Python based tools for working with LCA foreground model disclosures

Disclosure framework based on "Disclosure of product system models in life cycle assessment: achieving Transparency and Privacy" by Brandon Kuczenski in J Ind Ecol. (doi: 10.1111/jiec12810)

This package was originally authored by James Joyce. Code and testing related to Brightway 2 have been branched into the lca_disclosures_bw2 repository so as to remove dependencies from the core program. The original repository lca_disclosures is still operational.

The lca_disclosures repository supports the generation of a disclosure specification from the following software:

• lcopt
• Antelope Framework

Support (and testing) for Brightway2 is available in the original repository lca_disclosures.

To use the disclosure framework, instantiate the appropriate Disclosure subclass and supply it with the software-specific input data. The disclosure object can be used to answer questions about the model:

• listing foreground flows, background flows, and emissions;
• identifying cutoff flows (flows that exit the model boundary with no impacts);
• performing an aggregation of the model into a unit process

The disclosure object can also be serialized to JSON or saved as an Excel spreadsheet.

• pandas, to write to excel
• scipy, to perform matrix operations

A disclosure has six parts, including three lists of flows:

• i An ordered list of foreground flows, beginning with the functional unit of the system (length = p);
• ii An ordered list of flows from background processes or activities (length = n);
• iii An ordered list of environmental or social "emissions" (length = m).

And three sparse matrix specifications:

• iv A list of nonzero entries in the p x p sparse matrix defining the foreground;
• v A list of nonzero entries in the n x p sparse matrix defining the foreground's dependency on the background;
• vi A list of nonzero entries in the m x p sparse matrix defining the foreground's direct emissions.

A visual representation of the disclosure is found in Figure 3 of the above paper:

The base Disclosure class is abstract, and a subclass must be created for each supported software type. A subclass must populate the following methods:

• _prepare_efn() returns the evaluated filename used to store disclosure serializations.
• _prepare_disclosure() returns a 6-tuple corresponding to the 6 disclosure elements, in the order listed above.
• __init__() must be written to handle input data and then call the superclass __init__, which computes _prepare_disclosure() and stores the output.