KeyError: avg_order_size, which in features_config, yet not integrated into X in processor.py, but used at Line 207?
Joecth opened this issue · 2 comments
Hi, I get KeyError while running the QuickStart steps,
Could you please help to give a hint?
Thanks!
Following shows the Pic regarding my problem,
That is to say,
At line 207, In processor.py,
According to the Quick Start, the avg_order_size is from
features_config[“features”],
features_config = {
"features": {
"avg_order_size": {"type": "N"},
...
And, I can’t see anywhere X in processor.py, which is mainly from data_frame, integrates avg_order_size into itself,
Data_Frame:
row | mdp_id | sequence_number | context | decision | reward |
---|---|---|---|---|---|
1 | u123 | 1 | {"country": 2, "year": 1796} | female | 154.2 |
2 | u456 | 1 | {"country": 4, "year": 2017} | male | 170.9 |
which makes me don't know how to fix it by making sense...
Hey @Joecth - features
should match the features present in your context
. See example_feature_config.json
for an example.
Thanks for your reply, @econti!
Just now, I already simply replace features_config on QuickStart page with the content in example_feature_config.json, shown as the pic,
I can see the features
includes country
, year
, and decision
, which is already consistent with context
in DataFrame.
However, looks like it receives an empty array in preprocess_feature() in proprocessor.py, and thus raised ValueError...
Is it possible for you to guide a quick fix regarding what happened here?
(p.s. I just wanna smoothly and successfully run it up to get more idea abt it according to QuickStart.., but still not able to... )
Thanks~!