Gnucash
1 How to Use It:
2 How to Use It (part II):
2.1 Parsing
gnucash-read
2.2 A Note On Representations
2.3 Libs.rkt
3 Example.rkt
8.12

Gnucash🔗ℹ

John Clements <clements@racket-lang.org>

 (require gnucash) package: gnucash
This collection is intended to allow you to use Racket/DrRacket to operate on gnucash files. It’s flimsy but effective for me. It’s intended for Racket hackers, and is underdocumented, under-test-cased, and under-error-checking-ed. For heaven’s sake don’t try using this stuff unless you’ve already written a bunch of Racket code.
Olin Shivers talks about 80% solutions; I would call this a 60% solution. Here’s why you’d use it: it has example code, and it runs. Let me know if you improve it.

1 How to Use It:🔗ℹ

The very simplest way to see how it works is to take a look at the examples directory: it contains a simple example gnucash file, and a simple Racket file with plenty of comments that opens the gnucash file and pulls out a few of the transactions in the file. If you still have questions, come back and read this file.

2 How to Use It (part II):🔗ℹ

The library consists of two files. One reads in the data (parse.rkt), and one has some utility functions for handling it. Both of them are loaded automatically.

2.1 Parsing🔗ℹ

procedure

(gnucash-read gnucash-file gnucash-zo-file)

  (listof sxml-thing)
  gnucash-file : path-string?
  gnucash-zo-file : path-string?
The reader does a funny thing: since XML reading is so appallingly slow, it makes a .zo file that contains the compiled representation of the input file. Building this thing is horrifically slow, taking about 3 minutes to process my 3-megabyte gnucash file (actually, it’s 19M after unzipping it). This is all hidden from you, except that when you call the reader you must supply the name of a cache file.

If the source file is newer than the zo file, it goes and recompiles the darn thing.

You probably won’t need to call any other functions in parse.rkt.

2.2 A Note On Representations🔗ℹ

Darn near everything in this library is represented using plain-old Racket lists, and more specifically SXML representations. The up side of this is that you can "display" nearly everything transparently. The down side of this is that you will curse and scream at a bunch of errors that would be caught by any kind of type system or use of structures, e.g. using an account instead of an account id, etc.

Note from 2015: Let’s use TR! That’s definitely a better use of my time than grading all these exams....

2.3 Libs.rkt🔗ℹ

Here’s how to require it:

(require gnucash/libs)

I’m not even going to try to document a significant subset of the functions in lib.rkt; they’re mostly like this one:

> (transaction-splits t)

Returns the splits associated with the transaction

> (transaction-date t)

Returns the date associated with the transaction

Here’s one worth mentioning:

> (crossers transactions account-ids accounts)

For most kinds of graphing/reporting/munging/etc., you will want to choose a set of accounts and look for all transactions that "cross the line", in the sense that they transfer money from that set of accounts to something else. More accurately, you want transactions where at least one split is inside the set of accounts, and where at least one split is outside the set of accounts.

Why is this? Well, imagine you have a checking account and a credit card. You want to take a look at what you’re spending your money on. You’d like to lump together the checking account and the credit card account and take a look at the money coming into or out of these two accounts, while ignoring transfers between these two. This is what "crossers" is for.

Please do note that the "account-ids" is a list of ids (that is, strings), and not accounts. The example file shows this happening.

3 Example.rkt🔗ℹ

For an example of how to use this, take a look at "examples/examples.rkt".

Well, that’s about all I have to say for now. Let me know if you find it useful or if it makes you grind your teeth.

John Clements, 2007-08 (updated repeatedly, most recently 2018-07-01)