compox 3.0.1

No project description provided

Compox

Compox is a simple Python execution engine for Tescan applications. It's purpose is to run data processing algorithms written in Python on the server and report the results to the client application such as Tescan 3D Viewer or Picannto.

See the GitHub pages for full python API reference, docs and tutorials!

Installation

Compox can be installed through PyPI:

pip install compox

You can also install the server by cloning this repository and installing it from source.

Note that some additional dependencies, such as MinIO executable for your operating system, will be downloaded on the first run of the server.

Running the server

Once installed into your Python environment, Compox can be run using the compox command. This command is available in the virtual environment created during the setup process. Run the following command to list the available commands:

compox --help

It's recommended to start with generating an initial configuration file with compox generate-config. See the configuration options section below. To see the usage of the generate-config command, you can run:

compox generate-config --help

To run the Compox, use the following command:

compox run --config /path/to/config.yaml

Deploying algorithms

See the relevant section in the documentation for information about target structure of the deployable algorithms. You may also refer to our tutorials that will guide you through the preparation and deployment of selected algorithms.

Once your algorithm is ready, you can use the compox deploy-algorithms command to deploy it to your server. To see the usage of the deploy-algorithms command, you can run:

compox deploy-algorithms --help

By default, this command reads algorithm definitions from the folder specified in your configuration file with the deploy_algorithms_from key and deploys all deployable algorithm subfolders found there.

Examples:

compox deploy-algorithms --config /path/to/config.yaml

Deploy only selected algorithm folders:

compox deploy-algorithms --config /path/to/config.yaml --name foo --name bar

Override the algorithm root folder for a single run:

compox deploy-algorithms --config /path/to/config.yaml --path /path/to/algorithms

Note that the server does not need to be running in order to deploy the algorithms.

Configuration

The server uses pydantic settings for configuration. The options can be either set as runtime arguments, in a yaml file, or using the COMPOX__ environment variables, in corresponding order of precedence.

• Ensure all paths and URLs are correctly set before running the application.
• Adjust CUDA settings based on hardware capabilities.
• Logging paths should be accessible by the application to prevent errors.
• Use the logging section to control how much detail is shown in the console versus the log file.
• See the configuration reference below for detailed configuration options.

