A collection of reusable Python modules for AI services
Project description
CAEMA Utils
A collection of reusable Python modules for AI services.
Available Modules
| Module | Description | Status |
|---|---|---|
pose_estimation |
Body keypoint detection using YOLOv8 | Stable (v0.2.0) |
Installation
Step-by-Step Install (Recommended)
Run these commands one by one:
# Step 1: Install the package with all dependencies
pip install "caema-utils[all]"
# Step 2: Fix OpenCV for headless servers (Codespaces, Docker, cloud VMs)
pip uninstall opencv-python opencv-python-headless -y
pip install opencv-python-headless --force-reinstall --no-deps
# Step 3: Verify installation
pose-server --check
You should see: All dependencies OK!
Quick Install (One Command)
pip install "caema-utils[all]" && pip uninstall opencv-python opencv-python-headless -y && pip install opencv-python-headless --force-reinstall --no-deps && pose-server --check
Troubleshooting
If you see libGL.so.1: cannot open shared object file or No module named 'cv2':
pip uninstall opencv-python opencv-python-headless -y
pip install opencv-python-headless --force-reinstall --no-deps
From GitHub
pip install "caema-utils[all] @ git+https://github.com/msorozabal/Utils.git"
pip uninstall opencv-python opencv-python-headless -y
pip install opencv-python-headless --force-reinstall --no-deps
pose-server --check
Pose Estimation Module
Body keypoint detection using YOLOv8.
Features
- Detects 17 body keypoints (COCO format)
- Calculates angles between body segments
- Returns annotated images with skeleton overlay
- Multiple input formats: file path, bytes, numpy array
- Multiple output formats: bytes, base64, numpy array, file
Quick Start
from caema_utils.pose_estimation import PoseEstimator
# Initialize (downloads model on first run)
estimator = PoseEstimator(model_size="nano")
# Detect keypoints from file
result = estimator.analyze("photo.jpg")
if result['success']:
print(f"Confidence: {result['confidence']:.1%}")
print(f"Keypoints: {result['keypoints']}")
print(f"Angles: {result['angles']}")
# Save annotated image
with open("output.png", "wb") as f:
f.write(result['annotated_image_bytes'])
Model Sizes
| Size | Model | Speed | Accuracy | Use Case |
|---|---|---|---|---|
nano |
yolov8n-pose | Fastest | Good | Real-time, mobile |
small |
yolov8s-pose | Fast | Better | Balanced |
medium |
yolov8m-pose | Medium | High | Production |
large |
yolov8l-pose | Slow | Higher | Quality-critical |
xlarge |
yolov8x-pose | Slowest | Highest | Research |
API Reference
PoseEstimator
class PoseEstimator:
def __init__(self, model_size: str = "nano"):
"""
Initialize pose estimator.
Args:
model_size: "nano", "small", "medium", "large", or "xlarge"
"""
def analyze(
self,
image: Union[str, Path, bytes],
output_format: str = "bytes",
save_path: Optional[str] = None
) -> Dict[str, Any]:
"""
Detect body keypoints from image.
Args:
image: Path to image file or image bytes
output_format: "bytes", "base64", "array", or "file"
save_path: If output_format="file", save path for annotated image
Returns:
{
'success': bool,
'keypoints': Dict[str, List[float]], # name -> [x, y, confidence]
'angles': Dict[str, float], # angle_name -> degrees
'confidence': float, # 0-1
'annotated_image_*': ..., # Image in requested format
}
"""
Response Structure
{
'success': True,
'keypoints': {
'nose': [256.5, 128.3, 0.95],
'left_shoulder': [200.1, 180.2, 0.92],
'right_shoulder': [312.8, 178.9, 0.93],
# ... 17 keypoints total
},
'angles': {
'neck_angle': 5.2,
'shoulder_line_angle': 3.1,
'hip_line_angle': 2.8,
'torso_angle': 4.5,
'left_elbow_angle': 165.3,
'right_elbow_angle': 168.7,
'left_knee_angle': 175.2,
'right_knee_angle': 176.8,
},
'confidence': 0.93,
'annotated_image_bytes': b'...', # PNG image with skeleton overlay
}
HTTP API Server
The module includes a built-in FastAPI server for HTTP access.
