Paul Lo

Oh yeah, I didn't think of weekend issue. Just made some more changes on those two files using urllib3 consistently (previously opened by you in issue #7)

@toomore 針對你提的"六日或是沒開市日"的問題，我剛才加了一些程式碼: base on你原本的TWSEOpen d_day()延伸了一個recent_open_day() function，如果今天沒開市就往前一天去check，直到有開市的那天為止。 (原本在想會不會有infinite loop，但我還沒有試出這樣的日期，看起來d_day()支援的時間區間似乎滿完整的)

It works for me, @helenykwang do you mind providing the errors you run into?

Thanks for reaching out, @vincewu51 added _pretrain_ parameter just last month in #511, a quick workaround before upgrading causalml's version can be ignore the parameter (e.g. `estimate_ate(X=X, treatment=treatment, y=y)` )...

@jeongyoonlee I think this can be relevant to #413 that I can find that EconML supports hyperparameters method such as GridSearchCV (example in the [usage FAQs](https://econml.azurewebsites.net/spec/estimation/dr.html#usage-faqs)) and additional `kwargs` parameters...

Thanks for reaching out, @ya-stack Do you mind providing a complete error msg/ log for the ValueError? that would be helpful

if the error comes from that `estimate_ate()` line, let me know how the below works (one possible cause is [this](https://stackoverflow.com/questions/45346550/valueerror-unknown-label-type-unknown)): ``` slearner.estimate_ate(X, w_multi, y.astype('int')) ``` and I would want to...

hmm so it seems `fit_predict()` worked fine, can you click on the "5 frames" to expand the error msg for me? and have you tried passing in with `y_train.astype('int')`? thanks

@ZhangXInFD Thanks for reaching out, it sounds like a similar issue one of our colleagues ran into. To narrow down, can you try **passing in 1 or 2 for n_jobs...

Thanks for the details! Looks like it's actually a different issue that I didn't see it before. I can reproduce it when running the pytest. @jeongyoonlee do you know what...