Configuration reference

Compox Configuration Reference

Section	Field	Default	Description
	port	5481	The main server port used to make requests.
	deploy_algorithms_from	"./algorithms"	Directory for algorithm deployment sources.
	log_path	"LOG_DEFAULT:compox.log"	Path to the main log file (supports dynamic prefixes).
	config	None	Optional config path override.
logging	console_level	"INFO"	Minimum level shown in the interactive console. At DEBUG/TRACE, suppressed access and MinIO console logs are shown again.
logging	file_level	"INFO"	Minimum level written to the main log file.
info	product_name	"Tescan Compox Backend"	Product display name.
info	server_tags	[] (auto appends "compox")	Tags attached to the server. "compox" is added automatically.
info	group_name	"TESCAN GROUP, a.s."	Name of the corporate group.
info	organization_name	"TESCAN GROUP, a.s."	Full name of the organization.
info	organization_domain	"tescan.com"	Domain used in server configuration.
gui	algorithm_add_remove_in_menus	False	Enables/disables GUI menu for algorithm management.
gui	use_systray	False	Enables/disables systray GUI integration.
gui	icon_path	Path to the installed package resource	Path to the systray icon (supports dynamic prefixes).
inference	device	"cuda"	Device used for model inference ("cpu", "cuda", "mps").
inference	cuda_visible_devices	"0"	Comma-separated list of visible CUDA GPUs.
inference.backend_settings (fastapi)	executor	"fastapi_background_tasks"	Task executor selection.
inference.backend_settings (fastapi)	worker_number	1	Number of FastAPI background task workers.
inference.backend_settings (celery)	executor	"celery"	Task executor selection.
inference.backend_settings (celery)	worker_name	"compox_worker"	Celery worker name.
inference.backend_settings (celery)	broker_url	required	Broker URL for Celery (e.g. amqp://).
inference.backend_settings (celery)	result_backend	"rpc://"	Celery result backend.
inference.backend_settings (celery)	run_flower	False	Whether to run Celery Flower.
inference.backend_settings (celery)	flower_port	None	Optional Flower port override.
storage	collection_prefix	""	Prefix applied to object store collections (useful for AWS S3).
storage	data_store_expire_days	1	Days until stored datasets expire.
storage	execution_store_expire_days	30	Days until execution data expires.
storage	training_store_expire_days	30	Days until training data expires.
storage	deploy_store_expire_days	30	Days until deploy data expires.
storage	stop_requests_expire_days	7	Days until stop-requests expire.
storage	access_key_id	generated with UUIDv4	Generated access key for storage backend. If null is provided, random UUIDv4 is generated.
storage	secret_access_key	generated with UUIDv4	Generated secret key for storage backend. If null is provided, random UUIDv4 is generated.
storage.backend_settings (minio)	provider	"minio"	Selected backend provider.
storage.backend_settings (minio)	start_instance	True	Whether to start a local MinIO server.
storage.backend_settings (minio)	port	5483	MinIO service port.
storage.backend_settings (minio)	console_port	5482	MinIO admin console port.
storage.backend_settings (minio)	executable_path	"minio/minio_bin" or "minio/minio.exe"	MinIO binary path (OS-specific, supports dynamic prefixes).
storage.backend_settings (minio)	storage_path	"minio/compox_store"	Storage directory used by MinIO (supports dynamic prefixes).
storage.backend_settings (minio)	aws_region	None	Optional AWS compatibility region.
storage.backend_settings (minio)	s3_domain_name	None	Optional domain override for S3 compatibility.
storage.backend_settings (minio)	s3_endpoint_url	Derived from port	Computed as http://localhost:{port}.
storage.backend_settings (aws)	provider	"aws"	AWS backend selection.
storage.backend_settings (aws)	s3_endpoint_url	None	Optional override for S3 endpoint URL.
storage.backend_settings (aws)	aws_region	None	AWS region (e.g. us-east-1).
storage.backend_settings (aws)	s3_domain_name	None	Domain used for S3-style URLs.
ssl	use_ssl	False	Enable SSL on the server.
ssl	ssl_keyfile	None	SSL key file path (supports dynamic prefixes).
ssl	ssl_certfile	None	SSL certificate path (supports dynamic prefixes).
middleware	allow_origins	[]	CORS allowed origins.
middleware	allow_methods	["GET"]	CORS allowed methods.
middleware	allow_headers	[]	CORS allowed headers.
middleware	allow_credentials	False	CORS allow credentials.
middleware	expose_headers	[]	CORS exposed headers.
middleware	max_age	3600	CORS preflight max age in seconds.

Logging configuration

Compox uses separate console and file sinks. This lets you keep the terminal readable while still collecting more detailed logs in the file.

Example:

log_path: "LOG_DEFAULT:compox.log"

logging:
  console_level: "INFO"
  file_level: "DEBUG"

Recommended modes:

• Quiet day-to-day operation:

  logging:
    console_level: "INFO"
    file_level: "DEBUG"
  

• Full interactive debugging:

  logging:
    console_level: "DEBUG"
    file_level: "DEBUG"
  

At INFO console level, Compox suppresses some high-frequency console noise such as polling status requests, file transfer access logs, and routine MinIO subprocess output. These logs still remain available in the file sink. At DEBUG or TRACE, those console logs are shown again.

Some fields in the Compox configuration (such as log_path, icon_path, etc.) support dynamic prefixes that resolve to OS-specific or runtime-specific paths. This allows for portability across platforms (e.g., Windows, Linux) and between development and production environments.

Supported Prefixes

Prefix	Meaning (Resolved To...)
LOG_DEFAULT:	A platform-dependent log directory:
	- Windows: %TEMP%/<organization>/<product>
	- Linux/macOS: /var/log/<organization>/<product>
PROGRAMDATA_DEFAULT:	A system-wide data directory (Windows only):
	- e.g., %PROGRAMDATA%/<organization>/<product>
	- On Linux/macOS, defaults to "." (current dir)

These are resolved at runtime in the Settings.parse_paths() validator method.

Security Notice

While you can technically communicate with a remotely running Compox instance, the communication is currently not authenticated and done via standard HTTP. Do not expose Compox endpoints to network in case of sensitive code or data. Security improvements are currently work in progress.