Hello,
The 5th and 6th weeks of the coding period are almost over and it has been a fun time working on the project. During the first month, I initally worked on moving the scripts and their adding/editing functionalities from the retriever to the recipes repository. Later I added a CLI interface to the retriever-recipes repository for adding, deleting and editing scripts. I also wrote a test script to check the installation of modified or newly added scripts. Travis CI was integrated to run remote tests in a docker environment when pushing the code.
Currently, Retriever downloads all the json scripts at once during installation, or whenever the scripts folder in the home directory (~/.retriever directory) is empty. The goal of the second phase of the project was to instead download the scripts only when specifically needed. I mainly stuck with the proposal for this task and added two utility functions namely, get_script_upstream, and get_dataset_names_upstream. The former method is called whenever a script is not available locally. The latter is used for printing the results of retriever ls. It also supports searching of the upstream scripts for keyword and license query parameters, using Github's serach api. Also, necessary code changes were made throughout the project to support this feature.
I have opened a pull request for this task, which is currently in review and testing phase. Now, I would be working on the task of always using the latest version of scripts, i.e. prompting the user if a newer version of script is available upstream. Stay tuned!