Work Summary

Work Summary Generator

Create easy-to-understand summaries of code changes for non-technical stakeholders and founders.

Key Features

Git Diff Analysis: The feature analyzes Git diffs representing code changes to generate comprehensive summaries.
Natural Language Summaries: Generates detailed summaries in plain language, making technical changes understandable to non-technical stakeholders.
Sectional Breakdown: Provides a structured summary including high-level overview, feature highlights, and impact analysis.
Social Media Integration: Automatically generates Twitter and LinkedIn posts based on the work summary.
Customizable Output: Allows customization of summary depth and focus areas based on stakeholder needs.

How it Works

Provide a Git diff representing the code changes to be summarized.
The tool uses advanced language models to analyze the diff and generate a detailed summary.
The summary includes a high-level overview, sectional breakdown, plain language explanations, feature highlights, and impact analysis.
The tool ensures a consistent, readable structure and can generate social media posts based on the summary.

Work Summary Generation Process

You can find an example here (opens in a new tab)

Usage Guide

Here's a detailed step-by-step guide on how to use the Work Summary Generator:

Follow the initial setup guide here.

Create a new Python file and import the necessary modules:

from kaizen.reviewer.work_summarizer import WorkSummaryGenerator
import requests
from datetime import datetime, timedelta, timezone

Set up GitHub repository information and date range:

GITHUB_OWNER = "YourGitHubUsername"
GITHUB_REPO_NAME = "YourRepoName"
current_date = datetime.now(timezone.utc).date()
since_date = current_date - timedelta(days=14)
since_date_iso = since_date.isoformat()

Fetch commits and generate diff:

# Fetch commits
commits_url = f"https://api.github.com/repos/{GITHUB_OWNER}/{GITHUB_REPO_NAME}/commits"
params = {"since": since_date_iso}
commits_response = requests.get(commits_url, params=params)
commits = commits_response.json()
 
# Generate diff
first_commit_sha = commits[0]["sha"]
last_commit_sha = commits[-1]["sha"]
diff_url = f"https://api.github.com/repos/{GITHUB_OWNER}/{GITHUB_REPO_NAME}/compare/{last_commit_sha}...{first_commit_sha}"
diff_response = requests.get(diff_url, headers={"Accept": "application/vnd.github.v3+json"})
diff_data = diff_response.json()

Extract file diffs:

file_diffs = [
    {
        "file": file_dict["filename"],
        "patch": file_dict["patch"],
        "status": file_dict["status"],
    }
    for file_dict in diff_data["files"]
    if "patch" in file_dict
]

Generate work summary and social media posts:

work_summary_generator = WorkSummaryGenerator()
result = work_summary_generator.generate_work_summaries(file_diffs, user="YourUsername")
summary = result["summary"]
 
twitter_post, * = work_summary_generator.generate_twitter_post(summary, user="YourUsername")
linkedin_post, * = work_summary_generator.generate_linkedin_post(summary, user="YourUsername")

Display the results:

print(f"Work Summary:\n{summary}\n")
print(f"Twitter Post:\n{twitter_post}\n")
print(f"LinkedIn Post:\n{linkedin_post}\n")

Review the generated summary and social media posts, and share them with stakeholders as needed.

Example

To try out fast, here's a complete example to generate a work summary from GitHub commits:

from kaizen.reviewer.work_summarizer import WorkSummaryGenerator
import requests
from datetime import datetime, timedelta, timezone
 
# GitHub repository information
GITHUB_OWNER = "Cloud-Code-AI"
GITHUB_REPO_NAME = "kaizen"
 
# Get the current date and calculate the date 14 days ago
current_date = datetime.now(timezone.utc).date()
since_date = current_date - timedelta(days=14)
since_date_iso = since_date.isoformat()
 
# GitHub API endpoint for getting commits
commits_url = f"https://api.github.com/repos/{GITHUB_OWNER}/{GITHUB_REPO_NAME}/commits"
params = {"since": since_date_iso}
commits_response = requests.get(commits_url, params=params)
 
if commits_response.status_code != 200:
    print("ERROR: Could not get GitHub commits")
    exit(1)
 
commits = commits_response.json()
first_commit_sha = commits[0]["sha"]
last_commit_sha = commits[-1]["sha"]
 
# Get the diff between the first and last commits
diff_url = f"https://api.github.com/repos/{GITHUB_OWNER}/{GITHUB_REPO_NAME}/compare/{last_commit_sha}...{first_commit_sha}"
diff_response = requests.get(diff_url, headers={"Accept": "application/vnd.github.v3+json"})
diff_data = diff_response.json()
 
# Extract file diffs
file_diffs = [
    {
        "file": file_dict["filename"],
        "patch": file_dict["patch"],
        "status": file_dict["status"],
    }
    for file_dict in diff_data["files"]
    if "patch" in file_dict
]
 
# Generate work summary
work_summary_generator = WorkSummaryGenerator()
result = work_summary_generator.generate_work_summaries(file_diffs, user="oss_example")
summary = result["summary"]
 
# Generate social media posts
twitter_post, * = work_summary_generator.generate_twitter_post(summary, user="oss_example")
linkedin_post, * = work_summary_generator.generate_linkedin_post(summary, user="oss_example")
 
# Display results
print(f"Work Summary:\n{summary}\n")
print(f"Twitter Post:\n{twitter_post}\n")
print(f"LinkedIn Post:\n{linkedin_post}\n")

Supported Input

Git diffs
GitHub repository information

Benefits

Bridge the gap between technical and non-technical stakeholders
Keep stakeholders engaged and informed about development progress
Provide a user-friendly way for non-technical stakeholders to track development progress
Streamline the reporting process with automatically generated work summaries
Facilitate communication across different platforms with integrated social media post generation

Limitations

AI Limitations: May not fully capture complex technical nuances or project-specific context.
GitHub API Limitations: Rate limits and authentication requirements may affect large-scale usage.
Summary Depth: The level of detail in summaries may vary based on the complexity of code changes.

End to End Testing Configuring Models