feat: docs for pr review bot (#613)

# Motivation

<!-- Why is this change necessary? -->

# Content

<!-- Please include a summary of the change -->

# Testing

<!-- How was the change tested? -->

# Please check the following before marking your PR as ready for review

- [x ] I have added tests for my changes
- [ x] I have updated the documentation or added new documentation as
needed

---------

Co-authored-by: kopekC <[email protected]>

Author: kopekC

codegen-examples/examples/pr_review_bot/.env.template

@@ -0,0 +1,11 @@
+GITHUB_TOKEN="..."
+# ====[ MODEL PROVIDERS ]===
+
+ANTHROPIC_API_KEY="..."
+OPENAI_API_KEY="..."
+GOOGLE_API_KEY="..."
+GEMINI_API_KEY="..."
+
+# ====[ SLACK ]===
+SLACK_SIGNING_SECRET="..."
+SLACK_BOT_TOKEN="..."

codegen-examples/examples/pr_review_bot/README.md

@@ -1,129 +1,43 @@
 # AI-Powered Pull Request Review Bot
 
-This example demonstrates how to use Codegen to create an intelligent PR review bot that analyzes code changes and their dependencies to provide comprehensive code reviews. The bot uses GPT-4 to generate contextual feedback based on modified code and its relationships.
+This example project demonstrates how to deploy an agentic bot that automatically reviews GitHub Pull Requests. The bot analyzes code changes and their dependencies to provide comprehensive code reviews using AI, considering both direct modifications and their impact on the codebase.
 
-> [!NOTE]
-> This codemod helps development teams by providing automated, context-aware code reviews that consider both direct and indirect code dependencies.
+## Prerequisites
 
-## How the PR Review Bot Works
+Before running this application, you'll need the following API tokens and credentials:
 
-The script analyzes pull requests in several key steps:
+- GitHub API Token
+- Anthropic API Token
+- GitHub Repository Access
 
-1. **Symbol Analysis**
+## Setup
 
-   ```python
-   modified_symbols = codebase.get_modified_symbols_in_pr(pr_number)
-   for symbol in modified_symbols:
-       deps = codebase.get_symbol_dependencies(symbol, max_depth=2)
-       rev_deps = codebase.get_symbol_dependents(symbol, max_depth=2)
-   ```
-
-   - Identifies modified symbols in the PR
-   - Analyzes dependencies up to 2 levels deep
-   - Tracks reverse dependencies (symbols that depend on changes)
-
-1. **Context Building**
-
-   ```python
-   context = {"pr_title": pr.title, "pr_body": pr.body, "modified_symbols": [...], "context_symbols": [...]}
-   ```
-
-   - Gathers PR metadata
-   - Collects modified code content
-   - Includes relevant dependency context
-
-1. **AI Review Generation**
-
-   ```python
-   review = codebase.ai_client.llm_query_with_retry(messages=[...], model="gpt-4", max_tokens=2000)
-   ```
-
-   - Uses GPT-4 for analysis
-   - Generates comprehensive review feedback
-   - Considers full context of changes
-
-## Why This Makes Code Review Better
-
-1. **Context-Aware Analysis**
-
-   - Understands code dependencies
-   - Considers impact of changes
-   - Reviews code in proper context
-
-1. **Comprehensive Review**
-
-   - Analyzes direct modifications
-   - Evaluates dependency impact
-   - Suggests improvements
-
-1. **Consistent Feedback**
-
-   - Structured review format
-   - Thorough analysis every time
-   - Scalable review process
-
-## Review Output Format
-
-The bot provides structured feedback including:
+1. Clone the repository
+1. Set up your environment variables in a `.env` file:
 
