code2docs-ai-core

中文

GitHub Action: Code2Docs-ai Agent Workflow

This repository contains a GitHub Action workflow that allows you to create a new repository in the current organization based on an input repository URL.

Usage

To use this GitHub Action workflow, follow these steps:

1. Trigger the workflow with the following input parameters:

   • repoUrl: The GitHub repository URL in HTTPS format.
   • branchName: The branch name to use (default value is main).

2. The workflow will perform the following steps:

   • Print out the input parameters values in the workflow log.
   • Validate that the repoUrl is a GitHub.com address and a HTTPS address. If not, the job will fail.
   • Validate that the repoUrl exists on GitHub and is publicly accessible. If not, the job will fail.
   • Extract the orgName and repoName from the repoUrl, and concatenate them to form another parameter docs_repo_name = orgName_repoName.
   • Check if the repository docs_repo_name exists using the GitHub API.
   • If the repository exists, delete it before creating a new one.
   • Create a new repository in the current organization with the name docs_repo_name using the GitHub API.
   • Check if the local directory already exists and remove it if it exists before cloning the repository.
   • Clone the source from repoUrl at the end of the workflow, using the branchName input parameter to specify which branch to clone.
   • Clean up the working directory on the runner before running any steps.

Example

Here is an example of how to trigger the workflow:

on:
  workflow_dispatch:
    inputs:
      repoUrl:
        description: 'The GitHub repository URL in HTTPS format'
        required: true
      branchName:
        description: 'The branch name to use'
        default: 'main'

This will trigger the workflow and create a new repository in the current organization based on the input repoUrl.

Note

The workflow requires a GITHUB_TOKEN secret to authenticate and create the new repository.

Error Handling

The Create new repo step includes error handling to ensure that the job fails if the repository creation fails. The curl command used to create the repository includes the --fail option to ensure it fails on error, and || exit 1 to exit the job if the repository creation fails.

Additionally, the workflow now includes validation to check if the repository docs_repo_name exists before attempting to create it. If the repository exists, it will be deleted before creating a new one. This ensures that the repository creation process does not fail due to an existing repository with the same name.

The workflow also includes validation to check if the repoUrl exists on GitHub and is publicly accessible. If the repoUrl does not exist or is not publicly accessible, the job will fail. This ensures that the repository creation process does not fail due to an invalid or inaccessible repoUrl.

Additional Steps

The workflow now includes additional steps to activate the python environment, run the 'aise-cli' command, and parse the repo using 'aise-cli repo parse-repo':

1. Activate the python environment:

   source ~/source/aise-cli/.venv/bin/activate

2. Run the 'aise-cli' command to test if the python environment works:

   aise-cli

3. Delete the source repo and parse the repo using 'aise-cli repo parse-repo':

   # delete source repo
   aise-cli repo delete-repo source_repo
   # parse repo
   aise-cli repo parse-repo --repo_path  aise-cli repo parse-repo --repo_name <source_repo absolute>

Storing Workflow Run State

The workflow now includes functionality to store the state of workflow runs in a JSON file named workflow_runs.json. This file includes the following values for each workflow run:

• Workflow run id
• repoUrl and branchName from the input parameters
• docs_repo_name
• Current status and conclusion
• Workflow run created time

The workflow_runs.json file is created or updated during each workflow run, and the new record of the current workflow run is added to the top of the existing content of the file. The file is then committed and pushed to the repository.

Updating Workflow Run State

The workflow now includes a step at the end to update the workflow_runs.json file with the status set to 'completed' and the 'completed_at' field with the current time. This ensures that the workflow run state is accurately recorded and updated.