Syberia is a meta-framework for R that allows for on-the-fly
creation of concrete frameworks for constructing arbitrary software.
In its current formulation, the modeling engine provides an opiniated
framework for developing production-ready machine learning classifiers in R.
The more general vision for Syberia is still in progress, but aims
to unify the currently disparate realms of R packages, script codebases,
Shiny dashboards, R web apps, and reproducible analysis. In the
viewpoint of the author, R is syntactic sugar around LISP, which
enables arbitrary computation; Syberia is an attempt to support this
conjecture by allowing the construction of arbitrary software
projects within the R programming language, thereby finally outgrowing
its long-overdue misconception as a statistical tool.
Syberia is the development framework for R.
The original formulation, developed at Avant, is the modeling grammar that serves as a machine learning classifier development and deployment framework for solving arbitrary academic, research and business problems that require use of statistical methods.
The modeling engine provides an opinionated unified framework for fast iteration on classifier development and deployment. It has modularity and testability built in as a design assumption, is founded on convention-over-configuration, and aims to solve the problems of classifier-specific data preparation and classifier-specific modeling parameters.
The more general vision for Syberia is still in progress, but aims to unify the currently disparate realms of R packages, script codebases, Shiny dashboards, R web apps, and reproducible analysis. In the viewpoint of the author, R is syntactic sugar around LISP, which enables arbitrary computation; Syberia is an attempt to support this conjecture by allowing the construction of arbitrary software projects within the R programming language, thereby finally outgrowing its long-overdue misconception as a statistical tool.
The modeling engine, currently the main significant engine built in Syberia, is a framework for building, debugging, testing, and deploying classifiers developed in R. It includes support for importing data, preparing data, arbitrary statistical modeling methodologies, exporting trained classifiers, validating the results, and deploying as a REST service.
The timeline for future engines and information about how to contribute is listed at the Syberia roadmap.
To get started right away, try out the minimal example syberia project:
# Run this from your command line terminal. git clone firstname.lastname@example.org:syberia/example.sy.git && cd example.sy && R
This will open an R console (installing dependencies for the first time may take a while; for troubleshooting see the troubleshooting guide). You can then type:
run("example1") model$predict(iris[1:5, ]) # The first five scores from a trained classifier. #  5.005686 4.757667 4.773923 4.890092 5.055138
For more detailed instructions, see the installation guide.
Syberia relies on the following supplemental packages:
Additional packages used internally at Avant in conjunction with Syberia modeling projects include batchman, bettertrace, cachemeifyoucan, dokk, lockbox, microserver, objectdiff, Ramd, rocco, s3mpi, testthatsomemore, and treeskeleton.
To run the tests for the Syberia package, you will have to check out its git submodules.
git submodule update --init --recursive
This will pull in
inst/engines/base.sy from the base engine.
This project is licensed under the MIT License:
Copyright (c) 2014-2016 Robert Krzyzanowski
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
Syberia was originally created at Avant by Robert Krzyzanowski, email@example.com. Additional contributors include Ryland Ely, Peter Hurford, Elaine Lee, Tong Lu, and Kirill Sevastyanenko.
Functions in syberia
|ensure_no_global_variable_pollution||Ensure no global variables are polluted during an expression.|
|syberia;||Syberia is a development meta-framework for R.|
|build_engine||Build a syberia engine.|
|test_resource||Run the tests for a single resource.|
|find_test_hook||Fetch the test setup or teardown hook, if any exists.|
|ensure_resources_are_tested||Check that all mandatory tested resources have tests.|
|syberia_engine||Bootstrap a Syberia engine.|
|any_is_substring_of||Whether or not any substring of a string is any of a set of strings.|
|traverse_parent_directories||Perform an action repeatedly on parent directories until success or error.|
|infix_or||Non-null selection operator.|
|active_project||The current active Syberia project.|
|should_exclude||Whether to exclude a syberia engine from being used for resourcing.|
|test_resources||Run the tests on a given set of resources.|
|syberia_version||Fetch the current Syberia version.|
|test_engine||Run all tests in a syberia project or engine.|
|Remotes||syberia/director, syberia/stagerunner, robertzk/bettertrace, robertzk/testthatsomemore|
|License||MIT + file LICENSE|
|Roxygen||list(wrap = FALSE)|
Include our badge in your README