brief.Rmd

---
title: "brief"
author: "Paul Oldham"
date: "27 September 2016"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

### Get started with RStudio

RStudio works with projects. 

In the top right of the windown you will see a blue sign with R. Click on that.

1. New Project
2. Give the Project a Name (you will need this later and it will be public)
3. Select a location to save the project (such as desktop or my documents). Then OK. 
4. Go to `Build` in the top menu
5. Choose `Configure Build Tools` . Choose website. 

Ok... we are good to go.

## opsrdev session

We need to load some libraries. You can find each one in `Packages` and tick them or you can write the following in the `Console`.

```{r library, eval=FALSE}
library(opsrdev) # the opsr package
library(dplyr)
library(readr) 
```

## Authenticate

OPS will work for very limited searches without authenticating. 

When you signed up you were asked to obtain a `key` and a `secret`. You can save them in R to use later by typing the following in the console.

Note that the sign is `<-` and the key and secret must be inside quotes "".

```{r}
key <- "BiaLwkHwuus8WjXQPdvVajGGrLMkZnf9"
secret <- "wkvTOQeuLwUGYxcw"
```

Next we can use this information to authenticate

```{r}
library(opsrdev)
ops_auth(key, secret)
```

## What is the status of the service (it can be busy)

```{r}
library(opsrdev)
ops_status()
```

## How many results are there

We can only retrieve upto 2000 results at a time. 

```{r}
library(opsrdev)
ops_count("drones")
```

Ok that's fine. If we had larger results we would want to break them down by year. Let's try pizza.

```{r}
library(opsrdev)
ops_count("pizza")
```

```{r}
library(opsrdev)
ops_count("pizza", "ta", start = 1990, end = 2015)
```

Still too many. Let's try a narrower set. 

```{r}
library(opsrdev)
ops_count("pizza", "ta", start = 2010, end = 2015)
```

So, by experimenting we can arrive at a date range under the 2000 results that we could then retrieve. 

##APIs work with Urls

APIs by definition work with URls. 

The basic url for the EPO OPS service is 

[http://ops.epo.org/3.1/rest-services](http://ops.epo.org/3.1/rest-services)

We won't use this but just to show you how this works programatically. The ops_urls function will create the urls for you. 

```{r}
urls <- ops_urls(query = "pizza", type = "ti", start = 1990, end = 2000)
urls
```

We want to fetch back some bibliographic data. We can use a function called `ops_fetch_biblio()`. As the name suggests it fetches bibliographic data. It uses a set of controls where type = "ta" is titles and abstracts, while type = "ti"

```{r warning=FALSE}
library(opsrdev)
drones <- ops_fetch_biblio(query = "drones", type = "ta", service = "biblio", start = 1990, end = 2015, timer = 10)
```

Let's see that as a table in RStudio.

```{r}
View(drones)
```

We can also write this data to a .csv file. It will write the file to your project folder.

```{r}
write_csv(drones, "drones.csv") 
```

Don't forget quotes or .csv at the end. 

Congratulations, you have now accessed data from an API using R and downloaded the results.