analogcyan / itadakimasu Goto Github PK

In the current setup, Itadakimasu relies on three distinct services — the frontend hosted on Cloudflare, backend hosted on Google Cloud, and full recipe generation leveraging OpenAI API for AI capabilities. This configuration might lead to inefficiencies in managing and optimizing our resources. I propose we fully migrate our project to be managed entirely on Google Cloud Platform (GCP).

This move towards consolidating our services on Google Cloud Platform (GCP) could potentially simplify our project management and make our infrastructure more efficient. By adapting to a single cloud service provider, we can leverage GCP's robust infrastructure for both front-end and back-end services, enhancing consistency and reliability.

Moreover, rather than using the OpenAI API for the generation of recipes, we could consider utilizing Google Cloud's AI APIs, which offer powerful AI and machine learning capabilities directly integrated within the platform. This consolidation might lead to easier system configuration, more cohesive support, and potential infrastructure cost saving.

Shifting Itadakimasu entirely to GCP could also tackle existing reliability issues related to our setup. This restructuring simplifies cross-platform communication, reducing inconsistencies and helping to streamline problem troubleshooting.

Either via a selector by the generate button, or simply always random within a range.

In the current iteration of Itadakimasu, there's a notable inconsistency issue with the recipe generation API. It's been observed that this API doesn't consistently return expected results, which impacts the application's secondary functionality—generating full recipes. This inconsistency merits further investigation. This may also be an issue relating to the OpenAI API we rely on.

In the v2 update of Itadakimasu, we're encountering a persistence issue affecting initial and subsequent loads of the application when browser cache isn't manually cleared. This behavior hinders a satisfactory user experience as the application fails to load as expected. Further investigation is necessary to address and resolve this persistent issue.

Moving the food list to a dedicated text file would make it easier to manage. Ideally a simple text file with each item on its own line, imported on page load into the foods array (unless there's some other better way of doing this that's slipped my mind).

Replace font-size-adjust with something browser-agnostic so header displays as intended on all browsers.