---
title: "Plotting the fitted models"
author: "Umut Caglar, Claus O. Wilke"
date: "`r Sys.Date()`"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Plotting the fitted models}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---
  
```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```
  
The **sicegar** package provides a helpful function that can generate plots of the fitted models. This function is built on top of the **ggplot2** plot library.
  
```{r install packages, echo=FALSE, warning=FALSE, results='hide',message=FALSE}

###*****************************
# INITIAL COMMANDS TO RESET THE SYSTEM
seedNo=14159
set.seed(seedNo)
###*****************************

###*****************************
require("sicegar")
require("dplyr")
require("ggplot2")
require("cowplot")
###*****************************
```

To demonstrate figure generation, we first generate simulated sigmoidal and double-sigmoidal data and fit the respective models to these datasets.

```{r}
# simulate sigmoidal data
time <- seq(3, 24, 0.5)

noise_parameter <- 0.1
intensity_noise <- runif(n = length(time), min = 0, max = 1) * noise_parameter
intensity <- sigmoidalFitFormula(time, maximum = 4, slope = 1, midPoint = 8)
intensity <- intensity + intensity_noise
dataInputSigmoidal <- data.frame(intensity = intensity, time = time)

# simulate double-sigmoidal data
noise_parameter <- 0.2
intensity_noise <- runif(n = length(time),min = 0,max = 1) * noise_parameter
intensity <- doublesigmoidalFitFormula(time,
                                    finalAsymptoteIntensityRatio = .3,
                                    maximum = 4,
                                    slope1 = 1,
                                    midPoint1Param = 7,
                                    slope2 = 1,
                                    midPointDistanceParam = 8)
intensity <- intensity + intensity_noise
dataInputDoubleSigmoidal <- data.frame(intensity = intensity, time = time)

# fit models to both datasets
fitObj_sm <- fitAndCategorize(dataInput = dataInputSigmoidal)
fitObj_dsm <- fitAndCategorize(dataInput = dataInputDoubleSigmoidal)
```

Now we can plot the results using the function `figureModelCurves()`. 
The default function assumes that the lower asymptote, $h_0$, (when x is at negative infinity) is zero, but the argument `use_h0 = TRUE` can be used to estimate the lower asymptote.
This function returns a **ggplot2** plot that can be saved or displayed directly. 
The function has several different options.

First, we can plot only the raw input data.
```{r plot raw data, echo=TRUE, fig.height=4, fig.width=6}
# sigmoidal raw data only
figureModelCurves(dataInput = fitObj_sm$normalizedInput)

# double-sigmoidal raw data only
figureModelCurves(dataInput = fitObj_dsm$normalizedInput)
```

Second, we can plot the input data with the fitted lines.
```{r plot raw data and fit, echo=TRUE, message=FALSE, warning=FALSE, comment=FALSE, fig.height=4, fig.width=6}
# sigmoidal fit
figureModelCurves(dataInput = fitObj_sm$normalizedInput,
                  sigmoidalFitVector = fitObj_sm$sigmoidalModel)

# double-sigmoidal fit
figureModelCurves(dataInput = fitObj_dsm$normalizedInput,
                  doubleSigmoidalFitVector = fitObj_dsm$doubleSigmoidalModel)
```

Third, we can additionally visualize the parameter estimates, by setting `showParameterRelatedLines = TRUE`.
```{r plot raw data and fit with parameter related lines, echo=TRUE, message=FALSE, warning=FALSE, comment=FALSE, fig.height=4, fig.width=6}
# sigmoidal fit with parameter related lines
figureModelCurves(dataInput = fitObj_sm$normalizedInput,
                  sigmoidalFitVector = fitObj_sm$sigmoidalModel,
                  showParameterRelatedLines = TRUE)

# double-sigmoidal fit with parameter related lines
figureModelCurves(dataInput = fitObj_dsm$normalizedInput,
                  doubleSigmoidalFitVector = fitObj_dsm$doubleSigmoidalModel,
                  showParameterRelatedLines = TRUE)
```


Note that the last example only works for models that had additional parameters calculated using `parameterCalculation()`. This is done automatically when fitting with `fitFunction()`, but needs to be done manually when fitting with `multipleFitFunction()`.