Server-Side Coding

Basic Operation (for Python)

It will be easier to understand if you check the sample project below:

https://github.com/MasahideMori-SimpleAppli/delta_trace_db_py_server_example

Currently, only Python server-side samples are provided, but a similar system can be created in Dart as well. If you prefer to implement everything in Dart, please create equivalent server-side code in Dart.

Server-side programming with DeltaTraceDB is very simple.

  1. Authenticate the user using JWT or any other method.

  2. Recover the query from the request JSON.

Example:

from delta_trace_db import UtilQuery

query_json = await request.json()
query = UtilQuery.convert_from_json(query_json)

  1. Execute the query while setting database permissions based on the authenticated user’s privileges. Permissions are whitelisted by collection, so grant only the permissions you wish to allow.

collection_permissions = {
    "users": Permission([
        EnumQueryType.add,
        EnumQueryType.getAll,
        EnumQueryType.search
    ])
}
result = delta_trace_db.execute_query_object(
    query=query,
    collection_permissions=collection_permissions
)

Note

The Permission object is part of the serializable Actor class. You can create an Actor instance for each user and use the collection_permissions property within it to manage access rights more cleanly.

Note that execute_query_object is a method that automatically determines whether it is a QueryObject, TransactionQueryObject, or Dict and processes it appropriately.

Warning

On the server side, be sure to set Permission class that grant appropriate privileges to each user, or implement a permission granting program.

Also, when using Permission classes, be sure to use the Permission classes that exist on the server side, as the classes sent from the front end may have been tampered with.

For front-end convenience, if permissions are set to “None,” the DB will allow all operations.

  1. It stores the query, handles any errors, and then returns the results to the front end.

# After the preservation process etc.
return result.to_dict()

DeltaTraceDB uses JSON for its query structure, making it easy to store queries persistently. By saving the query data to disk, you can resume operations from any state, even after a crash. We recommend always saving queries to enable full traceability of all database operations.

As shown in the sample, you can also convert the entire DB to JSON using delta_trace_db.to_dict().

Basically, we recommend doing both of the following:

  • Regularly back up your entire database.

  • Always back up the queries.

For more information on saving strategies, see Best Practices.

Precautions for Use

When implementing server-side code for DeltaTraceDB, keep the following in mind:

  • DeltaTraceDB is single-threaded and designed to run on a single server instance.

  • It operates as an in-memory database, so if multiple instances are launched in parallel, each one will maintain its own isolated in-memory database.

For cloud environments, the simplest and most reliable configuration is to run a single DeltaTraceDB instance and use it as the shared database.

Alternatively, for lightweight setups such as cloud functions, you can load and use a user-specific database from storage on demand whenever a request is made. Alternatively, it is also cost-effective and useful to manage a user-specific database on the front-end side and periodically save it to cloud storage.

API