Mongo Reader

Hmmm, looking at the source code, I feel like you did it right?

https://github.com/jerryjliu/llama_index/blob/main/gpt_index/readers/mongo.py#L46

I see the file was updated 3 days ago though 👀

Yes, I always doubt my Python-foo in the first instance. Worthy of me raising a git issue?

I think so! Seems very sus

Oh wait

Did you try calling with the query dict as a kwarg instead? Like the notebook?

I feel like that shouldn't matter but worth trying

Yeah, I had tried that earlier, but that I get the
SimpleMongoReader.load_data() got multiple values for argument 'query_dict'
So I tried naming field_names but get
SimpleMongoReader.load_data() got an unexpected keyword argument 'field_names'
And at that point I think I have the wrong library version, but I've checked the installed files and they have the recent updates.

😵‍💫

wait, so you tried documents = reader.load_data(db_name, collection_name, field_names=field_names, query_dict=query_dict)

if so, sounds like some wacky ass bug lol

More likely me being a numpty . ..
I just searched the venv folder and found the ./readers/llamahub_modules/mongo/base.py file and then a ./readers/mongo.py file, both with different definitions of the load_data function.
I'm not sure the difference yet. These are installed using pip, so would assume it's correct.
Appreciate you taking the tipe to respond to earlier queries - you are one busy person! Cheers

haha no worries! I hope that fixes some stuff by switching! 🙏

haha, found the issue, was me being a numpty I guess.
I was using the llama_index.download_loader('SimpleMongoReader')
instead of a
from llama_index import SimpleMongoReader
Not sure why the difference, but now done.

noooo haha good catch!!!

Yea, there's a separate repo of loaders on llama hub, but the most common ones we have implemented locally too (and obviously they aren't at feature parity, oof)