Hi,
I've finally gotten (a) account structure I want for now (b) importers working for my 3 core accounts. I now want to get smart_importer working and want to understand the best strategy to bring this into the process.
My understanding from the smart_importer data is that I can covert all my importers to be "smart importers" using the hooks that smart_importer provides as noted in the readme file. I think that is straightforward. I can use this on any importer type bank, cc, etc.
So my core question then, is it best to only manually label 1-2 months of data and then use that labelled data to train the smart_importer and then run the smart_importer version on the rest of my data? For example, I have 1.5 years of OFX data from my bank in one OFX file. I can run the current non-smart importer on the entire file but then I have to manually label the entire thing. There isn't a totally different way to dynamically do this as I edit in emacs or use smart_importer in a different way? If I understand I should split into 1.3 years and 0.2 years for example. Manually label the 0.2 years and then take the remaining 1.3 years and run through the smart_importer version of the importer? Or is there some other way I should go about this?
Also want to say, I really appreciate the beancount tool, so many amazing contributions (e.g. Reds importers), and the new smart_importer tool and everyone helping answer questions.
thanks,
Jonathan