+```env
+GITHUB_TOKEN=your_github_token
+ANTHROPIC_API_KEY=your_anthropic_token
+GITHUB_REPO_OWNER=your_repo_owner
+GITHUB_REPO_NAME=your_repo_name
+GITHUB_PR_NUMBER=your_pr_number
 ```
-1. Overall Assessment
-   - High-level review of changes
-   - Impact analysis
-
-2. Specific Code Feedback
-   - Detailed code comments
-   - Style suggestions
-   - Best practices
-
-3. Potential Issues
-   - Security concerns
-   - Performance impacts
-   - Edge cases
-
-4. Dependency Analysis
-   - Impact on dependent code
-   - Breaking changes
-   - Integration considerations
-
-```
-
-## Key Benefits to Note
-
-1. **Better Code Quality**
-
-   - Thorough code analysis
-   - Consistent review standards
-   - Early issue detection
-
-1. **Time Savings**
-
-   - Automated initial review
-   - Quick feedback loop
-   - Reduced review burden
-
-1. **Knowledge Sharing**
-
-   - Educational feedback
-   - Best practice suggestions
-   - Team learning
-
-## Configuration Options
 
-You can customize the review by:
+## Features
 
-- Adjusting dependency depth
-- Modifying the AI prompt
-- Changing the review focus areas
-- Tuning the GPT-4 parameters
+- Automated PR code review using AI
+- Deep dependency analysis of code changes
+- Context-aware feedback generation
+- Structured review format with actionable insights
+- Integration with GitHub PR system
 
-## Learn More
+## Usage
 
-- [Codegen Documentation](https://docs.codegen.com)
-- [OpenAI API Documentation](https://platform.openai.com/docs/api-reference)
-- [GitHub API Documentation](https://docs.github.com/en/rest)
-- [Codegen llm integration](https://docs.codegen.com/building-with-codegen/calling-out-to-llms)
+1. `uv sync`
+1. `uv run modal deploy app.py`
+   - This will deploy a modal app that can be triggered to review PRs
+1. Create or update a PR to trigger the review bot
 
 ## Contributing
 
-Feel free to submit issues and enhancement requests! Contributions to improve the review bot's capabilities are welcome.
+Contributions are welcome! Please feel free to submit a Pull Request.

codegen-examples/examples/pr_review_bot/app.py

@@ -0,0 +1,66 @@
+import logging
+from logging import getLogger
+import modal
+from codegen.extensions.events.app import CodegenApp
+from fastapi import Request
+from codegen.extensions.github.types.events.pull_request import PullRequestLabeledEvent, PullRequestUnlabeledEvent
+from helpers import remove_bot_comments, pr_review_agent
+
+logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(name)s - %(levelname)s - %(message)s")
+logger = getLogger(__name__)
+
+REPO_URL = "https://github.com/codegen-sh/codegen-sdk.git"
+COMMIT_ID = "20ba52b263ba8bab552b5fb6f68ca3667c0309fb"
+
+base_image = (
+    modal.Image.debian_slim(python_version="3.12")
+    .apt_install("git")
+    .pip_install(
+        # =====[ Codegen ]=====
+        # "codegen>=0.18",
+        f"git+{REPO_URL}@{COMMIT_ID}",
+        # =====[ Rest ]=====
+        "openai>=1.1.0",
+        "fastapi[standard]",
+        "slack_sdk",
+    )
+)
+
+app = CodegenApp(name="github", image=base_image, modal_api_key="")
+
+
+@app.github.event("pull_request:labeled")
+def handle_labeled(event: PullRequestLabeledEvent):
+    logger.info("[PULL_REQUEST:LABELED] Received pull request labeled event")
+    logger.info(f"PR #{event.number} labeled with: {event.label.name}")
+    logger.info(f"PR title: {event.pull_request.title}")
+    # app.slack.client.chat_postMessage(
+    #     channel="C08DPPSL1CG",
+    #     text=f"PR #{event.number} labeled with: {event.label.name}",
+    # )
+    if event.label.name == "Codegen":
+        app.slack.client.chat_postMessage(
+            channel="C08DPPSL1CG",
+            text=f"PR #{event.number} labeled with: {event.label.name}, waking up CodegenBot and starting review",
+        )
+
+        logger.info(f"PR ID: {event.pull_request.id}")
+        logger.info(f"PR title: {event.pull_request.title}")
+        logger.info(f"pr number: {event.number}, ")
+        pr_review_agent(event)
+
+
+@app.github.event("pull_request:unlabeled")
+def handle_unlabeled(event: PullRequestUnlabeledEvent):
+    logger.info("unlabeled")
