Added local_folder param to get_result_df (#394)

[shivamsanju]

This PR addresses issue #394 :
Previously users were not getting an option to download the result dataframe on the local directory because a temporary directory was getting created which was cleaned up after reading the dataframe.

Changes:
Now an optional parameter called local_folder is added where users can pass the local directory name where they want to download the result. If the user does not pass this parameter, the function will create a temporary directory and clean it after reading the dataframe.

[xiaoyongzhu]

@shivamsanju thanks for the contribution! Overall this looks good to me. I put a few comments inline.

Can you also merge main branch back? There might be some typos that we just fixed in the main branch which might be impactful for this PR.

[xiaoyongzhu]

@shivamsanju can you merge the main branch of Feathr into this PR? Seems like those are the root cause for those test failures

[shivamsanju]

@shivamsanju can you merge the main branch of Feathr into this PR? Seems like those are the root cause for those test failures

Hi @xiaoyongzhu, I have already merged the main branch of linkedin/feathr to this PR (commit b2ee907). It says already up to date when I do a git pull from linkedin/feathr.

[xiaoyongzhu]

@shivamsanju can you merge the main branch of Feathr into this PR? Seems like those are the root cause for those test failures

Hi @xiaoyongzhu, I have already merged the main branch of linkedin/feathr to this PR (commit b2ee907). It says already up to date when I do a git pull from linkedin/feathr.

Ah OK so it might be a small typo that I've pointed out (feathr_spark_laucher to feathr_spark_launcher)

[Yuqing-cat]

It would be better to have a log to record the local folder path if not None.