This data release corresponds to the following 2022 data release: "Predictions of lake water temperatures for eight reservoirs in Missouri US, 1980-2021" (doi:10.5066/P9FY8JQC). The modeling to support this data release is contained in lake-temperature-missouri-models.

code

Need to have non-CRAN packagesdssecrets, meddle, and scipiper installed, among other packages. Need to have CRAN package sbtools installed

• Create a data release or create a child item on sciencebase to experiment on
• Add "write" permissions on the release item for cidamanager (this is the dssecrets service account)
• Change the files and the functions in src/ to what you need
• Edit data release information in in_text/text_data_release.yml to fit your data release and your file names and contents
• modify the sciencebase identifier to your parent data release identifier (should be a string that is something like "5faaac68d34eb413d5df1f22")
• run scmake()
• validate your out_xml/fgdc_metadata.xml file with the validator tool
• fix any validation errors (usually this requires filling in metadata information in the in_text/text_data_release.yml and perhaps looking a the metadata template)
• win

remake.yml details

This slim template is designed to keep everything in a single remake yaml. So all data munging, manipulation, and file writing happens there, in addition to the sciencebase uploads.

This is the single push to sciencebase, it does the xml (metadata) and data at the same time. Because the upload step uses an internal task table, you can specify all files that should be pushed to the same sbid at one time. Again, because the upload step uses an internal task table, data files aren't replaced everytime you fix a metadata typo or add information to a metadata field. The result of the sciencebase push step is a file with timestamps for when each file got pushed that can be checked into GitHub. Having the file with timestamps in GitHub will clearly show when updates were made and will render nicely without having to build the object target locally. ```yaml targets: all: depends: - log/sbpostedfiles.csv

```

Read in (st_read) and manipulate data here. extract_feature() is from the meddle package and builds a list of structured spatial information for use in metadata files.

```yaml sfspatialdata: command: stread('exampledata/exampleshapefile/exampleshapefile.shp') depends: - exampledata/exampleshapefile/exampleshapefile.dbf - exampledata/exampleshapefile/exampleshapefile.prj - exampledata/exampleshapefile/example_shapefile.shx

spatialmetadata: command: extractfeature(sfspatialdata) write final metadata files as you want them to appear in the data release. yaml outdata/cars.csv: command: file.copy(from = "exampledata/examplecars.csv", to = targetname, overwrite = TRUE)

outdata/spatial.zip: command: sftozip(zipfilename = targetname, sfobject = sfspatialdata, layername = I('spatialdata'))

outxml/fgdcmetadata.xml: command: render(filename = targetname, "intext/textdatarelease.yml", spatial_metadata) ```

Push the files to sciencebase using the sb_replace_files function from src/sb_utils.R of this repo template. If you are uploading many files at once using sb_replace_files (either in a file hash or just multiple files passed in through ...), it is recommended to use a task table to do so. Internal task table methods are enabled by default. If you do not want to use an internal task table, set the use_task_table = FALSE when using sb_replace_files. You must also specify the file(s) where the sb_replace_files functions exist using the argument, sources. You can specify multiple files and a file hash file in one call to sb_replace_files. Currently, each sb_replace_files function can only push to one sbid.

yaml log/sb_posted_files.csv: command: sb_replace_files(filename = target_name, sb_id = I('5faaac68d34eb413d5df1f22'), "out_data/spatial.zip", "out_data/cars.csv", "out_xml/fgdc_metadata.xml", sources = "src/sb_utils.R")

Acknowledgements

This repository was derived from the slim-data-release-template