Update fetch_data.py
Browse files- fetch_data.py +49 -44
fetch_data.py
CHANGED
@@ -1,45 +1,50 @@
|
|
1 |
-
from arvix import extract_arxiv_data
|
2 |
-
from pmc import extract_pmc_data
|
3 |
-
import json
|
4 |
-
import dotenv
|
5 |
-
import os
|
6 |
-
from concurrent.futures import ThreadPoolExecutor
|
7 |
-
|
8 |
-
dotenv.load_dotenv()
|
9 |
-
ACCESS_KEY = os.getenv("ACCESS_KEY")
|
10 |
-
|
11 |
-
def fetch_arxiv_data():
|
12 |
-
return json.loads(extract_arxiv_data())
|
13 |
-
|
14 |
-
def fetch_pmc_data():
|
15 |
-
return json.loads(extract_pmc_data())
|
16 |
-
|
17 |
-
def fetch_data(user_access_key):
|
18 |
-
if user_access_key != ACCESS_KEY:
|
19 |
-
papers_data = {"
|
20 |
-
else:
|
21 |
-
papers_data = {}
|
22 |
-
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
|
|
|
|
|
|
|
|
|
|
45 |
f.write(data)
|
|
|
1 |
+
from arvix import extract_arxiv_data
|
2 |
+
from pmc import extract_pmc_data
|
3 |
+
import json
|
4 |
+
import dotenv
|
5 |
+
import os
|
6 |
+
from concurrent.futures import ThreadPoolExecutor
|
7 |
+
|
8 |
+
dotenv.load_dotenv()
|
9 |
+
ACCESS_KEY = os.getenv("ACCESS_KEY")
|
10 |
+
|
11 |
+
def fetch_arxiv_data():
|
12 |
+
return json.loads(extract_arxiv_data())
|
13 |
+
|
14 |
+
def fetch_pmc_data():
|
15 |
+
return json.loads(extract_pmc_data())
|
16 |
+
|
17 |
+
def fetch_data(user_access_key):
|
18 |
+
if user_access_key != ACCESS_KEY:
|
19 |
+
papers_data = {"status": "Invalid access key"}
|
20 |
+
else:
|
21 |
+
papers_data = {}
|
22 |
+
try:
|
23 |
+
papers_data['status'] = 'success'
|
24 |
+
with ThreadPoolExecutor() as executor:
|
25 |
+
arxiv_future = executor.submit(fetch_arxiv_data)
|
26 |
+
pmc_future = executor.submit(fetch_pmc_data)
|
27 |
+
arxiv_data = arxiv_future.result()
|
28 |
+
pmc_data = pmc_future.result()
|
29 |
+
for topic, topic_data in arxiv_data.items():
|
30 |
+
if topic_data['count'] == 0:
|
31 |
+
continue
|
32 |
+
else:
|
33 |
+
papers_data[topic] = {}
|
34 |
+
papers_data[topic]['ids'] = topic_data['ids']
|
35 |
+
for topic, topic_data in pmc_data.items():
|
36 |
+
if topic_data['count'] == 0:
|
37 |
+
continue
|
38 |
+
else:
|
39 |
+
papers_data[topic] = {}
|
40 |
+
papers_data[topic]['ids'] = topic_data['ids']
|
41 |
+
except Exception as e:
|
42 |
+
print(str(e))
|
43 |
+
papers_data['status'] = 'error'
|
44 |
+
data = json.dumps(papers_data, indent=4, ensure_ascii=False)
|
45 |
+
return data
|
46 |
+
|
47 |
+
if __name__ == '__main__':
|
48 |
+
data = fetch_data()
|
49 |
+
with open('data.json', 'w') as f:
|
50 |
f.write(data)
|