Start the Server
# Check dependencies first
pose-server --check
# Start server (localhost only)
pose-server --port 8005
# Accept connections from other machines
pose-server --host 0.0.0.0 --port 8005
# With auto-reload for development
pose-server --host 0.0.0.0 --port 8005 --reload
API Endpoints
| Endpoint | Method | Description |
|---|---|---|
/ |
GET | API info |
/health |
GET | Health check |
/detect |
POST | Full detection (JSON + base64 image) |
/detect/image |
POST | Returns annotated image directly |
/detect/json-only |
POST | Detection without image |
/docs |
GET | OpenAPI documentation |
curl Examples
# Health check
curl http://localhost:8005/health
# Full detection with annotated image (base64)
curl -X POST "http://localhost:8005/detect" \
-H "accept: application/json" \
-F "file=@photo.jpg"
# Get annotated image directly (save to file)
curl -X POST "http://localhost:8005/detect/image" \
-F "file=@photo.jpg" \
--output annotated.png
# JSON only (faster, no image)
curl -X POST "http://localhost:8005/detect/json-only" \
-F "file=@photo.jpg"
Response Example
{
"success": true,
"filename": "photo.jpg",
"keypoints": {
"nose": [256.5, 128.3, 0.95],
"left_shoulder": [200.1, 180.2, 0.92]
},
"angles": {
"neck_angle": 5.2,
"shoulder_line_angle": 3.1
},
"confidence": 0.93,
"annotated_image_base64": "iVBORw0KGgoAAAANS..."
}
Keypoint Reference
The module detects 17 body keypoints in COCO format:
| Index | Name | Description |
|---|---|---|
| 0 | nose | Nose tip |
| 1 | left_eye | Left eye |
| 2 | right_eye | Right eye |
| 3 | left_ear | Left ear |
| 4 | right_ear | Right ear |
| 5 | left_shoulder | Left shoulder |
| 6 | right_shoulder | Right shoulder |
| 7 | left_elbow | Left elbow |
| 8 | right_elbow | Right elbow |
| 9 | left_wrist | Left wrist |
| 10 | right_wrist | Right wrist |
| 11 | left_hip | Left hip |
| 12 | right_hip | Right hip |
| 13 | left_knee | Left knee |
| 14 | right_knee | Right knee |
| 15 | left_ankle | Left ankle |
| 16 | right_ankle | Right ankle |
Calculated Angles
| Angle | Description |
|---|---|
neck_angle |
Angle of nose-to-mid-shoulder line relative to vertical |
shoulder_line_angle |
Angle of shoulder line relative to horizontal |
hip_line_angle |
Angle of hip line relative to horizontal |
torso_angle |
Angle of mid-shoulder-to-mid-hip line relative to vertical |
left_elbow_angle |
Angle at left elbow joint |
right_elbow_angle |
Angle at right elbow joint |
left_knee_angle |
Angle at left knee joint |
right_knee_angle |
Angle at right knee joint |
Development
Running Tests
# Install dev dependencies
pip install -e ".[all,dev]"
# Run tests
pytest tests/ -v
# Run with coverage
pytest tests/ --cov=caema_utils --cov-report=term-missing
Code Quality
# Format code
black src/ tests/
# Lint
ruff src/ tests/
# Type check
mypy src/
Requirements
- Python 3.9+
- For pose estimation:
- numpy < 2.0.0
- ultralytics >= 8.1.0
- opencv-python-headless >= 4.9.0
- For HTTP server:
- fastapi >= 0.100.0
- uvicorn >= 0.23.0
- python-multipart >= 0.0.6
License
MIT License - see LICENSE file.
Contributing
- Fork the repository
- Create a feature branch (
git checkout -b feature/amazing-feature) - Commit changes (
git commit -m 'Add amazing feature') - Push to branch (
git push origin feature/amazing-feature) - Open a Pull Request
Support
- Issues: GitHub Issues
- Documentation: GitHub Wiki
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file caema_utils-0.2.0.tar.gz.
File metadata
- Download URL: caema_utils-0.2.0.tar.gz
- Upload date:
- Size: 18.3 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.12.1
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
b4c4405bff081d499f17e6f126cb28e71415097864ff1cf15ba460487ee3d180
|
|
| MD5 |
bbb2b628338268b1d77721b8a2de6d45
|
|
| BLAKE2b-256 |
32edc0d7febb19eaca597e765019da5975e68d81b6c3dd3f41fc55d77bd2e092
|
File details
Details for the file caema_utils-0.2.0-py3-none-any.whl.
File metadata
- Download URL: caema_utils-0.2.0-py3-none-any.whl
- Upload date:
- Size: 15.4 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.12.1
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
a5ba7b735336c82c42c7d9703a94eec455e23f489142c2dc705f22817c582600
|
|
| MD5 |
22b9fa6b6186f060c26fe4508f103cf3
|
|
| BLAKE2b-256 |
eb0358f0d75176a4874f375fb9dad1f0372e3bf9154879a51f69f2f64eaa2b72
|