---
title: "Introduction to the sarp.snowprofile package"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{sarp.snowprofile}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---


Functions to import/export various formats of snow profiles, with basic formatting and visualization functions.

```{r}
library(sarp.snowprofile)
```


## 1. Snowprofile objects

The package uses S3 classes for individual snow profiles (class `snowprofile`) or lists of multiple snow profiles (class `snowprofileSet`). Objects with these classes can be created, manipulated, and visualized.

A `snowprofile` object contains data about a snow stratigraphy profile. It is structured as a `list` with metadata (e.g. profile name and location) and a `data.frame` containing layer properties. Mandatory parts of a `snowprofile` object include:

- `station` and `station_id` provide a profile name
- `datetime` timestamp
- `latlon`, `elev`, `angle`, `aspect` location information
- `hs` total snow height
- `layers` a `data.frame` of class `snowprofileLayers` that contains layer properties (each row is a layer and each column is a property such as depth or grain size). 

A `snowprofileSet` object is simply a `list` of multiple `snowprofile` objects.

The package includes sample data packaged into three `snowprofileSet` objects:

- `SPgroup` contains 12 profiles from different locations with the same timestamp
- `SPtimeline` contains 11 profiles from the same location with different timestamps
- `SPpairs` contains various other snow profiles

```{r}
## Print the structure of a single `snowprofile` object
Profile <- SPpairs$C_day1
str(Profile)
```

Check out an empty snowprofile object to understand which information can potentially go where:
```{r}
snowprofile(dropNAs = FALSE)
```

## 2. Profile creation

The package includes functions to create a `snowprofile` object by importing common file formats as well as provides constructor functions to manually create a `snowprofile`.

### 2.1 Import profiles from file

Import functions for generic snow profiles include::

- CAAML (`snowprofileCaaml`)
- csv (`snowprofileCsv`)

and import functions for simulated profiles produced with the snow cover model [SNOWPACK](https://models.slf.ch/p/snowpack/) include:

- prf (`snowprofilePrf`)
- pro (`snowprofilePro`)
- sno (`snowprofileSno`)

Note that prf and pro files contain multiple profiles and thus the import functions return a `snowprofileSet`, while CAAML, csv, and sno files return a single `snowprofile`.

```{r}
## Import a CAAML file
# Filename <- "path/to/file.caaml"
# Profile <- snowprofileCaaml(Filename)

# ## Import all profiles from a directory of CAAML files and create a `snowprofileSet`
# CaamlFiles <- list.files('path/to/caamlprofiles', full.names = T)
# Profiles <- lapply(CaamlFiles, snowprofileCaaml)
# Profiles <- snowprofileSet(Profiles)
```

An additional parser `readSmet` is also provided to read  other input and output files from SNOWPACK.

### 2.2 Manually construct profiles

To manually create a `snowprofile` object see the help pages for the constructor functions `snowprofile()` and `snowprofileLayers()` where metadata and layer properties are provided as function arguments.

### 2.3 Formatting profiles

Import and constructor functions perform several validation checks for consistent structure (e.g. variable names, consistent layer thickness/depth/height). Sometimes profiles may be malformatted (e.g. files have different formatting than the functions in this package, future changes to this R package), so this package provides functions for checking profiles for formatting discrepancies and reformatting them if necessary. `validate_snowprofile` raises errors (or silently print error messages) in case of formatting discrepancies and `reformat_snowprofile` can conveniently correct data types or rename metadata / layer properties. See examples in the help files for these functions for examples of how it identify and correct errors in malformatted profiles 


## 3. Profile manipulation

`print`, `summary`, and `rbind` methods exist to summarize and extract contents from `snowprofile` and `snowprofileSet` objects.

### 3.1 Print

Print contents of a `snowprofile` to the console.

```{r}
Profile <- SPpairs$A_manual
print(Profile)
```

### 3.2 Summary

Extract metadata from a single profile or set of profiles into a `data.frame`.

```{r}
summary(Profile)
summary(SPgroup)
```

Summary methods are useful to extract subsets of a `snowprofileSet` based on some attribute (e.g. location, time).

```{r}
## Subset all profiles from SPgroup with elevation > 2000 m
Metadata <- summary(SPgroup)
Alpine <- SPgroup[Metadata$elev > 2000]
print(paste(length(Alpine), 'of', length(SPgroup), 'profiles in SPgroup are above 2000 m'))
```

### 3.3 Rbind

`rbind` methods merge metadata and layer properties from one or many profiles into a large `data.frame` that is convenient for analysis tasks.

```{r}
## Rbind SPgroup 
TabularProfile <- rbind(SPgroup)
names(TabularProfile)

## Tabulate all grain types
table(TabularProfile$gtype)

## Get elevations of profiles with SH layers > 5 mm
SH_layers <- subset(TabularProfile, gtype == 'SH' & gsize > 5)
sort(SH_layers$elev)
```



## 4. Profile visualization

Plot methods exist for individual profiles and sets of profiles:

- `plot.snowprofile` produces a hardness profile for a single `snowprofile` object
- `plot.snowprofileSet` produces several types of plots for a `snowprofileSet` including timelines and side-by-side stratigraphies

```{r}
## Plot a single hardness profile
plot(Profile)
## Plot a timeline
plot(SPtimeline)
## Plot all profiles in the group and sort by height
plot(SPgroup, SortMethod = 'hs')
```

The package also includes several `getColour...` and `setColour...` functions to define colour palettes for various profile properties.