Skip to content

QUICK_START.md

Latest commit

 

History

History
218 lines (166 loc) · 4.98 KB

QUICK_START.md

File metadata and controls

218 lines (166 loc) · 4.98 KB

DDI Skill - Quick Start Guide

Status: ✅ WORKING - Pipeline fixed and tested Last Updated: 2026-02-09

Choose Your Implementation

Python SDK

Option 1: Use the Working Pipeline (RECOMMENDED)

# Import from either file (both work)
from python_implementation import DDIAnalyzer
# or: from ddi_pipeline import DDIAnalyzer

# Initialize analyzer
analyzer = DDIAnalyzer()

# Analyze drug pair
report = analyzer.analyze("warfarin", "amoxicillin")

# Report automatically saved to: DDI_report_warfarin_amoxicillin.md
print(f"Risk Score: {report['risk_score']}/100")
print(f"Severity: {report['severity']}")

Option 2: Use Individual Tools

from tooluniverse import ToolUniverse

tu = ToolUniverse()
tu.load_tools()

# Drug identification (RxNorm)
result = tu.tools.RxNorm_get_drug_names(drug_name="warfarin")  # ✅ Correct parameter

# Drug interactions (DrugBank)
result = tu.tools.drugbank_get_drug_interactions_by_drug_name_or_id(
    query="warfarin",      # ✅ Correct parameter
    case_sensitive=False,
    exact_match=False,
    limit=50
)

# Pharmacology (DrugBank)
result = tu.tools.drugbank_get_pharmacology_by_drug_name_or_drugbank_id(
    query="warfarin",      # ✅ Correct parameter
    case_sensitive=False,
    exact_match=False,
    limit=1
)

# FDA Labels (DailyMed - 2 steps)
search = tu.tools.DailyMed_search_spls(query="warfarin")
setid = search['data']['spls'][0]['setid']
interactions = tu.tools.DailyMed_parse_drug_interactions(setid=setid)

# Literature (PubMed)
result = tu.tools.PubMed_search_articles(
    query='"warfarin" AND "drug interaction"',
    max_results=10
)

# Adverse events (FAERS)
result = tu.tools.FAERS_count_reactions_by_drug_event(
    medicinalproduct="warfarin",  # ✅ Correct parameter (not 'drug_name')
    event_name="drug interaction"
)

MCP (Model Context Protocol)

Option 1: Conversational (Claude Desktop or Compatible Client)

Tell Claude:

"Analyze drug-drug interactions between warfarin and amoxicillin using ToolUniverse"

Claude will follow the workflow from SKILL.md and use these tools:

  1. RxNorm_get_drug_names - Drug identification
  2. drugbank_get_drug_interactions - Interaction data
  3. DailyMed_parse_drug_interactions - FDA labels
  4. PubMed_search_articles - Literature evidence
  5. FAERS_count_reactions - Adverse events

Option 2: Direct Tool Calls

Step 1: Drug Identification

Tool: RxNorm_get_drug_names
Parameters:
{
  "drug_name": "warfarin"
}

Step 2: Interaction Analysis

Tool: drugbank_get_drug_interactions_by_drug_name_or_id
Parameters:
{
  "query": "warfarin",
  "case_sensitive": false,
  "exact_match": false,
  "limit": 50
}

Step 3: FDA Label Warnings

Tool: DailyMed_search_spls
Parameters:
{
  "query": "warfarin"
}

Then use the returned setid:

Tool: DailyMed_parse_drug_interactions
Parameters:
{
  "setid": "[setid from previous call]"
}

Step 4: Literature Evidence

Tool: PubMed_search_articles
Parameters:
{
  "query": "\"warfarin\" AND \"drug interaction\"",
  "max_results": 10
}

Step 5: Adverse Events

Tool: FAERS_count_reactions_by_drug_event
Parameters:
{
  "medicinalproduct": "warfarin",
  "event_name": "drug interaction"
}

Run Examples (Python SDK)

# Run the working pipeline
python ddi_pipeline.py

# Generates reports:
#   - DDI_report_warfarin_amoxicillin.md
#   - DDI_report_simvastatin_ketoconazole.md

What Works ✅

  • ✅ Drug identification (RxNorm)
  • ✅ DrugBank integration (correct parameters)
  • ✅ DailyMed FDA labels
  • ✅ PubMed literature search
  • ✅ FAERS adverse events
  • ✅ Report generation (markdown)
  • ✅ Risk scoring algorithm
  • ✅ Clinical recommendations

Known Limitations

⚠️ Data Availability: Some tools return empty results:

  • DrugBank interactions may not find all drug pairs
  • FAERS data may be limited
  • This is a data availability issue, not a code issue

⚠️ Response Formats: Some tools return lists instead of dicts:

  • Pipeline handles this gracefully
  • Reports still generate successfully

Tool Parameters (All Implementations)

These parameter names apply to both Python SDK and MCP:

Tool Parameter Correct Name Notes
RxNorm_get_drug_names Drug name drug_name NOT query
drugbank_* Drug query query All DrugBank tools use this
FAERS_count_reactions_by_drug_event Drug name medicinalproduct NOT drug_name
DailyMed_search_spls Search query query Step 1 of 2
DailyMed_parse_drug_interactions SetID setid Step 2 of 2

Note: Whether using Python SDK or MCP, the parameter names are the same

Files

  • ddi_pipeline.py - Complete working pipeline ✅
  • SKILL.md - Skill documentation (framework)
  • EXAMPLES.md - Clinical scenarios (documentation)
  • README.md - Original readme
  • QUICK_START.md - This file

Fixed: 2026-02-09 - Pipeline now uses correct ToolUniverse tool parameters