In ropensci/rfishbase: R Interface to 'FishBase'

rfishbase

Welcome to rfishbase 4. This is the fourth rewrite of the original rfishbase package described in Boettiger et al. (2012).

rfishbase 1.0 relied on parsing of XML pages served directly from Fishbase.org.
rfishbase 2.0 relied on calls to a ruby-based API, fishbaseapi, that provided access to SQL snapshots of about 20 of the more popular tables in FishBase or SeaLifeBase.
rfishbase 3.0 side-stepped the API by making queries which directly downloaded compressed csv tables from a static web host. This substantially improved performance a reliability, particularly for large queries. The release largely remained backwards compatible with 2.0, and added more tables.
rfishbase 4.0 extends the static model and interface. Static tables are distributed in parquet and accessed through a provenance-based identifier. While old functions are retained, a new interface is introduced to provide easy access to all fishbase tables.

We welcome any feedback, issues or questions that users may encounter through our issues tracker on GitHub: https://github.com/ropensci/rfishbase/issues

knitr::opts_chunk$set(warning=FALSE, comment=NA)
options("rfishbase_local_db" = FALSE)

Installation

remotes::install_github("ropensci/rfishbase")

library("rfishbase")
library("dplyr") # convenient but not required

Getting started

Generic table interface

All fishbase tables can be accessed by name using the fb_tbl() function:

fb_tbl("ecosystem")

You can see all the tables using fb_tables() to see a list of all the table names (specify sealifebase if desired). Careful, there are a lot of them! The fishbase databases have grown a lot in the decades, and were not intended to be used directly by most end-users, so you may have considerable work to determine what's what. Keep in mind that many variables can be estimated in different ways (e.g. trophic level), and thus may report different values in different tables. Also note that species is name (or SpecCode) is not always the primary key for a table -- many tables are specific to stocks or even individual samples, and some tables are reference lists that are not species focused at all, but meant to be joined to other tables (faoareas, etc). Compare tables against what you see on fishbase.org, or ask on our issues forum for advice!

fish <- c("Oreochromis niloticus", "Salmo trutta")

fb_tbl("species") %>% 
  mutate(sci_name = paste(Genus, Species)) %>%
  filter(sci_name %in% fish) %>% 
  select(sci_name, FBname, Length)

SeaLifeBase

SeaLifeBase.org is maintained by the same organization and largely parallels the database structure of Fishbase. As such, almost all rfishbase functions can instead be instructed to address the

fb_tbl("species", "sealifebase")

Versions and importing all tables

By default, tables are downloaded the first time they are used. rfishbase defaults to download the latest available snapshot; be aware that the most recent snapshot may be months behind the latest data on fishbase.org. Check available releases:

available_releases()

Low-memory environments

If you have very limited RAM (e.g. <= 1 GB available) it may be helpful to use fishbase tables in remote form by setting collect = FALSE. This allows the tables to remain on disk, while the user is still able to use almost all dplyr functions (see the dbplyr vignette). Once the table is appropriately subset, the user will need to call dplyr::collect() to use generic non-dplyr functions, such as plotting commands.

fb_tbl("occurrence")

Local copy

Set the option "rfishbase_local_db" = TRUE to create a local copy, otherwise will use a remote copy. Local copy will get better performance after initial import, but may experience conflicts when duckdb is upgraded or when multiple sessions attempt to access the directory. Remove the default storage directory (given by db_dir()) after upgrading duckdb if using a local copy.

options("rfishbase_local_db" = TRUE)
db_disconnect() # close previous remote connection

conn <- fb_conn()
conn

Users can trigger a one-time download of all fishbase tables (or a list of desired tables) using fb_import(). This will ensure later use of any function can operate smoothly even when no internet connection is available. Any table already downloaded will not be re-downloaded. (Note: fb_import() also returns a remote duckdb database connection to the tables, for users who prefer to work with the remote data objects.)

fb_import()

db_disconnect(conn)

Interactive RStudio pane

RStudio users can also browse all fishbase tables interactively in the RStudio connection browser by using the function fisbase_pane(). Note that this function will first download a complete set of the fishbase tables.

Backwards compatibility

rfishbase 4.0 tries to maintain as much backwards compatibility as possible with rfishbase 3.0. Because parquet preserves native data types, some encoded types may differ from earlier versions. As before, these are not always the native type -- e.g. fishbase encodes some boolean (logical TRUE/FALSE) values as integer (-1, 0) or character types. Use as.logical() to coerce into the appropriate type in that case.

Toggling between fishbase and sealifebase servers using an environmental variable, FISHBASE_API, is now deprecated.

Note that fishbase will store downloaded files by hash in the app directory, given by db_dir(). The default location can be set by configuring the desired path in the environmental variable, FISHBASE_HOME.

Please note that this package is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.