lipd2neotomaDemo.Rmd

---
title: "lipd2neotoma Demo"
author: "Dave Edge"
date: "2022-12-05"
output: html_document
---

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

## This is a demonstration of the new `lipd2neotoma()` function

### I will use a neotoma dataset, called Bambili 2, for this demonstration

First, we will load this record and convert into lipd format

```{r}
B <- neotoma2::get_sites(sitename = "Bambili 2")
D <- neotoma2::get_downloads(B)
L <- neotoma2lipd(D)
```

### LiPD

Let's see a summary of the record in LiPD

```{r warning=FALSE}
summary(L)
```


### Neotoma

Now we will convert back to neotoma and observe the changes

```{r}
N <- lipd2neotoma(L)

D@sites[[1]]
N

```

The site printout looks good

Site level metadata

```{r, error=TRUE}
D
D@sites
D@sites[[1]]
N



D@sites[[1]]@siteid
N@siteid


D@sites[[1]]@sitename
N@sitename

D@sites[[1]]$geography
N@geography

D@sites[[1]]$altitude
N@altitude

D@sites[[1]]$geopolitical
N@geopolitical

D@sites[[1]]$area
N@area

D@sites[[1]]$notes
N@notes

D@sites[[1]]$description
N@description

D@sites[[1]]$collunits
N@collunits
```

We do pretty good here with the exception of "geopolitical", which could be added to the lipd under "geo"

We also see an error in trying to print the "collunit", which is detailed later

### Sample datum

```{r}
D@sites[[1]]@collunits@collunits[[1]]@datasets@datasets[[1]]@samples@samples[[1]]@datum
N@collunits@collunits[[1]]@datasets@datasets[[1]]@samples@samples[[1]]@datum
```

note that `D` is the original neotoma dataset and `N` is the new one
`N` is a `site` rather than a `sites`

Aside from some loss in the `variablename`, these data frames look identical.
Improved `variablename` fidelity is possible with alteration to the `neotoma2lipd()` function

What's the difference between element and elementtype?

Let's jump to the last `datum` to see if this fidelity holds

```{r}
length(D@sites[[1]]@collunits@collunits[[1]]@datasets@datasets[[1]]@samples@samples)
D@sites[[1]]@collunits@collunits[[1]]@datasets@datasets[[1]]@samples@samples[[140]]@datum
N$collunits@collunits[[1]]@datasets@datasets[[1]]@samples@samples[[140]]@datum
```

again, we see some minor changes to `variablename`, but otherwise the data frame is identical

### Chronologies

Let's have a look at the chronology data

```{r}
D@sites[[1]]@collunits@collunits[[1]]@chronologies@chronologies[[1]]@chroncontrols
N$collunits@collunits[[1]]@chronologies@chronologies[[1]]@chroncontrols
```


I could not identify the sorting of the original chronology table to follow

The tables are otherwise identical


### Outstanding Issues

Collection unit print doesn't work 

```{r error=TRUE}
D@sites[[1]]@collunits@collunits[[1]]
N$collunits@collunits[[1]]

new("site")
new("collunits")
new("collunit")
```

creating a new "collunit" is an issue?


Some of the chronology metadata isn't retained, lipd doesn't store metadata at this level

```{r error=TRUE}
D@sites[[1]]@collunits@collunits[[1]]@chronologies@chronologies[[1]]
N$collunits@collunits[[1]]@chronologies@chronologies[[1]]
```

Same for "collunit" metadata

```{r error=TRUE}
D@sites[[1]]@collunits@collunits[[1]]
N$collunits@collunits[[1]]
```