checkPRs_gitHub.py

import requests
import datetime
import json

# Set the GitHub API endpoint and headers
api_url = "https://api.github.com"
headers = {
    "Accept": "application/vnd.github.v3+json"
}

# Set the repository owner and name
repo_owner = "cms-sw"
repo_name = "cmssw"

# Set the label and the username of the "dqm" user
label = "dqm-approved"
dqm_user = "dqm_user"

# Set the start and end dates for the time range
start_date = datetime.datetime(2023, 1, 1)  # Specify your desired start date
end_date = datetime.datetime(2023, 12, 31)  # Specify your desired end date

# Construct the search query for PRs with the "dqm" label within the time range
search_query = f"repo:{repo_owner}/{repo_name} is:pr label:{label} is:closed created:{start_date.date()}..{end_date.date()}"

# Make the API request to search for matching PRs
page = 1
pr_list = []

while True:
    response = requests.get(f"{api_url}/search/issues", params={"q": search_query, "page": page}, headers=headers)
    data = response.json()
    
    if "items" not in data:
        break
    
    pr_list.extend(data["items"])
    page += 1

# Process the PRs and calculate the time to first sign by "dqm"
for pr in pr_list:
    pr_number = pr["number"]

    creation_time = datetime.datetime.strptime(pr["created_at"], "%Y-%m-%dT%H:%M:%SZ")
    closure_time = datetime.datetime.strptime(pr["closed_at"], "%Y-%m-%dT%H:%M:%SZ")
    #print(creation_time)
    #print(closure_time)
    print(f"PR #{pr_number}",pr["title"],"from",pr["user"]["login"]," took: ", closure_time-creation_time)

    #print(f"PR #{pr_number}",pr["title"],"from",pr["user"]["login"])

    #pretty_json = json.dumps(pr, indent=4)
    #print(pretty_json)

    # Make the API request to get the PR details
    #pr_response = requests.get(f"{api_url}/repos/{repo_owner}/{repo_name}/pulls/{pr_number}", headers=headers)
    #pr_data = pr_response.json()
    
    #print(pr_data)

    # # Get the list of reviews for the PR
    # reviews_response = requests.get(f"{api_url}/repos/{repo_owner}/{repo_name}/pulls/{pr_number}/reviews", headers=headers)
    # reviews_data = reviews_response.json()
    
    # # Find the first review from "dqm" and calculate the time to first sign
    # first_sign_time = None
    # for review in reviews_data:
    #     if review["user"]["login"] == dqm_user:
    #         first_sign_time = datetime.datetime.strptime(review["submitted_at"], "%Y-%m-%dT%H:%M:%SZ")
    #         break
    
    # if first_sign_time:
    #     time_to_sign = first_sign_time - datetime.datetime.strptime(pr_data["created_at"], "%Y-%m-%dT%H:%M:%SZ")
    #     print(f"PR #{pr_number}: Time to first sign by 'dqm': {time_to_sign}")
    # else:
    #     print(f"PR #{pr_number}: Not signed by 'dqm'")

checkPRs_gitHub_v2.py

import requests
import datetime
import json
import matplotlib.pyplot as plt

def fill_time_histogram(listOfTimes, start_time, end_time):
    # Calculate the time duration in minutes
    time_delta = end_time - start_time
    total_hours = time_delta.total_seconds() / (60 * 60)
    
    print("took:",total_hours,"hours")
    # Fill the histogram with the time duration
    listOfTimes.append(total_hours)

# Set the GitHub API endpoint and headers
api_url = "https://api.github.com"
headers = {
    "Accept": "application/vnd.github.v3+json"
}

# Set the repository owner and name
repo_owner = "cms-sw"
repo_name = "cmssw"

# Set the label and the username of the "dqm" user
label = "dqm-approved"
dqm_user = "dqm_user"

# Set the start and end dates for the time range
start_date = datetime.datetime(2023, 1, 1)  # Specify your desired start date
end_date = datetime.datetime(2023, 12, 31)  # Specify your desired end date

# Construct the search query for PRs with the "dqm" label within the time range
search_query = f"repo:{repo_owner}/{repo_name} is:pr label:{label} is:closed created:{start_date.date()}..{end_date.date()}"

# Make the API request to search for matching PRs
page = 1
pr_list = []

listOfTimes=[]

while True:
    response = requests.get(f"{api_url}/search/issues", params={"q": search_query, "page": page}, headers=headers)
    data = response.json()
    
    if "items" not in data:
        break
    
    pr_list.extend(data["items"])
    page += 1

# Process the PRs and calculate the time to first sign by "dqm"
for pr in pr_list:
    pr_number = pr["number"]

    creation_time = datetime.datetime.strptime(pr["created_at"], "%Y-%m-%dT%H:%M:%SZ")
    closure_time = datetime.datetime.strptime(pr["closed_at"], "%Y-%m-%dT%H:%M:%SZ")
    #print(creation_time)
    #print(closure_time)
    print(f"PR #{pr_number}",pr["title"],"from",pr["user"]["login"]," took: ", closure_time-creation_time)

    # Fill the histogram with the time duration
    fill_time_histogram(listOfTimes, creation_time, closure_time)

    #print(f"PR #{pr_number}",pr["title"],"from",pr["user"]["login"])

    #pretty_json = json.dumps(pr, indent=4)
    #print(pretty_json)

    # Make the API request to get the PR details
    #pr_response = requests.get(f"{api_url}/repos/{repo_owner}/{repo_name}/pulls/{pr_number}", headers=headers)
    #pr_data = pr_response.json()
    
    #print(pr_data)

    # # Get the list of reviews for the PR
    # reviews_response = requests.get(f"{api_url}/repos/{repo_owner}/{repo_name}/pulls/{pr_number}/reviews", headers=headers)
    # reviews_data = reviews_response.json()
    
    # # Find the first review from "dqm" and calculate the time to first sign
    # first_sign_time = None
    # for review in reviews_data:
    #     if review["user"]["login"] == dqm_user:
    #         first_sign_time = datetime.datetime.strptime(review["submitted_at"], "%Y-%m-%dT%H:%M:%SZ")
    #         break
    
    # if first_sign_time:
    #     time_to_sign = first_sign_time - datetime.datetime.strptime(pr_data["created_at"], "%Y-%m-%dT%H:%M:%SZ")
    #     print(f"PR #{pr_number}: Time to first sign by 'dqm': {time_to_sign}")
    # else:
    #     print(f"PR #{pr_number}: Not signed by 'dqm'")

# Display the histogram
# Create a histogram with suitable range and binning
hist, bins, _ = plt.hist(listOfTimes, bins=100, range=(0, 400))

plt.xlabel('Time (hours)')
plt.ylabel('PR number')
plt.title('Time Duration Histogram')
# Save the figure as a PNG file
plt.savefig('histogram.png')