SKILL: Bohrium LKM (Large Knowledge Model)

Overview

LKM endpoints on open.bohrium.com provide scientific knowledge graph search, claim verification with evidence chains, variable relationship queries, and batch paper OCR.

Core capabilities:

Use when:

• Verifying whether a scientific conclusion has literature support
• Querying relationships between two variables (positive/negative/none)
• Searching knowledge nodes in a specific domain
• Batch OCR of papers for structured data extraction

Don't use for:

• General paper keyword search → bohrium-paper-search
• Knowledge base file management → bohrium-knowledge-base
• Single PDF parsing → bohrium-pdf-parser

No CLI support — HTTP API only.

Auth configuration

"bohrium-lkm": {
  "enabled": true,
  "apiKey": "YOUR_ACCESS_KEY",
  "env": {
    "ACCESS_KEY": "YOUR_ACCESS_KEY"
  }
}

Common template

import os, requests

AK = os.environ["ACCESS_KEY"]
BASE = "https://open.bohrium.com/openapi/v1/lkm"
H = {"accessKey": AK, "Content-Type": "application/json"}

1. Knowledge graph search — /lkm/search

r = requests.post(f"{BASE}/search", headers=H, json={
    "query": "effect of temperature on lithium ion battery degradation",
    "limit": 10
})
data = r.json()
print(data)

Parameters:

2. Claim matching — /lkm/claims/match

Submit a scientific claim, get back evidence that supports or refutes it (with source papers and relevance scores).

r = requests.post(f"{BASE}/claims/match", headers=H, json={
    "text": "Graphene oxide improves the mechanical strength of concrete",
    "limit": 5
})
data = r.json()
# data["data"]["variables"] contains matched claims
# data["data"]["papers"] contains related paper details
# data["data"]["new_claim_likely"] indicates if this might be a novel claim
for item in data.get("data", {}).get("variables", []):
    print(f"  Claim ID: {item['id']}")
    print(f"  Role: {item.get('role')}")  # premise / conclusion
    print(f"  Score: {item.get('score')}")
    print(f"  Content: {item.get('content')[:100]}...")

Parameters:

Response fields:

3. Evidence chain — /lkm/claims/:id/evidence

Get detailed evidence for a specific claim ID (source papers, experimental data, reasoning paths).

claim_id = "abc123"
r = requests.get(f"{BASE}/claims/{claim_id}/evidence", headers=H)
data = r.json()
for ev in data.get("data", []):
    print(f"  Paper: {ev.get('paper_title')}")
    print(f"  Evidence: {ev.get('text')}")
    print(f"  Type: {ev.get('evidence_type')}")

4. Variable batch query — /lkm/variables/batch

Batch query variable details by ID. Variable IDs can be obtained from /lkm/search or /lkm/claims/match responses.

r = requests.post(f"{BASE}/variables/batch", headers=H, json={
    "ids": ["gcn_b2bf079b541a4fa0", "gcn_5cecd02c3d8a4e61"]
})
data = r.json()
for var in data.get("data", {}).get("variables", []):
    print(f"  ID: {var['id']}")
    print(f"  Content: {var.get('content')[:100]}...")
# data["data"]["not_found"] lists IDs that were not found

Parameters:

5. Batch paper OCR — /lkm/papers/ocr/batch

Batch OCR extraction from papers.

r = requests.post(f"{BASE}/papers/ocr/batch", headers=H, json={
    "paper_ids": ["doi:10.1038/s41586-021-03819-2", "doi:10.1126/science.abf3041"]
})
data = r.json()
for paper in data.get("data", []):
    print(f"  Paper: {paper.get('title')}")
    print(f"  Status: {paper.get('status')}")

Parameters:

curl examples

AK="YOUR_ACCESS_KEY"

# Knowledge graph search
curl -s -X POST "https://open.bohrium.com/openapi/v1/lkm/search" \
  -H "accessKey: $AK" -H "Content-Type: application/json" \
  -d '{"query":"lithium battery degradation mechanism","limit":10}' | jq .

# Claim matching
curl -s -X POST "https://open.bohrium.com/openapi/v1/lkm/claims/match" \
  -H "accessKey: $AK" -H "Content-Type: application/json" \
  -d '{"text":"MoS2 is a promising catalyst for hydrogen evolution","limit":5}' | jq .

# Evidence chain
curl -s -X GET "https://open.bohrium.com/openapi/v1/lkm/claims/abc123/evidence" \
  -H "accessKey: $AK" | jq .

# Variable batch query (IDs from search/claims results)
curl -s -X POST "https://open.bohrium.com/openapi/v1/lkm/variables/batch" \
  -H "accessKey: $AK" -H "Content-Type: application/json" \
  -d '{"ids":["gcn_b2bf079b541a4fa0","gcn_5cecd02c3d8a4e61"]}' | jq .

# Batch OCR
curl -s -X POST "https://open.bohrium.com/openapi/v1/lkm/papers/ocr/batch" \
  -H "accessKey: $AK" -H "Content-Type: application/json" \
  -d '{"paper_ids":["doi:10.1038/s41586-021-03819-2"]}' | jq .

Troubleshooting

Pairs well with

• lkm verify claim → paper-search to find original full paper
• lkm query variable relationships → mol-search for related molecular structures
• lkm batch OCR → knowledge-base to store extracted results