06_ModelOutputs.qmd

---
title: "Model Outputs"
output:
  html_document:
    theme: journal
---

```{css, echo=F}
pre {
  overflow-x: scroll
}


pre code {
  white-space: pre;
}
```

```{r setup, include=FALSE}
knitr::opts_knit$set(root.dir=here::here())
knitr::opts_chunk$set(warning=F, message=F, eval=F)
```

# Main Objectives

- Understand the components of a jjm output object
- Be able to make individual plots using the `jjmR` package
- Be able to make custom plots using jjm model outputs

# Output Files

Once you've successfully run the assessment model for `h1` and `h2`, these are the files you should have (assuming you selected the `pdf=TRUE` option in the `runit()` function).

```{r}
- h1_1.00_1_R.rep # Named report file that is read directly as a list into R
- h1_1.00.cor # Parameter correlation matrix
- h1_1.00.par # Best estimates of the parameters within the model
- h1_1.00.pdf # PDF of plots
- h1_1.00.rep # Predicted quantities
- h1_1.00.std # Estimated parameter values and standard deviations
- h1_1.00.yld # Profile of stock, yield, and recruitment over F

---

- h2_1.00_1_R.rep
- h2_1.00_2_R.rep # Note second R.rep file for the second stock
- h2_1.00.cor
- h2_1.00.par
- h2_1.00.pdf
- h2_1.00.rep
- h2_1.00.std
- h2_1.00.yld
```

The `readJJM()` function only requires a subset of these files to work:

```
- R.rep
- .par
- .rep
- .std
- .yld
```

# Reading the Output into R

- We have already most of this code for this in the Input Files doc.

```{r, eval=F}
library(jjmR)

h1.mod <- readJJM("h1_1.00", path = "config", input = "input")

h2.mod <- readJJM("h2_1.00",path = "config", input = "input")

names(h1.mod[[1]]$output[[1]])
```

- These names correspond to values in the TPL file, so you can search that if you get stuck.

- JC Quiroz wrote a script that pulls out quantities to make the tables in the technical annex (`jjm/assessment/R/constructor_input_tables.R`)

## Things to Note

- One of the most important tables (in my opinion) in the output object is `msy_mt`. It contains a matrix with the time series of derived quantities and estimated reference points.
    - Note to self: Set colnames in jjmR output function for this table

```{r, eval=F}
# // Yr Fspr 1-Fspr F/Fmsy Fmsy F Fsprmsy MSY MSYL Bmsy Bzero SSB B/Bmsy

h1.mod[[1]]$output[[1]]$msy_mt
```


- `FW_*` denotes the Francis Weights for each fleet.

```{r}
$FW_N_Chile
0.915323
$FW_SC_Chile_PS
0.910934
$FW_Offshore_Trawl
0.681234
$FW_Chile_AcousCS
0.92536
$FW_Chile_AcousN
0.91344
$FW_DEPM
1.07482
```

- `SSB_fut_1` to `SSB_fut_5` denotes future spawning stock biomass projected under five different fishing scenarios (more on this later).

# Making Plots

- The PDF generated by `runit()` generally contains all the plots 

- The `diagnostics` function generates plots from the jjm object.

```{r, eval=F}
# Setting plot=T will create the entire suite of plots
h1.diag <- diagnostics(h1.mod, plot=F)

h2.diag <- diagnostics(h2.mod, plot=F)

#You can then plot individual plots by changing the `var` variable
plot(h1.diag, var = "fishedUnfishedBiomass")

plot(h2.diag, var = "summarySheet")

```

- You can find the different `var` options deep within the jjmR [code](https://github.com/SPRFMO/jjmr/blob/master/R/jjm.diag-internal.R) 
    - Might be worth documenting separately?

- Examples can also be found in `jjm/assessment/R/AnnexPlots.R`


# Exercises
- Pick two plots from the technical annex and make them from scratch
    - Find the variables using `names(h1.mod[[1]]$output[[1]])`
    - Do this for `h1` and `h2`
    - Bonus if you can graph them creatively

- Think of suggestions for other diagnostic plots we can use?