cudb (a.k.a. cuda++)

Simple document-based noSQL DBMS modelled after MongoDB. (Has nothing to do with CUDA, has a lot to do with the Cooper Union and DataBases.)

ECE464 Final project by Jonathan Lam, Derek Lee, Victor Zhang

API documentation

Examples

TODO

Functional design

CRUD operations

TODO

Operations on collections

TODO

Documents and data types

TODO

Query semantics

TODO: policies on missing index values

Architectural design

Persistence / filesystem backing store

TODO

Serialization

TODO

Documents and values

TODO

Indices

TODO: managing indices, matching queries to indices, b-trees

Build instructions

cargo is required to build this project.

$ cargo build                                 # build project
$ cargo test -- --nocapture --test-threads=1  # run unit tests
$ cargo doc --no-deps                         # build documentation
$ cargo clean                                 # delete build artifacts

Open documentation in browser

$ cargo doc --no-deps --open

Build documentation to docs/

$ ./build_docs.sh

Test config

$ cargo test -- --nocapture --test-threads=1

• --no-capture: Show stdout even when the test fails.
• --test-threads=1: Prevent running in parallel. This is to prevent multiple concurrent access to the database file.

To run a single test:

$ cargo test [TEST_FN_NAME] -- --nocapture

where [TEST_FN_NAME] is one of the tests defined in the tests/ directory, e.g., test_pool_fetch.

Build without unused warnings

$ RUSTFLAGS="$RUSTFLAGS -A unused_variables -A dead_code" cargo build

These flags can be used for the run and test targets as well.