Asset Allocation Classes-- TypeError: unsupported operand type(s) for /: 'str' and 'int'

[kantskernel]

I received the below error when trying to access "Asset Allocation Classes" in Fava Investor.

I was able to get around it by casting meta_value to a Decimal on line 122 and 123.

Exception on /generic-users-finances/extension/Investor/ [GET]
Traceback (most recent call last):
  File "/Users/generic-user/.virtualenvs/beans/lib/python3.9/site-packages/flask/app.py", line 2525, in wsgi_app
    response = self.full_dispatch_request()
  File "/Users/generic-user/.virtualenvs/beans/lib/python3.9/site-packages/flask/app.py", line 1822, in full_dispatch_request
    rv = self.handle_user_exception(e)
  File "/Users/generic-user/.virtualenvs/beans/lib/python3.9/site-packages/flask/app.py", line 1820, in full_dispatch_request
    rv = self.dispatch_request()
  File "/Users/generic-user/.virtualenvs/beans/lib/python3.9/site-packages/flask/app.py", line 1796, in dispatch_request
    return self.ensure_sync(self.view_functions[rule.endpoint])(**view_args)
  File "/Users/generic-user/.virtualenvs/beans/lib/python3.9/site-packages/fava/application.py", line 360, in extension_report
    content = Markup(render_template_string(template, extension=extension))
  File "/Users/generic-user/.virtualenvs/beans/lib/python3.9/site-packages/flask/templating.py", line 159, in render_template_string
    return _render(app, template, context)
  File "/Users/generic-user/.virtualenvs/beans/lib/python3.9/site-packages/flask/templating.py", line 130, in _render
    rv = template.render(context)
  File "/Users/generic-user/.virtualenvs/beans/lib/python3.9/site-packages/jinja2/environment.py", line 1301, in render
    self.environment.handle_exception()
  File "/Users/generic-user/.virtualenvs/beans/lib/python3.9/site-packages/jinja2/environment.py", line 936, in handle_exception
    raise rewrite_traceback_stack(source=source)
  File "<template>", line 239, in top-level template code
  File "/Users/generic-user/.virtualenvs/beans/lib/python3.9/site-packages/fava_investor/__init__.py", line 22, in build_assetalloc_by_class
    return libassetalloc.assetalloc(accapi, self.config.get('asset_alloc_by_class', {}))
  File "/Users/generic-user/.virtualenvs/beans/lib/python3.9/site-packages/fava_investor/modules/assetalloc_class/libassetalloc.py", line 203, in assetalloc
    asset_buckets = bucketize(vbalance, accapi)
  File "/Users/genric-user/.virtualenvs/beans/lib/python3.9/site-packages/fava_investor/modules/assetalloc_class/libassetalloc.py", line 122, in bucketize
    asset_buckets[bucket] += amount.number * (meta_value / 100)
TypeError: unsupported operand type(s) for /: 'str' and 'int'

I recently moved to using OFX and the corresponding reds importer for OFX files and it introduced some Assets:Zero-Sum-Accounts:Transfers that I believe created some "unknown" asset_alloc_by_class. I added this to my fava_config to get unknown to zero but even after doing that it looks like I need to cast meta_value to decimal to get Asset Allocation Classes to load

[redstreet]

This casting should not be required. Casting is likely hiding your true problem. And hmm, postings to Assets:Zero-Sum-Accounts:Transfers may not be related at all. The unknown warning occurs when your commodity metatadata is not fully filled out, or non-existent for a certain commodity. See here.

The only requirement is that the metadata field name begins with the prefix asset_allocation_, and has a number for its value that is a percentage, corresponding to the percentage of the commodity belonging to that asset class. The set of all asset classes for a commodity should add up to a 100. When they do not, the reporter will pad the remaining with the 'unknown' class.

Check and let me know if that solved it?

[kantskernel]

Yep all solved. Thanks for pointing me in the right direction.

TypeError: unsupported operand type(s) for /: 'str' and 'int'

I actually had some asset_allocation defined in my commodities file as str accidentally (e.g. asset_allocation_x "100" instead of asset_allocation_x 100).

Very embarrassing.

[redstreet]

Great! Thanks for sharing the issue too. Heh, while it's awesome that Beancount allows arbitrary metadata types, I've tripped up these a few times myself. The code should probably catch this exception and provide a helpful warning to the user (eg: "Your metadata turned out to be a string, while an integer was expected" rather than just a stack trace. Will do at some point.