A “Provider” is a python class that understands how to access a certain kind of LCA data source. Providers implement the BasicArchive (antelope_core.archives.basic_archive.BasicArchive) interface, which (though poorly specified) provides a set of useful functions for storing and retrieving entities.

The EntityStore

The EntityStore is essentially a big hash table that allows you to retrieve an “entity” from a reference string. An entity store has a handful of properties, of which the most useful are:

• ref - the semantic reference for the data collection (this becomes the origin of a data set)
• source - a file, directory, or URL that contains the data collection
• static - a boolean property that indicates (if true) that the entire data collection must be loaded at once.

Its core utilities are:

• retrieve_or_fetch_entity(entity_id), which loads an entity by its id, and stores it in local memory
• _add(entity, ref), which stores entities in a way that handles validation and retrieval
• __getitem__, which retrieves an already-loaded entity by its reference or UUID
• load_all(), which loads all content
• serialize(), which writes the entity store to a JSON file.

It is a partially-abstract class, and the following routines must be implemented by providers:

• _load_all() which is called by load_all()
• _fetch() which is called by retrieve_or_fetch_entity() when the entity is not known locally

In addition, each provider is responsible for constructing valid entities.

The provider infrastructure is some of the oldest python code in the archive. please don’t judge.

The BasicArchive implements the key features of the EntityStore for flows and quantities, as well as introduces a search() function. A BasicArchive can be saved and restored from JSON and forms the base class for all other providers. The LcArchive adds processes to the list of supported entities, and is used for all Life Cycle data sources that contain processes.

Default Providers

The providers available built-in in antelope_core are:

• BasicArchive (generic container for quantities and flows)
• LcArchive (adds processes)
• Background (the “trivial” background engine, for accessing files containing rolled-up datasets)
• OpenLcaJsonLdArchive (for OpenLCA .zip files)
• EcoinventLcia (for ecoinvent-issued LCIA tables)
• Traci21Factors (for accessing TRACI 2.1 spreadsheet)
• XdbClient (for connecting to xdb background data servers)

Adding lxml support ($ pip install lxml) allows XML-based providers to be loaded:

• EcospoldV1Archive for ecospold v1, including ecoinvent 2.2 and old-style US LCI
• EcospoldV2Archive for ecospold v2, including ecoinvent 3.x databases
• IlcdArchive for ILCD datasets (note: this has not been maintained for some time)
• IlcdLcia is a subclass of IlcdArchive and adds the capability to read stored LCIA results

Adding antelope_background introduces Tarjan ordering:

• TarjanBackground performs partial ordering of databases and constructs allocated LCI matrices

Adding antelope_foreground introduces foreground modeling capacity:

• LcForeground provides the capability to create and save fragments, and also stores catalog references to other data sources
• OryxClient (for connecting to oryx foreground data servers)