How shashank975/ml_productionready_project works
US Visa Classification is an ML decision-support application: an operator can launch model training, while a user submits application attributes for an approval/denial prediction. Training orchestrates ingestion, validation, transformation, training, evaluation, and conditional model publishing; prediction loads the published model and applies its preprocessing. The graph reflects README capabilities and indexed imports; unsampled implementation details and web-template wiring are not inferred.
Main entry points: training_pipeline.py Training orchestrator · main_utils.py ML utilities · app.py Web app · data_ingestion.py Data ingestion
Components 21
Application
- Web app
app.py
Model training
- Training orchestrator
US_visa/pipline/training_pipeline.py - Data ingestion
US_visa/components/data_ingestion.py - Data validation
US_visa/components/data_validation.py - Feature transformation
US_visa/components/data_transformation.py - Model trainer
US_visa/components/model_trainer.py - Model evaluation
US_visa/components/model_evaluation.py - Model publisher
US_visa/components/model_pusher.py
Prediction serving
- Prediction pipeline
US_visa/pipline/prediction_pipeline.py - S3 estimator
US_visa/entity/s3_estimator.py - Model and preprocessing
US_visa/entity/estimator.py
Data and model infrastructure
- Visa data access
US_visa/data_access/usvisa_data.py - Mongo connection
US_visa/configuration/mongo_db_connection.py - MongoDB
- S3 storage
US_visa/cloud_storage/aws_storage.py - AWS connection
US_visa/configuration/aws_connection.py - AWS S3
- Stage artifacts
US_visa/entity/artifact_entity.py - Pipeline configuration
US_visa/entity/config_entity.py - ML utilities
US_visa/utils/main_utils.py
Other components
- Operator
Connections 24
- Operator → Web app: uses
README.md - Web app → Training orchestrator: starts training
app.py - Web app → Prediction pipeline: requests prediction
app.py - Training orchestrator → Data ingestion: runs ingestion
US_visa/pipline/training_pipeline.py - Training orchestrator → Data validation: runs validation
US_visa/pipline/training_pipeline.py - Training orchestrator → Feature transformation: runs transformation
US_visa/pipline/training_pipeline.py - Training orchestrator → Model trainer: runs training
US_visa/pipline/training_pipeline.py - Training orchestrator → Model evaluation: runs evaluation
US_visa/pipline/training_pipeline.py - Training orchestrator → Model publisher: publishes accepted model
US_visa/pipline/training_pipeline.py - Training orchestrator → Stage artifacts: passes stage results
US_visa/pipline/training_pipeline.py - Data ingestion → Visa data access: exports collection
US_visa/components/data_ingestion.py - Visa data access → Mongo connection: uses connection
US_visa/data_access/usvisa_data.py - Mongo connection → MongoDB: connects to
US_visa/configuration/mongo_db_connection.py - Prediction pipeline → S3 estimator: loads and predicts
US_visa/pipline/prediction_pipeline.py - S3 estimator → S3 storage: loads model
US_visa/entity/s3_estimator.py - S3 estimator → Model and preprocessing: predicts with
US_visa/entity/s3_estimator.py - Model publisher → S3 storage: uploads model
US_visa/components/model_pusher.py - S3 storage → AWS connection: creates client
US_visa/cloud_storage/aws_storage.py - AWS connection → AWS S3: connects to
US_visa/configuration/aws_connection.py - Data ingestion → Pipeline configuration: uses settings
US_visa/components/data_ingestion.py - Data validation → ML utilities: uses helpers
US_visa/components/data_validation.py - Feature transformation → ML utilities: uses helpers
US_visa/components/data_transformation.py - Model trainer → ML utilities: uses helpers
US_visa/components/model_trainer.py - Prediction pipeline → ML utilities: loads YAML
US_visa/pipline/prediction_pipeline.py