/referia

Reviewing framework initially created for reviewing REF papers then heavily modified.

Primary LanguagePython

referia

Tests

codecov

The referia library provides tools for assisting with assessment, originally written as an aide for 2021 REF Assessment, code has gone through many addtions and then a major restructuring for version 0.2.0

The library uses jupyter notebook as an interface.

The library builds on functionality provided in the lynguine data oriented architecture library. The main difference between the two is that functionality that is general for the flow-based model the code follows sits in lynguine. The referia code provides convenience functionality for ease of creation of reviewing software.

To install use

%pip install referia

Configuration is stored in a local file, _referia.yml.

This file provides the source and format of input data (for assessment), the location and format of annotations to store on the assessment as well of details of how to view files and urls associated with the assessment.

The configuration has the following fields

input

This contains the input data to be assessed.

viewer

Provides information at the top of the score sheet that gives background (for example reviewing instructions).

editpdf

Lists pdfs to copy and allow the user to edit (for example to make notes on a submitted thesis).

Subfields are field which contains the filename, sourcedirectory and storedirectory. Also provides ability to specify pages from the source, so that we have something like

field: ColumnName0
sourcedirectory: ./
pages:
  first: ColumnName1
  last: ColumnName2
storedirectory: ./pdfs

Functionality is provided in system.Sys.edit_files.

urls

Lists urls that should be opened for providing additional information on the review.

output

This specifies how the annotation information is to be stored.

scored

Specifies how the code should "count up" how many reviews are complete.

series

This specifies how to store annotation information which is available with a subindex (such as a time series).

review

This specifies a set of widgets to create for human reviewing.

compute

Specifies fields that should be filled in by computations.

documents

Allows for the creation of documents (such as word docx or emails) that summarise the provided information.

summary_documents

Similar to documents, but this section provides documents that summarise all the data. Can be useful for summarising a sequence of reviews.