dcycle / docker-ledger

Dockerized version of https://www.ledger-cli.org/

Dockerized version of the Ledger-cli accounting software, automatically updated weekly.

Features

• Free and open-source.
• No GUI, this project is strictly command-line.
• Double-entry accounting system: data entry will fail if your numbers do not add up.
• This docker image is automatically rebuilt weekly with the latest code, and made available on the Docker hub.

Examples

Download some example files and run:

docker run --rm -v $PWD/examples:/data dcycle/ledger:1 -f /data/sample.dat reg
docker run --rm -v $PWD/examples:/data dcycle/ledger:1 -f /data/drewr3.dat reg
docker run --rm -v $PWD/examples:/data dcycle/ledger:1 -f /data/sample.dat balance
docker run --rm -v $PWD/examples:/data dcycle/ledger:1 -f /data/drewr3.dat balance
docker run --rm -v $PWD/examples:/data dcycle/ledger:1 -f /data/drewr3.dat register Groceries
docker run --rm -v $PWD/examples:/data dcycle/ledger:1 -f /data/drewr3.dat balance Groceries
docker run --rm -v $PWD/examples:/data dcycle/ledger:1 -f /data/drewr3.dat register payee "Organic"
docker run --rm -v $PWD/examples:/data dcycle/ledger:1 -f /data/drewr3.dat cleared

Resources and links:

• What if accounting worked like software development? Command-line, double-entry accounting for agencies, Dcycle Blog, March 4, 2019
• Ledger 3 manual.
• dcycle/ledger on the Docker Hub.
• dcycle/ledger on the Github.
• Command-line, double-entry accounting for agencies.

docker-ledger's Issues

Docker file no longer builds with "ledger (no such package)"

See details in ledger/ledger#2091

Multi-arch image that contains both Intel and ARM versions

The current (Intel-based) docker image works fine on M1 (ARM-based) Macs, but you get an ugly warning about this being run under emulation.

Would it be possible to build a multi-arch image that contains both Intel and ARM versions?

https://www.docker.com/blog/multi-arch-build-and-images-the-simple-way/

(The warning can be silenced by adding a --platform linux/amd64 flag to the run command, but of course, the emulation still takes place)

Determine if it's useful to have configurable timezone, and how best to do it

The timezone is hard-coded. Does this have repercussions on users? If so, how? If so, could we configure the timezone instead of hard-coding.

Version with Python support?

Would it be possible to get a version of this image with Python support? Or would that depend on getting a version with Python support built as an Alpine package?

I assume not everyone would want to include Python support, so I'd be content with using a separate tag for that.

Reduce Image Size

I was wondering if it would be possible to minimize the image size. The current one seems to be based on ubuntu and the total download is >500MB.

Compared to ledger executable, which is <6MB, this is a significant overhead.

Would it be possible to create an (official) image based on alpine linux image?

I noticed some difference between running a Windows version and the current GitHub version of Docker and would really like to have the same version running on different machines and platforms. Docker seems like a potential option but 500MB images are an overkill for that scenario.

I've seen some images (Dockerfiles) based on alpine but was unable to build them due to technical issues in the environment. I will try to resolve this and build an image but would still prefer if there was an official image, like this one.

Cheers and thanks for any consideration!

Image improvements

As suggested in a ledger discussion, use HTTPS URLs for the repositories, and try one of the official mirrors. The main repo (dl-cdn) does not offer HTTPS and nl. server is not on the list (?) but is referenced in the official documentation.

https://uk.alpinelinux.org/alpine/edge/main/
https://uk.alpinelinux.org/alpine/edge/testing/

Support request: can't pull docker container with "latest" tag, must use tag "1" instead

docker pull dcycle/ledger
Using default tag: latest
Error response from daemon: manifest for dcycle/ledger:latest not found: manifest unknown: manifest unknown