GitStudy

Lazy cat OIDC login has been integrated. ##Functional characteristics - GitHub project search: Enter keywords to search public repositories and quickly select projects to learn. - Introduction to core code: Automatically grab the core document of the project and generate Chinese explanations, structured understanding paths and pre-knowledge. - Code question and answer and question determination: Support multiple rounds of questioning around the current document, and also support automatic question determination and feedback on code questions. - AI classroom and article mode: Project content can be organized into teaching slides, technical articles and graphical learning materials. - Browser Lab: Built-in TensorFlow.js interactive experiments such as attention, diffusion, CNN, and decision boundaries to help map abstract concepts with real source code. - Learning data precipitation: The learning history, collections, wrong questions books and rankings are all saved in the `data/`directory of the server and can still be retained after restarting. ###Required Configuration 1. AI API is required; if it is not configured, GitStudy can still open the page, but core capabilities such as AI interpretation, question and answer, question determination, and slide show cannot work. 2. `GITHUB_TOKEN` is optional; if you encounter GitHub rate limit when searching or reading large warehouses at high frequencies, you can add a token in "AI Settings". 3. `GITHUB_GIST_TOKEN` is also optional; only needed if you want to use the "Open in Colab" feature, which will be used to create or update a private Gist. ###How to get GitHub Token It is recommended to use GitHub's official Fine-grained personal access token first, which makes it easier to tighten permissions. 1. Log in to GitHub, click the avatar in the upper right corner, and enter `Settings`. 2. Enter Developer settings on the left. 3. Open `Personal access tokens`->`Fine-grained tokens`. 4. Click `Generate new token`, and the name can be filled in as `GitStudy GitHub API`. 5. Resource owner Choose your own GitHub account. 6. Repository access Select on demand: - If you mainly study public repositories, according to GitHub official instructions, fine-grained token will by default include read-only access to all public repositories. - If you still want GitStudy to read your own private repository, add the corresponding repository to Selected Repositories. 7. If you want to read the contents of your own private repository, it is recommended to grant an additional grant of 'Repository permissions'->'Contents: Read-only'. 8. Copy the token after generation and fill in it in the `GitHub Token` field in GitStudy's "AI Settings"; if you are accustomed to using environment variables, you can also write `GITHUB_TOKEN`. 9. Note: Fine-grained token usually starts with 'github_pat_'; the 'ghp_...' in the input box It is just a place-holder example and does not affect storage and use. The main purpose of this token is to increase the GitHub API limit and reduce the probability of encountering API rate limit exceeded when searching for projects, reading README, and code files. ###How to get GitHub Gist Token The "Open in Colab" feature calls the GitHub Gist API to create or update a private `. ipynb` file for the current project; therefore a token with Gist write permissions is required. 1. Log in to GitHub, click the avatar in the upper right corner, and enter `Settings`. 2. Enter Developer settings on the left. 3. Open `Personal access tokens`->`Fine-grained tokens`. 4. Click `Generate new token`, and the name can be filled in as `GitStudy Colab Gist`. 5. Resource owner Choose your own GitHub account. 6. Find Account permissions in Permissions and set Gists to Write. 7. Copy the token after generation and fill in the `GitHub Gist Token` field in GitStudy's "AI Settings"; if you want to use environment variables, you can also write `GITHUB_GIST_TOKEN`. 8. Note: Fine-grained tokens usually start with 'github_pat_'; GitStudy can be used normally as long as the permissions are correct.