Skip to content

VECTOR_SEARCH_QUICK_REF.md

Latest commit

 

History

History
99 lines (74 loc) · 2.27 KB

VECTOR_SEARCH_QUICK_REF.md

File metadata and controls

99 lines (74 loc) · 2.27 KB

Vector Search Quick Reference

API Endpoints

Search by Vibe

GET /api/v1/search/vibe?query=dark+moody+loop

Parameters:

  • query (required): Natural language search
  • limit (20): Max results (1-100)
  • min_similarity (0.7): Threshold (0-1)
  • bpm_min, bpm_max: BPM range
  • genre: Genre filter
  • energy_min, energy_max: Energy (0-1)
  • danceability_min, danceability_max: Danceability (0-1)

Similar Samples

GET /api/v1/search/similar/123

Parameters:

  • limit (10): Max results (1-50)
  • min_similarity (0.8): Threshold (0-1)

Example Queries

# Basic search
curl "http://localhost:8100/api/v1/search/vibe?query=energetic+trap+drums"

# With filters
curl "http://localhost:8100/api/v1/search/vibe?query=chill+jazz&bpm_min=80&bpm_max=100&energy_max=0.6"

# Similar samples
curl "http://localhost:8100/api/v1/search/similar/456?limit=5"

# Pretty JSON
curl "http://localhost:8100/api/v1/search/vibe?query=dark+bass" | jq .

Cost

  • Per query: ~$0.000001 (0.0001¢)
  • 10K queries: ~$0.01 (1¢)
  • 1M queries: ~$1.00

Python Usage

from app.services.embedding_service import EmbeddingService
from app.services.vibe_search_service import VibeSearchService

# Generate embedding
embedding = await embedding_service.generate_embedding("dark moody loop")

# Search samples
results = await vibe_search_service.search_by_vibe(
    query="energetic drums",
    limit=10,
    bpm_min=120,
    bpm_max=140
)

# Similar samples
similar = await vibe_search_service.get_similar_samples(
    sample_id=123,
    limit=5
)

Files

  • Service: backend/app/services/embedding_service.py
  • Search: backend/app/services/vibe_search_service.py
  • Endpoint: backend/app/api/v1/endpoints/vibe_search.py
  • Config: backend/app/core/config.py (vector search settings)
  • Test: backend/scripts/test_vibe_search.py

Testing

# Test embeddings
./venv/bin/python backend/scripts/test_vibe_search.py

# Start server
./venv/bin/python backend/run.py

# Test endpoint
curl "http://localhost:8100/api/v1/search/vibe?query=test&limit=5"

Next Steps (Data Processing)

  1. Generate embeddings for 2,328 existing samples
  2. Insert into Turso sample_embeddings table
  3. Verify vector search returns results