README.md

In mienkoja/oliveR: A package for extracting data from the oliver SMS replica and creating measurements

oliveR

Database access and measurement development for the Oliver SMS. The package is designed to work with the oliver-opencpu-server-docker and the oliver_replica psql database.

Prerequisites for Mac and Mac-ish environments

1. A public/private key pair with the public key shared to [e-mail] and residing on the reverse tunnel target.
2. A valid username on the reverse tunnel target. Can be set by contacting [e-mail].
3. Authy account information should be forwarded to \email{[e-mail]} prior to trying to establish a connection to the replica with the command string or access will be denied. Access to oliver_replica requires two-factor authentication. Authy (\url{https://www.authy.com/}) is used for this purpose.
4. A connection to the oliver_replica database. This is accomplished with a complex ssh statement, a sample of which is as follows: ssh -i /Users/mienkoja/.ssh/id_rsa -p 5431 -L 10.200.10.1:5432:oliver-replica.criploulbgnu.us-west-2.rds.amazonaws.com:5432 -N mienkoja@52.90.57.218. This statement can be generated for a specific user from within oliveR using the create_ssh_command() function.
5. A functioning instance of oliver-opencpu-server-docker.
6. Unless you want to specify your credentials manually, the build_all_metrics() function is set to pull connection parameters for oliver_replica from the environment. These can be set according within your favorite shell environment or can be set within R using Sys.setenv() as shown below:

Sys.setenv(OLIVER_REPLICA_DBNAME = "oliver_replica"
           ,OLIVER_REPLICA_HOST = "10.200.10.1"
           ,OLIVER_REPLICA_USER = "mienkoja"
           ,OLIVER_REPLICA_PORT = "5432"
           ,OLIVER_REPLICA_PASSWORD = "my_password")

Prerequisites for Windows

• Placeholder for Windows instructions

Installation

oliveR can be installed from from github with:

```{r, eval = FALSE}

install.packages("devtools")

devtools::install_github("hadley/pkgdown")

## Usage

The interaction between [oliver-opencpu-server-docker](https://github.com/pocdata/oliver-opencpu-server-docker) and oliveR relies on the present of one or more "metric" objects. On the initial install of oliveR, no metric objects are present as the data contained within may be sensitive. The first time you load oliveR (and each time you want to refresh the metric objects), you should update the metric objects with: 

```{r, eval = FALSE}
oliveR::build_all_metrics()

This will drop the objects into an rds formatted file in the oliveR directory. The objects are then loaded into the global environment with:

```{r, eval = FALSE} oliveR::load_measurement_objects()

Metric objects are used within oliveR to create lists of metric information - values and geometry. This is accomplished with running methods from a `metric_group`. A `metric_group` object serves as a container for one or more metric objects. For our purposes we are currently only dealing with `metric_performance_provider` objects, an R6 object which currently contains two methods: 1. `get_value` (a method to produce a single numeric value for a given metric), and 2. `get_donut` (a method to produce svg geometry representing a donut plot. The function, `get_metric_list()`, is used to generate a nested list which contains the returns from the `get_value` and `get_donut` methods on all metric objects within a given metric group by:

```{r, eval = FALSE}
oliveR::get_metric_list(metric_group = my_metric_group, case_id = my_case_id)

There are multiple ways that this can be replicated within oliver-opencpu-server-docker. The following demonstrates how this can be accomplished to convert the list output of get_metric_list to json.

curl http://localhost/ocpu/library/oliveR/R/get_metric_list/json -H "Content-Type: application/json" -d '{"case_id":my_case_id}'

mienkoja/oliveR documentation built on May 6, 2019, 6:01 p.m.