Júlio César Batista
Júlio César Batista
No strong opinion as well. I think it wold be a nice addition, though we need to be careful on what will be our threshold of accepted utility dependencies. Optional...
I think we can work something here. Mostly because of the API. When I call `loader.load_item()` I'd expect all the fields I defined to be there, even if I set...
Indeed, I'm in favor of having a flag or specialized `ItemLoader` for this behavior. I think it's weird to `loader.add_value('field', None)` and not have the field in the output. Even...
This is a good idea. Probably it would be easier than jsonschema to write some validations and checks :smile: . Since, we don't need to bother about names, I'd suggest...
@manycoding , my idea here is, if it is only a config for `report_all`, then it would be better to set it only in `report_all`. Since most of the time...
IMHO, they should be different. I might accept a diff of 0.3 in categories but only 0.05 for coverage for example. So ``` a = Arche(source, target) a.report_all( coverage_threshold=(0.1, 0.05),...
Here is an implementation https://github.com/heylouiz/scrapy-sticky-meta-params
Just adding an idea here. What if we had a flag `FEED_EXPORT_WARN_DATA_LOSS` (bool). Then, if there was any data loss (different fields then the header) it would log a message....