|
1
|
|
|
#!/usr/bin/env python |
|
2
|
|
|
# -*- coding: utf-8 -*- |
|
3
|
|
|
""" |
|
4
|
|
|
|
|
5
|
|
|
""" |
|
6
|
|
|
from __future__ import print_function |
|
7
|
|
|
import argparse |
|
8
|
|
|
from icecream import ic |
|
9
|
|
|
import sys |
|
10
|
|
|
ic.configureOutput(outputFunction=lambda *a: print(*a, file=sys.stderr)) |
|
11
|
|
|
ic.configureOutput(prefix='> ') |
|
12
|
|
|
|
|
13
|
|
|
|
|
14
|
|
|
def get_parser(): |
|
15
|
|
|
parser = argparse.ArgumentParser( |
|
16
|
|
|
description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter) |
|
17
|
|
|
#parser.add_argument('-', "--", help="", default="") |
|
18
|
|
|
parser.add_argument("-v", "--verbose", |
|
19
|
|
|
action="store_true", help="be verbose") |
|
20
|
|
|
parser.add_argument("file", help="", default="", nargs='+') |
|
21
|
|
|
return parser |
|
22
|
|
|
|
|
23
|
|
|
import requests |
|
24
|
|
|
|
|
25
|
|
|
def fetch_pdb_header(pdb_id): |
|
26
|
|
|
""" |
|
27
|
|
|
Fetches the header information for a given PDB ID using the RCSB PDB GraphQL API. |
|
28
|
|
|
|
|
29
|
|
|
:param pdb_id: The PDB ID to query. |
|
30
|
|
|
:return: A dictionary containing the header information or an error message. |
|
31
|
|
|
""" |
|
32
|
|
|
# Define the GraphQL query |
|
33
|
|
|
query = """ |
|
34
|
|
|
query getPDBHeader($id: String!) { |
|
35
|
|
|
entry(entry_id: $id) { |
|
36
|
|
|
rcsb_id |
|
37
|
|
|
struct { |
|
38
|
|
|
title |
|
39
|
|
|
} |
|
40
|
|
|
exptl { |
|
41
|
|
|
method |
|
42
|
|
|
} |
|
43
|
|
|
rcsb_accession_info { |
|
44
|
|
|
deposit_date |
|
45
|
|
|
initial_release_date |
|
46
|
|
|
} |
|
47
|
|
|
citation { |
|
48
|
|
|
title |
|
49
|
|
|
rcsb_authors |
|
50
|
|
|
year |
|
51
|
|
|
pdbx_database_id_DOI |
|
52
|
|
|
} |
|
53
|
|
|
} |
|
54
|
|
|
} |
|
55
|
|
|
""" |
|
56
|
|
|
|
|
57
|
|
|
# Define the variables |
|
58
|
|
|
variables = {"id": pdb_id} |
|
59
|
|
|
|
|
60
|
|
|
# API endpoint |
|
61
|
|
|
url = "https://data.rcsb.org/graphql" |
|
62
|
|
|
headers = {"Content-Type": "application/json"} |
|
63
|
|
|
|
|
64
|
|
|
try: |
|
65
|
|
|
# Send the POST request |
|
66
|
|
|
response = requests.post(url, json={"query": query, "variables": variables}, headers=headers) |
|
67
|
|
|
response.raise_for_status() # Raise HTTPError for bad responses (4xx and 5xx) |
|
68
|
|
|
data = response.json() |
|
69
|
|
|
|
|
70
|
|
|
# Return the data or error |
|
71
|
|
|
if "data" in data: |
|
72
|
|
|
return data["data"] |
|
73
|
|
|
else: |
|
74
|
|
|
return {"error": "No data returned. Check the PDB ID or API response."} |
|
75
|
|
|
|
|
76
|
|
|
except requests.exceptions.RequestException as e: |
|
77
|
|
|
return {"error": str(e)} |
|
78
|
|
|
|
|
79
|
|
|
|
|
80
|
|
|
if __name__ == '__main__': |
|
81
|
|
|
parser = get_parser() |
|
82
|
|
|
args = parser.parse_args() |
|
83
|
|
|
|
|
84
|
|
|
if list != type(args.file): |
|
85
|
|
|
args.file = args.file |
|
86
|
|
|
|
|
87
|
|
|
for pdb_id in [f.replace('.pdb', '').replace('.cif', '') for f in args.file]: |
|
88
|
|
|
header_info = fetch_pdb_header(pdb_id) |
|
89
|
|
|
title = header_info.get('entry', {}).get('struct', {}).get('title', 'Title not found') |
|
90
|
|
|
# Print the extracted title |
|
91
|
|
|
print(pdb_id, title) |
|
92
|
|
|
|
mmagnus /
rna-tools
