Updated readme and workflow.

.github/workflows/code_review.yml

@@ -4,17 +4,24 @@ on:
     types:
       - opened
       - synchronize
+      - ready_for_review
 permissions: write-all
 jobs:
+  if: '! github.event.pull_request.draft'
   code_review:
+    if: '! github.event.pull_request.draft'
     runs-on: ubuntu-latest
     steps:
       - name: Checkout repository
         uses: actions/checkout@v3
       - name: Code Review
-        uses: freeedcom/ai-codereviewer@main
+        uses: lukehollenback/ai-codereviewer@main-luke
         with:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
           OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
-          OPENAI_API_MODEL: "gpt-4-1106-preview"
-          exclude: "yarn.lock,dist/**"
+          OPENAI_API_MODEL: "gpt-3.5-turbo"
+          exclude: "yarn.lock, dist/**, **/*.json, **/*.md, **/*.yaml, **/*.xml"
+          custom_prompts: |
+            Do not worry about the verbosity of variable names, as long as they are somewhat descriptive.
+            Be sure to call out concurrency issues and potential race conditions.
+            Do not worry about things a static analyzer would catch in real-time during development.

README.md

@@ -1,14 +1,15 @@
 # AI Code Reviewer
 
-AI Code Reviewer is a GitHub Action that leverages OpenAI's GPT-4 API to provide intelligent feedback and suggestions on
-your pull requests. This powerful tool helps improve code quality and saves developers time by automating the code
-review process.
+AI Code Reviewer is a GitHub Action that leverages OpenAI's large language models to provide intelligent feedback and
+suggestions on  your pull requests. This powerful tool helps improve code quality and saves developers time by
+automating the code review process.
 
 ## Features
 
-- Reviews pull requests using OpenAI's GPT-4 API.
+- Reviews pull requests using OpenAI's large language models.
 - Provides intelligent comments and suggestions for improving your code.
 - Filters out files that match specified exclude patterns.
+- Supports custom prompting to hone the analysis in on specific things, and avoid others.
 - Easy to set up and integrate into your GitHub workflow.
 
 ## Setup
@@ -29,35 +30,55 @@ on:
     types:
       - opened
       - synchronize
+      - ready_for_review
 permissions: write-all
 jobs:
   review:
+    if: '! github.event.pull_request.draft'                                       # Only run on non-draft pull requests. Must be coupled with the ready_for_review event above.
     runs-on: ubuntu-latest
     steps:
       - name: Checkout Repo
         uses: actions/checkout@v3
 
       - name: AI Code Reviewer
-        uses: your-username/ai-code-reviewer@main
+        uses: your-username/ai-codereviewer@main
         with:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} # The GITHUB_TOKEN is there by default so you just need to keep it like it is and not necessarily need to add it as secret as it will throw an error. [More Details](https://docs.github.com/en/actions/security-guides/automatic-token-authentication#about-the-github_token-secret)
-          OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
-          OPENAI_API_MODEL: "gpt-4" # Optional: defaults to "gpt-4"
-          exclude: "**/*.json, **/*.md" # Optional: exclude patterns separated by commas
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}                               # Keep this line verbatim. GitHub sets secrets.GITHUB_TOKEN automatically, it just needs to be plumbed into the action. See https://docs.github.com/en/actions/security-guides/automatic-token-authentication#about-the-github_token-secret for more information.
+          OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}                           # Your OpenAI key. Create one at https://platform.openai.com/api-keys. Set it in your repository under Settings -> Secrets and variables -> Actions -> Repository Secrets.
+          OPENAI_API_MODEL: "gpt-4"                                               # Optional: The OpenAI model (https://platform.openai.com/docs/models) to use. Defaults to "gpt-4". Prefer models with JSON Mode (https://platform.openai.com/docs/guides/text-generation/json-mode) support.
+          exclude: "yarn.lock, dist/**, **/*.json, **/*.md, **/*.yaml, **/*.xml"  # Optional: File patterns, separated by commas, to exclude from analysis.
+          custom_prompts: |                                                       # Optional: Custom commands to add to the prompt used to analyze code. This is a multiline string, where each line is an individual command.
+             Do not worry about the verbosity of variable names, as long as they are somewhat descriptive.
+             Be sure to call out potential null pointer exceptions.
+             Be sure to call out concurrency issues and potential race conditions.
 ```
 
 4. Replace `your-username` with your GitHub username or organization name where the AI Code Reviewer repository is
-   located.
+   located. If using the [original](https://github.com/freeedcom/ai-codereviewer), replace it with `freeedcom`.
 
 5. Customize the `exclude` input if you want to ignore certain file patterns from being reviewed.
 
-6. Commit the changes to your repository, and AI Code Reviewer will start working on your future pull requests.
+6. Customize or remove the `custom_prompts` input.
+
+7. Commit the changes to your repository, and AI Code Reviewer will start working on your future pull requests.
 
 ## How It Works
 
 The AI Code Reviewer GitHub Action retrieves the pull request diff, filters out excluded files, and sends code chunks to
 the OpenAI API. It then generates review comments based on the AI's response and adds them to the pull request.
 
+## Tips & Tricks
+
+* The `gpt-4` model is powerful, but [relatively expensive](https://openai.com/pricing) and does not support
+  [JSON Mode](https://platform.openai.com/docs/guides/text-generation/json-mode) at the time of writing. Consider using
+  one of the turbo models — such as `gpt-4-turbo-preview` or `gpt-3.5-turbo` — to find the right balance between price
+  and performance.
+* Use `custom_prompts` to hone the action's analysis in on things you care about, and to avoid things you don't. There
+  is no point, for example, in having the action repeat what your static analysis tooling does for you in real-time
+  while coding.
+* Use an `if: '! github.event.pull_request.draft'` job conditional and `ready_for_review` trigger to control costs and
+  ensure that you are only running the action on pull requests that are truly ready to be reviewed.
+
 ## Contributing
 
 Contributions are welcome! Please feel free to submit issues or pull requests to improve the AI Code Reviewer GitHub

action.yml

@@ -11,10 +11,18 @@ inputs:
     description: "OpenAI API model."
     required: false
     default: "gpt-4"
+  max_tokens:
+    description: "Maximum number of tokens that can be generated per analysis."
+    required: false
+    default: "700"
   exclude:
     description: "Glob patterns to exclude files from the diff analysis"
     required: false
     default: ""
+  custom_prompts:
+    description: "Custom commands to augment the agent's prompts with. Each line is an individual command."
+    required: false
+    default: ""
 runs:
   using: "node16"
   main: "dist/index.js"