Accounts Assessor - introduction

This repository hosts practical research into leveraging logic programming to solve accounting problems. The core logic runs in SWI-Prolog, and is aided by several smaller python codebases.

We use it at http://www.nobleaccounting.com.au to automate reporting and auditing tasks.

Several services are available:

• investment calculator
• hirepurchase agreement
• depreciation
• livestock (standalone)
• division 7A loan calculator

investment calculator

The most complex endpoint is the investment calculator; it validates and processes financial data of a financial entity for a given period:

• bank statements
• general ledger transactions
• forex and investments
• Australian SMSF accounting
• livestock accounting

It automates some accounting procedures, like tax calculations, and generates balance sheets, trial balances, investment report and other types of reports.

livestock calculator

depreciation calculator

hirepurchase calculator

(todo: UI)

Given a hire purchase arrangement, it can track the balance of a hire purchase account through time, the total payment and the total interest.

Division 7A Loan calculator

chatbot services

• determine tax residency by carrying out a dialog with the user
• determine small business entity status by carrying out a dialog with the user

in detail

running the server with docker

1. git clone --recurse-submodules https://github.com/lodgeit-labs/accounts-assessor/
2. docker_scripts/up.sh

usage

with Excel and LSU plugin

https://github.com/koo5/accounts-assessor-public-wiki/blob/master/excel_usage/README.md

with OneDrive explorer

navigate to https://robust1.ueueeu.eu/static/onedrive-explorer-js/ and log in. Log into OneDrive, the choose a file and click "run investment calculator".

through file upload form

1. Load http://localhost:8877/view/upload_form in your browser
2. upload one of the example input files

with curl

curl -F file1=@'tests2/endpoint_tests/loan/single_step/loan-0/request/request.xml' http://localhost:8877/upload

with test runner

see tests2/runner/README.md

through a custom GPT

we have an experimental endpoint for custom GPT, which you can register as an "action" - https://robust1.ueueeu.eu/ai3/openapi.json - see [/wiki/CustomGPT.md]

input files

endpoint_tests

• tests/endpoint_tests/**/request/*
• todo make sure that this includes whatever we generate from rdf templates - Robust Input Example 8.7.2021

how to generate an input file from template

documentation

todo, this is all private at the moment:

most endpoints should have some resources in lodgeit_private/doc/. Introductions to individual concepts can be found in videos on dropbox.

videos: https://www.dropbox.com/sh/prgubjzoo9zpkhp/AACd6YmoWxf9IUi5CriihKlLa?dl=0 https://www.dropbox.com/sh/o5ck3qm79zwgpc5/AABD9jUcWiNpWMb2fxsmeVfia?dl=0

wiki: https://github.com/lodgeit-labs/accounts-assessor-wiki/

architecture

there are 4 main components:

services

various helper functions that prolog invokes over http/rpc

workers

a remoulade worker that:

• wraps prolog and spawns prolog on request
• talks to the triplestore

frontend

lets users upload request files and triggers workers.

apache

serves static files and proxies requests to frontend

`

directory structure

• source/lib - prolog source code
• tests ** plunit - contains queries that test the functionality of the main Prolog program ** endpoint_tests - contains test requests for the web endpoint as well as expected reponses
• misc - contains the stuff that does not yet clearly fit into a category
• server_root - this directory is served by the prolog server ** tmp - each request gets its own directory here ** taxonomy - contains all xbrl taxonomy files. ** schemas - xsd schemas

version 2.0

A new version is planned, using a constrained logic programming language, aiming for these features:

Derives, validates, and corrects the financial information that it is given. The program uses redundancy to carry out its validations and corrections. By this it is meant that knowledge of parts of a company's financial data imposes certain constraints on the company's other financial data. If the program is given a company's ledger, then it knows what the balance sheet should look like. If the program is given a company's balance sheet, then it has a rough idea of what the ledger should look like.

• Given a hire purchase arrangement and ledger, it can guess what the erroneous transactions are
• Given a hire purchase arrangement and ledger, it can generate correction transactions to fix the erroneous transactions ...

todo: comparisons to other projects

• https://github.com/johannesgerer/buchhaltung
• gnu cash ...

Contributors ✨

Thanks goes to these wonderful people (emoji key):

stoopkid 🚇 ⚠️ 💻

Schwitter 🚇 ⚠️ 💻

salamt2 🚇 ⚠️ 💻

Murisi Tarusenga 🚇 ⚠️ 💻

koo5 🚇 ⚠️ 💻

This project follows the all-contributors specification. Contributions of any kind welcome!