References

Note

This document has been machine translated.

Web API

Event data acquisition API

• eventdata: Obtain event data

Map acquisition API

• riskmap: Obtain risk map
• predictmap: Obtain prediction map
• contour: Get contour polygon of precipitation data

Route Search API

• dijkstra: Search for a safe and close route based on the risk at the specified date and time on the specified risk map.
• calcrisk: Calculate the risk exposure for the specified route from the start point to the end point.

Data Loader API

• overview.
• registerloader:Initialize the table to populate with event data by addrecord.
• deleteloader:Deletes a table for data input.
• getloderlist:Display list of data input tables
• initrecord:Get a template for event data submission for use with addrecord.
• addrecord: submit event data

File Retrieval API

• getfile: Obtain information on files in event data saved by data loader WebAPI addrecord.

Alert Notification API

• Overview](alert_overview.md)
• Alert Settings: Register and subscribe to risk themes you wish to monitor
• Change alerts](alert_alter_alert.md): Change conditions for alerts on monitored risk themes
• Alert Notification: Receive alert notification messages
• Alert Notify API Client: JavaScript implementation of the Alert Notify API client

Provenance API

• Operation for DDC
  • prov.count_ddc_records: count the number of records
  • prov.ddc_exists: DDC existence check
  • prov.get_ddc_columns: get column information
  • prov.resolve_ddc: Get DDC metadata
  • prov.process: Open API process execution
  • prov.get_ddc_activities: Get DDC process history
  • prov.get_ddc_last_activity: Get DDC's last activity
  • prov.get_ddc_processing_status: Get DDC's execution status
• Processing execution control
• table view operations
  • prov.new_ddc: create table
  • prov.get_ddc_columns: get column information
  • prov.count_ddc_records: count the number of records
  • prov.set_ddc: partial selection of records/columns
  • prov.put_ddc_records: add records
  • prov.join_ddc: join tables
  • prov.delete_ddc: delete table
  • prov.begin_session, prov.resume_session, prov.begin_or_resume_session, prov.rollback, prov.commit: Transaction
  • prov.set_ddc, prov.resolve_ddc: conversion between table name and ddc
• Configuration information

Session API

• Session API List
  • session.open_session: start a session
  • session.verify_session_token: check token validity
  • session.close_session: close session

Module Execution API

• dataproc.exec: Execute a method of a module.
• dataproc.get_definitions: get module definition information
• dataproc.get_server_version: get server version of module execution API

Event Data Analysis API

• Extracting local correlation rules
  • overview
  • train_fpgrowth: Correlation pattern extraction by FP-Growth
  • import_category: auxiliary API for FP-Growth, register category definitions
  • extract_items: Auxiliary API for FP-Growth, symbolic conversion of transaction table based on category definitions
  • shuim: Frequent itemset extraction with SHUIM
  • extract_distance: Auxiliary API for SHUIM, calculate the distance between each point in the transaction table
• Deep learning method for spatio-temporal correlation patterns
  • Overview
  • train_crnn: CRNN model training
  • predict_crnn: apply trained CRNN model to predict
  • train_lgbm: LGBM model training
  • predict_lgbm: apply trained LGBM model for prediction
• Compound event analysis
  • overview
  • train_3dcnn: 3D-CNN model training
  • predict_3dcnn: Apply trained 3D-CNN model for prediction

Federated Learning API

• Federated Learning API

  • fl.get_server_version: get server version of Federated Learning API
  • fl.initialize_model_store: create model store
  • fl.exists_store: Check existence of model store
  • fl.delete_store: Delete a model store
  • fl.put_model: Store a model in the given model store
  • fl.put_aggregated_model: Store models in the given model store
  • fl.feedback: Feedback specified model to upstream model store
  • fl.proxy_feedback: send the given model to the upstream model store feedback
  • fl.load_model: Download a given model in a given model store
  • fl.request_transfer: retrieve the latest aggregated model from upstream model store and store it in the given model store
  • fl.proxy_request_transfer_model: retrieve the latest aggregated model from upstream model store
  • fl.proxy_request_transfer_info: Get the latest aggregated model information from the upstream model store
  • fl.get_model_info: get information about a given model from a given model store
  • fl.exists_model: Check for the existence of a given model in a given model store
  • fl.delete_model: Delete a given model in a given model store
  • fl.delete_all_model: Delete all models in the given model store
  • fl.get_models: Retrieve all models information for a given model store
  • fl.get_model_info_latest: Get latest model information for a given model store
  • fl.get_latest_feedbacked_models: get list of latest fedbacked models information from a given model store
  • fl.get_latest_aggregated_model: Get the latest aggregated model information from the given model store
  • fl.get_latest_transferred_model: Get the latest transferred model information from the given model store
  • fl.get_parties: Get parties from upstream
  • fl.create_party: register a party in the upstream
  • fl.delete_party: delete a party from the upstream

• xData Aggregation FW

  • class Manager: Class that manages the configuration of Aggregator and Party
  • class Aggregator: Class that executes a series of processes of federated learning as Aggregator
  • class Trainer: Class that executes a series of processes of the associative learning as a Party
  • class Serializer: Class that describes persistence and restoration methods when sending and receiving models
  • class RemoteStore: Model store with Federated Learning Web API as backend

Algorithm Module

• Module definition file