How prerna2415/fabricnet works
Project type and purpose
• This is a containerized Python application whose primary goal is to package and run a pre-trained Keras “FabricNet” model inside Docker.
• Key indicators: presence of Dockerfile, .devcontainer folder, Python scripts (main_app.py), Keras model files (*.keras), requirements.txt, runtime.txt.
File structure
Top-level:
• .devcontainer/ – VS Code devcontainer configuration (dev workflow)
• Dockerfile – instructions to build a Docker image
• main_app.py + requirements.txt + runtime.txt – entry point and dependency lists
• fabricnet_best_fold_3.keras – model file at root (duplicate in main/)
• main/ – likely source folder with the same artifacts (model, app and requirements)
Patterns:
• Single container app (no microservices)
• Dev environment (VS Code Remote-Containers) and production container share same Dockerfile
README
• Very brief: “FABRIC NET DOCKERIZATION CODE”
• Implies focus on container orchestration of the model rather than multi-service orchestration
Main components (to be shown
a. Developer Workstation
– VS Code with Remote-Containers extension reading .devcontainer/devcontainer.json
b. Docker Build
– Docker daemon using Dockerfile + requirements.txt + runtime.txt
c. Docker Image
– Base Python runtime (e.g. from runtime.txt)
– Installed Python dependencies (from requirements.txt)
– Keras model file (fabricnet_best_fold_3.keras) copied into image
– main_app.py copied into image as container entrypoint
d. Docker Container at runtime
– Runs main_app.py which loads the FabricNet model and serves inference
e. (Optional) External Resources
– GPU or CPU hardware
– Logging or monitoring if configured
Relationships & data flow
• Developer Workstation → builds → Docker Image
• Docker Image → instantiated as → Docker Container
• Docker Container reads → model file and Python dependencies
• main_app.py processes inputs → runs inference → outputs predictions
Architectural patterns & design principles
• Single-container deployment pattern
• Clear separation between development environment definition (.devcontainer) and production Dockerfile
• Immutable infrastructure: build once, run anywhere