// the find
metalalive/e_commerce
E-commerce backend platform implemented in Python / C / Rust
A learning-focused e-commerce backend split into six independent services (user management, product catalog, storefront, media processing, order processing, payment). Each service picks its own tech stack — Python/Django for most, C for the media service, Rust for order/payment. The author is explicit that this is not production-ready.
The media service is genuinely ambitious: hand-rolled HTTP server in C with FFmpeg transcoding, HLS output, multipart upload, and an RPC consumer — not the usual CRUD tutorial fare. The hexagonal architecture is applied consistently in the Rust and C services with clear adapter/domain separation. CI workflows exist per-service with real integration test infrastructure (RabbitMQ setup, DB init scripts, Stripe test workflow). The common Python and Rust libraries are factored out properly instead of copy-pasted across services.
One star, one fork, and the README license placeholder says 'Your License Here' — this is a solo portfolio project with no community or production validation whatsoever. The polyglot stack (Python + C + Rust) massively increases the operational surface area for no stated reason beyond learning; anyone trying to run this in a team faces three separate build toolchains and mental models. The inter-service communication relies on Celery/AMQP RPC, which means tight coupling through message contracts and no typed service boundaries — the 'service-oriented' claim is undermined by this. Documentation for actually standing up and running the full system end-to-end is scattered across per-service READMEs with no single working quickstart.