+    logger.info(event.action)
+    logger.info(event.label.name)
+    if event.label.name == "Codegen":
+        remove_bot_comments(event)
+
+
+@app.function(secrets=[modal.Secret.from_dotenv()])
+@modal.web_endpoint(method="POST")
+def entrypoint(event: dict, request: Request):
+    logger.info("[OUTER] Received GitHub webhook")
+    return app.github.handle(event, request)

codegen-examples/examples/pr_review_bot/helpers.py

@@ -0,0 +1,90 @@
+from github import Github
+from codegen.extensions.github.types.events.pull_request import PullRequestUnlabeledEvent
+from logging import getLogger
+
+import os
+
+from codegen import Codebase
+
+from codegen.extensions.github.types.events.pull_request import PullRequestLabeledEvent
+from codegen.configs.models.secrets import SecretsConfig
+from codegen import CodeAgent
+
+from codegen.extensions.langchain.tools import (
+    # Github
+    GithubViewPRTool,
+    GithubCreatePRCommentTool,
+    GithubCreatePRReviewCommentTool,
+)
+
+from dotenv import load_dotenv
+import logging
+
+load_dotenv()
+
+logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(name)s - %(levelname)s - %(message)s")
+logger = getLogger(__name__)
+
+
+def remove_bot_comments(event: PullRequestUnlabeledEvent):
+    g = Github(os.getenv("GITHUB_API_KEY"))
+    logger.info(f"{event.organization.login}/{event.repository.name}")
+    repo = g.get_repo(f"{event.organization.login}/{event.repository.name}")
+
+    pr = repo.get_pull(int(event.number))
+    comments = pr.get_comments()
+    if comments:
+        for comment in comments:
+            logger.info("removing comment")
+            logger.info(comment.user.login)
+            if comment.user.login == "codegen-team":
+                comment.delete()
+    reviews = pr.get_reviews()
+
+    if reviews:
+        for review in reviews:
+            logger.info("removing review")
+            logger.info(review.user.login)
+            if review.user.login == "codegen-team":
+                review.delete()
+
+    issue_comments = pr.get_issue_comments()
+    if issue_comments:
+        for comment in issue_comments:
+            logger.info("removing comment")
+            logger.info(comment.user.login)
+            if comment.user.login == "codegen-team":
+                comment.delete()
+
+
+def pr_review_agent(event: PullRequestLabeledEvent) -> None:
+    # Pull a subset of SWE bench
+    repo_str = f"{event.organization.login}/{event.repository.name}"
+    codebase = Codebase.from_repo(repo_str, language="python", secrets=SecretsConfig(github_token=os.environ["GITHUB_TOKEN"]))
+    review_atention_message = "CodegenBot is starting to review the PR please wait..."
+    comment = codebase._op.create_pr_comment(event.number, review_atention_message)
+    # Define tools first
+    pr_tools = [
+        GithubViewPRTool(codebase),
+        GithubCreatePRCommentTool(codebase),
+        GithubCreatePRReviewCommentTool(codebase),
+    ]
+
+    # Create agent with the defined tools
+    agent = CodeAgent(codebase=codebase, tools=pr_tools)
+
+    # Using a prompt from SWE Bench
+    prompt = f"""
+Hey CodegenBot!
+
+Here's a SWE task for you. Please Review this pull request!
+{event.pull_request.url}
+Do not terminate until have reviewed the pull request and are satisfied with your review.
+
+Review this Pull request like the señor ingenier you are
+be explicit about the changes, produce a short summary, and point out possible improvements where pressent dont be self congratulatory stick to the facts
+use the tools at your disposal to create propper pr reviews include code snippets if needed, and suggest improvements if feel its necesary
+"""
+    # Run the agent
+    agent.run(prompt)
+    comment.delete()

codegen-examples/examples/pr_review_bot/pyproject.toml

@@ -0,0 +1,12 @@
+[project]
+name = "ticket-to-pr"
+version = "0.1.0"
+description = "A example implementation of a agentic linear bot that can create PRs from linear tickets"
+readme = "README.md"
+requires-python = ">=3.12, <3.14"
+dependencies = [
+  "codegen==0.31.1",
+  "fastapi>=0.115.8",
+  "modal>=0.73.51",
+  "pydantic>=2.10.6",
+]