| jupytext | kernelspec | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
|
import pandas as pd
URL to use GH API and query trainer repository for all issues, both open and closed up to the maximum (100)
trainer_issues_url = 'https://api.github.com/repos/carpentries/trainers/issues?per_page=100&state=all'
read them into a dataframe
df_r = pd.read_json(trainer_issues_url)
df_r.head()
df_r.columns
df_r.loc[0]
unpack the columns that have json objects inthem still
unpack_cols = ['user','pull_request','reactions','labels']
#
final_col = {True: lambda col,ser: ser.apply(pd.Series).rename(columns=lambda c: '_'.join([col,str(c)])),
False: lambda col,ser: ser}
df_pieces = [final_col[col_i in unpack_cols](col_i,df_r[col_i]) for col_i in df_r.columns]
df = pd.concat(df_pieces,axis=1)
df.shape
df_r.shape
df.columns
the label columns are still objects, unpack those more
label_cols = ['labels_0','labels_1']
#
label_cols = [df[l].apply(pd.Series).rename(columns=lambda c: '_'.join([l,str(c)])) for l in label_cols]
df = pd.concat([df]+label_cols,axis=1)
df.columns
create filters
approved = df['labels_1_name']=='approved'
proposal = df['labels_0_name']=='Proposal'
create subsets of the issues
df_approved = df[approved]
df_proposal = df[proposal]
add a column to the approved dataframes with a link to their page
mdlink = lambda r: '['+ r['title'] + ']('+ r['html_url'] + ')'
df_approved['md'] = df_approved.apply(mdlink,axis=1)
print them out for use in the report
for approved_t in df_approved['md']:
print(approved_t)
df_proposal = df[proposal]
df_proposal['md'] = df_proposal.apply(mdlink,axis=1)
df_open = df_proposal[df_proposal['labels_1_name']!='approved']
for approved_t in df_open['md']:
print(approved_t)