Blarify Integration - Quick Start¶

Get started with blarify code graph integration in 5 minutes.

Prerequisites¶

1. Neo4j Running:

# Check if Neo4j is running
docker ps | grep neo4j

1. Python Environment:

   # Already installed with amplihack
   pip install -e .
   

Option 1: Test Without Blarify (Recommended First)¶

Test the integration using sample data (no blarify installation needed):

python scripts/test_blarify_integration.py

Expected output:

✓ Connected to Neo4j
✓ PASS: Schema initialization
✓ PASS: Sample import
✓ PASS: Code-memory relationships
✓ PASS: Query functionality
✓ PASS: Incremental update

Results: 5/5 tests passed
🎉 All tests passed! Blarify integration is working.

Option 2: Install Blarify and Import Real Code¶

Install Blarify¶

# Install blarify
pip install blarify

# Optional: Install SCIP for 330x speed boost
npm install -g @sourcegraph/scip-python

Import Your Codebase¶

# Import entire src/ directory
python scripts/import_codebase_to_neo4j.py

# Or import specific directory
python scripts/import_codebase_to_neo4j.py --path ./src/amplihack

Expected output:

Step 1: Running blarify on src/
Step 2: Connecting to Neo4j
Step 3: Initializing code graph schema
Step 4: Importing blarify output to Neo4j
  - Files:         150
  - Classes:       45
  - Functions:     320
  - Imports:       280
  - Relationships: 450
Step 5: Linking code to memories
  Created 25 code-memory relationships
Step 6: Code graph statistics
  - Total files:     150
  - Total classes:   45
  - Total functions: 320
  - Total lines:     15000

Quick Examples¶

Example 1: Query Code Files¶

from amplihack.memory.neo4j import Neo4jConnector, BlarifyIntegration

with Neo4jConnector() as conn:
    integration = BlarifyIntegration(conn)
    stats = integration.get_code_stats()
    print(f"Files: {stats['file_count']}, Functions: {stats['function_count']}")

Example 2: Get Code Context for Memory¶

from amplihack.memory.neo4j import Neo4jConnector, BlarifyIntegration

with Neo4jConnector() as conn:
    integration = BlarifyIntegration(conn)

    # Get code context for a memory
    context = integration.query_code_context("memory-id-here")

    for func in context["functions"]:
        print(f"{func['name']} in {func['file_path']}")

Example 3: Link Memory to Code¶

from amplihack.memory.neo4j import Neo4jConnector, MemoryStore, BlarifyIntegration

with Neo4jConnector() as conn:
    memory_store = MemoryStore(conn)
    integration = BlarifyIntegration(conn)

    # Create memory with file reference
    memory_id = memory_store.create_memory(
        content="Always use circuit breaker for external calls",
        agent_type="architect",
        metadata={"file": "connector.py"},  # Link to file
    )

    # Link code to memories
    integration.link_code_to_memories()

    # Query context
    context = integration.query_code_context(memory_id)
    print(f"Linked to {len(context['files'])} files")

Neo4j Browser Queries¶

Open Neo4j Browser at http://localhost:7474 and try these queries:

View All Code Files¶

MATCH (cf:CodeFile)
RETURN cf.path, cf.language, cf.lines_of_code
ORDER BY cf.lines_of_code DESC
LIMIT 10

View Function Call Graph¶

MATCH (source:Function)-[:CALLS]->(target:Function)
RETURN source.name as caller, target.name as callee
LIMIT 20

View Code-Memory Relationships¶

MATCH (m:Memory)-[:RELATES_TO_FILE]->(cf:CodeFile)
RETURN m.content, cf.path
LIMIT 10

Find Complex Functions Without Memories¶

MATCH (f:Function)
WHERE f.complexity > 10
OPTIONAL MATCH (f)<-[:RELATES_TO_FUNCTION]-(m:Memory)
RETURN f.name, f.complexity,
       CASE WHEN m IS NULL THEN 'No documentation' ELSE m.content END
ORDER BY f.complexity DESC
LIMIT 10

Common Issues¶

Issue: "Cannot connect to Neo4j"¶

Solution: Start Neo4j container

# Use amplihack memory system tools
python -c "from amplihack.memory.neo4j import ensure_neo4j_running; ensure_neo4j_running()"

Issue: "blarify not found"¶

Solution: Either install blarify or use test with sample data

# Option 1: Install blarify
pip install blarify

# Option 2: Use sample data (no blarify needed)
python scripts/test_blarify_integration.py

Issue: "Import failed - invalid JSON"¶

Solution: Check blarify output format

# View blarify output
cat .amplihack/blarify_output.json | python -m json.tool

Next Steps¶

1. Read Full Documentation: docs/blarify_integration.md
2. Explore Neo4j Browser: http://localhost:7474
3. Run Advanced Queries: See use cases in documentation
4. Integrate with Agents: Use code context in agent decision-making

Performance Tips¶

1. Use SCIP for Speed: npm install -g @sourcegraph/scip-python (330x faster)
2. Incremental Updates: --incremental flag for changed files only
3. Filter Languages: --languages python to reduce parsing time

Support¶

• Test suite: python scripts/test_blarify_integration.py
• Full docs: docs/blarify_integration.md
• Memory system: docs/neo4j_memory_system.md

Ready to go! Start with the test suite, then import your real codebase.