Shell Python Gnuplot
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
README.md
WTFPL.txt
ledgerrc
report
reportgraph.py
script.gnuplot

README.md

ledger-reports

Generate informational graphs from ledger-cli accounting files.

reportgraph.py

Generate a directed graph (in GraphViz markup) of how money moves within the file.

May not work with multiple commodities, fixes are welcome.

report

Creates multiple interesting graphs from the data via gnuplot and piechart.

Designed to be easily extendable (by editing it) with queries interesting to yourself.

Example outputs

A pie chart, showing asset split

A line chart, showing monthly expenses

A bar chart, showing monthly food expenses

Requirements

  • At least ledger 3
  • gnuplot
  • piechart
  • The files ledgerrc and script.gnuplot from this repository

Usage

Create a directory reports/.

Run as ./report <ledger file> [periods <period expressions>]

If no ledger file is specified, main.ledger is assumed.

The resulting graphs will be placed in the reports/ directory.

The script uses the following categories in the current state

  • Assets
  • Assets:Liquid
  • Expenses
  • Expenses:Food
  • Liabilities
  • Income

If your ledger file uses a different layout or you want to create new graphs, you will have to edit the report script. The drawing tool invocations are abstracted away to a large degree via bash functions.

Note that the ledgerrc file currently contains the strict option, which requires that all accounts and commodities be pre-announced. If you don't want that, remove the option from the file.

The periodic feature

The periodic feature is experimental and might not work exactly as expected.

Reporting periods must be exactly equal-width buckets, meaning that the query result must have the exact same number of lines.

This also includes things such as dates without a transaction (a problem which may be sidestepped by using --empty) and leap years, making --daily probably a bad window. --weekly --empty is probably a good minimum interval.

The periodic report will create the same type of graphs from the same queries as the regular report, but will try to plot the data for every period as its own line.

Data which can not reliably be segmented into equal-width periods should be plotted with the \_aperiodic functions, which always produce the same plot.

Example invocations

./report

./report whatever.ledger

./report main.ledger periods 2015 2016