|
|
|
140 |
|
141 |
def filter_fn(self, example):
|
142 |
seq = example["seq"]
|
143 |
+
if self.config.length_match and len(seq)!=self.design_length:
|
144 |
return False
|
145 |
if self.config.filter_N and "N" in seq:
|
146 |
return False
|
|
|
151 |
from datasets import load_dataset
|
152 |
dataset = load_dataset("jolma.py", name="ERR173157", split="all")
|
153 |
# dataset = load_dataset("jolma.py", name="ERR173157", split="all", data_dir="/root/autodl-fs/ERP001824-461")
|
|
|
|