obj_inspector.Rmd

---
title: "Suggested packages"
output: github_document
---

`R` has the issue that some `R` objects that don't work correctly if their package is not attached (common to `xts`, `tibble`, and `data.table`).  This is, unfortunately, easier to do in `R` than in other languages (for example `Python`'s `pickle` will attach packages).

A solution is: `saveRDS()` could be augmented to add the `suggested_packages()` as an attribute of what it writes out, say ".suggested_packages`. Then `readRDS()` could look for this attribute and issue a warning if any of them are not attached during the read.

To explore the feasibility of the above, let's experiment with finding suggested packages (without package order) for `R` objects.

First: define the suggestion function.

```{r setup}
source("find_pkgs.R")
```

Example 1: an `xts` object.

```{r}
library("tibble")
library("xts")

data(sample_matrix)
sample.xts <- as.xts(sample_matrix, descr='my new xts object')

suggested_packages(sample.xts)
```

However, notice if `data.table` is attached the advice changes.

```{r}
library("data.table")

suggested_packages(sample.xts)
```

We can ask for more details to see why this is.

```{r}

suggested_packages(sample.xts, show_details = "data.table")
```


Example 2: a `tibble`.

```{r}
d <- as_tibble(data.frame(x = 1))

suggested_packages(d)
```

Example 3: a `data.table`.

```{r}
dt <- data.table(x = 2)

suggested_packages(dt)
```

Example 4: `data.frame`

```{r}
df <- data.frame(x = 4)

suggested_packages(df)
```

Example 5: nested stuff.

```{r}
df2 <- data.frame(x = 1)
df2$y <- list(tibble(x = 5))

class(df2)

suggested_packages(df2)
```




Example 6: `tidyverse` adding packages to the outcome

```{r}
library("tidyverse") # extra packages to show interference effects

d <- as_tibble(data.frame(x = 1))

suggested_packages(d)

suggested_packages(d, show_details = c("dplyr", "ggplot2", "tidyr"))
```