Spaces:
Sleeping
Sleeping
Update genesis/api_clients/umls_api.py
Browse files- genesis/api_clients/umls_api.py +55 -62
genesis/api_clients/umls_api.py
CHANGED
@@ -1,93 +1,86 @@
|
|
1 |
# genesis/api_clients/umls_api.py
|
2 |
import os
|
3 |
import requests
|
|
|
4 |
|
5 |
-
|
6 |
-
|
7 |
-
|
8 |
|
|
|
|
|
9 |
|
10 |
-
|
|
|
11 |
"""
|
12 |
-
|
13 |
"""
|
14 |
-
|
15 |
-
r = requests.post(UMLS_AUTH_ENDPOINT, data=params)
|
16 |
r.raise_for_status()
|
|
|
|
|
|
|
17 |
|
18 |
-
# Extract TGT from HTML response
|
19 |
-
tgt_url = r.text.split('action="')[1].split('"')[0]
|
20 |
-
return tgt_url
|
21 |
|
22 |
-
|
23 |
-
def get_service_ticket(tgt):
|
24 |
"""
|
25 |
-
Get
|
26 |
"""
|
27 |
-
|
28 |
-
r = requests.post(tgt, data=params)
|
29 |
r.raise_for_status()
|
30 |
return r.text
|
31 |
|
32 |
|
33 |
-
def search_umls(
|
34 |
"""
|
35 |
-
Search UMLS for a term and return
|
36 |
"""
|
37 |
tgt = get_tgt()
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
r = requests.get(url, params=params)
|
43 |
r.raise_for_status()
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
"
|
51 |
-
|
52 |
-
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
-
def get_concept_details(cui: str):
|
59 |
"""
|
60 |
-
|
|
|
61 |
"""
|
62 |
tgt = get_tgt()
|
63 |
-
|
64 |
-
|
65 |
-
|
66 |
-
params = {"ticket": st}
|
67 |
-
r = requests.get(url, params=params)
|
68 |
r.raise_for_status()
|
69 |
-
return r.json()
|
70 |
|
71 |
|
72 |
-
def
|
73 |
"""
|
74 |
-
|
75 |
"""
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
-
details = get_concept_details(cui)
|
83 |
-
entity = details.get("result", {})
|
84 |
|
85 |
-
mapped.append({
|
86 |
-
"cui": cui,
|
87 |
-
"name": entity.get("name", ""),
|
88 |
-
"synonyms": [syn.get("name", "") for syn in entity.get("atoms", [])],
|
89 |
-
"semantic_types": [st.get("name", "") for st in entity.get("semanticTypes", [])],
|
90 |
-
"rootSource": c.get("rootSource", "")
|
91 |
-
})
|
92 |
|
93 |
-
|
|
|
|
|
|
|
|
|
|
|
|
1 |
# genesis/api_clients/umls_api.py
|
2 |
import os
|
3 |
import requests
|
4 |
+
from typing import Dict, List, Optional
|
5 |
|
6 |
+
# UMLS API
|
7 |
+
UMLS_BASE = "https://uts-ws.nlm.nih.gov/rest"
|
8 |
+
UMLS_API_KEY = os.getenv("UMLS_API_KEY") # Add this to your Hugging Face secrets or .env
|
9 |
|
10 |
+
if not UMLS_API_KEY:
|
11 |
+
raise EnvironmentError("UMLS_API_KEY is missing. Add it to your environment variables or Hugging Face secrets.")
|
12 |
|
13 |
+
# UMLS requires ticket-granting service for authentication
|
14 |
+
def get_tgt() -> str:
|
15 |
"""
|
16 |
+
Obtain a Ticket Granting Ticket (TGT) for UMLS authentication.
|
17 |
"""
|
18 |
+
r = requests.post("https://utslogin.nlm.nih.gov/cas/v1/api-key", data={"apikey": UMLS_API_KEY})
|
|
|
19 |
r.raise_for_status()
|
20 |
+
from xml.etree import ElementTree
|
21 |
+
root = ElementTree.fromstring(r.text)
|
22 |
+
return root.attrib.get("action")
|
23 |
|
|
|
|
|
|
|
24 |
|
25 |
+
def get_service_ticket(tgt: str) -> str:
|
|
|
26 |
"""
|
27 |
+
Get a single-use service ticket for API requests.
|
28 |
"""
|
29 |
+
r = requests.post(tgt, data={"service": "http://umlsks.nlm.nih.gov"})
|
|
|
30 |
r.raise_for_status()
|
31 |
return r.text
|
32 |
|
33 |
|
34 |
+
def search_umls(query: str, max_results: int = 10) -> List[Dict]:
|
35 |
"""
|
36 |
+
Search UMLS Metathesaurus for a term and return CUIs and names.
|
37 |
"""
|
38 |
tgt = get_tgt()
|
39 |
+
ticket = get_service_ticket(tgt)
|
40 |
+
params = {"string": query, "ticket": ticket, "pageSize": max_results}
|
41 |
+
|
42 |
+
r = requests.get(f"{UMLS_BASE}/search/current", params=params)
|
|
|
43 |
r.raise_for_status()
|
44 |
+
results = r.json().get("result", {}).get("results", [])
|
45 |
+
|
46 |
+
return [
|
47 |
+
{
|
48 |
+
"ui": item.get("ui"),
|
49 |
+
"name": item.get("name"),
|
50 |
+
"rootSource": item.get("rootSource")
|
51 |
+
}
|
52 |
+
for item in results if item.get("ui") != "NONE"
|
53 |
+
]
|
54 |
+
|
55 |
+
|
56 |
+
def get_concept_details(cui: str) -> Dict:
|
|
|
|
|
57 |
"""
|
58 |
+
Get full details for a given CUI (Concept Unique Identifier).
|
59 |
+
Includes synonyms, definitions, and mappings.
|
60 |
"""
|
61 |
tgt = get_tgt()
|
62 |
+
ticket = get_service_ticket(tgt)
|
63 |
+
|
64 |
+
r = requests.get(f"{UMLS_BASE}/content/current/CUI/{cui}", params={"ticket": ticket})
|
|
|
|
|
65 |
r.raise_for_status()
|
66 |
+
return r.json().get("result", {})
|
67 |
|
68 |
|
69 |
+
def map_between_ontologies(cui: str) -> List[Dict]:
|
70 |
"""
|
71 |
+
Given a CUI, return cross-ontology mappings (e.g., SNOMED β MeSH β ICD).
|
72 |
"""
|
73 |
+
tgt = get_tgt()
|
74 |
+
ticket = get_service_ticket(tgt)
|
75 |
+
|
76 |
+
r = requests.get(f"{UMLS_BASE}/crosswalk/current/source/{cui}", params={"ticket": ticket})
|
77 |
+
r.raise_for_status()
|
78 |
+
return r.json().get("result", [])
|
|
|
|
|
79 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
80 |
|
81 |
+
def normalize_term(term: str) -> Optional[Dict]:
|
82 |
+
"""
|
83 |
+
Normalize a free-text biomedical term to its canonical CUI and return metadata.
|
84 |
+
"""
|
85 |
+
matches = search_umls(term, max_results=1)
|
86 |
+
return matches[0] if matches else None
|