Import modules



In [ ]:

    
import os
from tinydb import TinyDB
import pandas as pd
import time
from DashPykpi.kpistats import KpiStats, GitURLs, GraphKPIs

Create DB

Need to cover the repos assoicated with the below github user accounts

✔ Indicates the repos are covered by rc-softdev-admin account...



In [ ]:

    
# or... use a list of URLS fetched from the GitURLs class
url_fetch = GitURLs()
urls = url_fetch.urls
print("Retrieved {0} urls.".format(len(urls)))



In [ ]:

    
# These projects don't exist anymore and break get_repo_stats
urls.remove('https://github.com/UCL/ucl')
urls.remove('https://github.com/UCL-RITS/ucl-rits')
urls



In [ ]:

    
# needed = [
#     "https://github.com/UCL-RITS",
#     "https://github.com/astro-informatics",
#     "https://github.com/bempp",
#     "https://github.com/DCPROGS",
#     "https://github.com/OPTIMET",
#     "https://github.com/UCL-Biochemical-Engineering",
#     "https://github.com/UCL-CSS",
#     "https://github.com/UCLProgrammingHub",
#     "https://github.com/UCL-HIC",
#     "https://github.com/EIT-team",
#     "https://github.com/Astrophysics-UCL",
#     "https://github.com/Bahler-Lab",
#     "https://github.com/CCPPETMR",
#     "https://github.com/Euclid-OULE3",
#     "https://github.com/FieldingChemistryUCL",
#     "https://github.com/PancreaticProject",
#     "https://github.com/ShapsUcl",
#     "https://github.com/msslsolar",
#     "https://github.com/mssl-plasma-group",
#     "https://github.com/UCL-ShippingGroup"]
    
# need_accounts = [n.split('/')[3] for n in needed]
# got_accounts = [url.split('/')[-2] for url in urls]



In [ ]:

    
# from github3 import GitHub
# fn = open("secret_key")
# g = GitHub(token=fn.read().split()[0])

# for thing in g.iter_user_repos(login='benlaken'):
#     print(thing)
    
# for need in need_accounts:
#     if not need in got_accounts:
#         print("Missing {0} repos ".format(need, need in got_accounts))
#         print("    Attempting to access - will work if public")
        
#         for thing in g.iter_user_repos(login=need):
#             print("    {0}".format(thing))



In [ ]:

    
#for url in urls:
#    user_str, repo_str = url.split('/')[-2:]
#    print(url, user_str, repo_str)



In [ ]:

    
# Run the code to generate a dababase if needed...
test = KpiStats(urls=urls)
test.work(verbose=False, debug=False, status=True)

Plotting section



In [ ]:

    
db = TinyDB('tinydb_for_KPI.json')



In [ ]:

    
df = pd.DataFrame(db.all())



In [ ]:

    
df.head()



In [ ]:

    
#for n in df['repo_owner']:
#    print(n)



In [ ]:

    
running = 0
num_repos = 0
tmp_hold = {}
for n, weekly in enumerate(df['weekly_commits']):
    if sum(weekly) > 1:
        tmp_hold[df['repo_name'][n]] = weekly
        running += sum(weekly)
        num_repos += 1
print("{0:3,} commits, in {1} active repos (out of {2} total repos), during past 52 weeks".format(
        running, num_repos, len(df)))



In [ ]:

    
import numpy as np



In [ ]:

    
tmp = []
for n, weekly in enumerate(df['weekly_commits']):
    if sum(weekly) > 1:
        tmp.append(weekly)

tmp = np.array(tmp)
tmp = tmp.sum(axis=0)
all_weekly_commits = {"All repos": tmp}



In [ ]:

    
width=4
data = tmp

result = data[:(data.size // width) * width].reshape(-1, width).sum(axis=1)
result
print(52/4)



In [ ]:

    
all_weekly_commits

Stacked area chart



In [ ]:

    
#from bokeh.charts import Area, show, output_notebook, defaults
#import numpy as np
from bokeh.charts import show, output_notebook
output_notebook()



In [ ]:

    
bk = GraphKPIs()
show(bk.weekly_activity(per_repo=False, verbose=True, bin=4))