How-To add new datasets

How-To add new datasets#

The EUREC\(^4\)A-Intake catalog is the dataset collection of the EUREC\(^4\)A and ATOMIC field campaign. It is a collection of yaml-files that contain references to the dataset storage locations.

Datasets should be added by following these steps:

via the command line#

Cloning the EUREC\(^4\)A-Intake catalog repository with
```
git clone git@github.com:eurec4a/eurec4a-intake.git
```
You might be asked to set up an SSH key for your GitHub account.
Create a new branch
```
git checkout -b <my_new_dataset>
```
Add new catalog entry

The catalog contains two types of entries:
- references to sub-catalogs
- references to a dataset
The surface flux dataset from the research vessel Meteor is accessible via
```
cat.Meteor.surface_fluxes
```
and is saved in the file Meteor/main.yaml.

The radar dataset, which is more complex and contains several subsets is accessible via
```
cat.Meteor.LIMRAD94.low_res
cat.Meteor.LIMRAD94.high_res
```
For the creation of the LIMRAD94 radar subset, a sub-catalog reference has been created in Meteor/main.yaml. The final reference to the dataset is added in Meteor/LIMRAD94.yaml

Depending on the complexity of the dataset, an entry can be directly added to the main.yaml file of the respective platform/simulation
```
vi <platform>/main.yaml
```
or to a (new) instrument specific file
```
vi <platform>/<instrument>.yaml
```
The reference has the following format:
```
sources:
  <dataset_name>:
    description: <short description of the dataset>
    driver: opendap
    args:
      auth: null
      urlpath: <url_to_dataset>
      chunks: {}
      engine: netcdf4
```
In case your dataset has been published on AERIS, the THREDDS link to your dataset can be determined by finding your dataset at https://observations.ipsl.fr/aeris/eurec4a-data/. To retrieve the THREDDS link, replace https://observations.ipsl.fr/aeris/eurec4a-data/ with https://observations.ipsl.fr/thredds/dodsC/EUREC4A/. You can check if the link is correct and the data correctly formatted by opening it e.g. directly with xarray.open_dataset() or Panoply by opening a Remote Dataset.

A sub-catalog reference can be created with
```
sources:
  <instrument/model_name>:
    args:
      path: "{{CATALOG_DIR}}/<instrument/model_name>.yaml"
    description: '<Instrument/model description>'
    driver: yaml_file_cat
    metadata: {}
```
Please also check out entries already present in the EUREC\(^4\)A-Intake catalog.

Finally those changes need to be staged and committed.
```
git add -p
```
```
git commit -m "<Adding my_new_dataset>"
```

Push branch to GitHub

git push --set-upstream origin <my_new_dataset>

Create pull request on GitHub A pull request can be started on the GitHub webpage. After the pull request has been submitted, the review process will start. To accelerate the process, please make sure all tests for your pull request succeed. The status of the tests are shown at the bottom of your pull request.

via GitHub web interface#

Visit EUREC\(^4\)A-Intake catalog repository Go to eurec4a/eurec4a-intake
Login with your GitHub credentials
Select the platform/simulation/product for which a new dataset entry shall be added

Edit main.yaml and add a reference to the dataset if it is simple and does not contain different subsets (e.g. resolutions, frequencies, sensors, dimensions):

sources:
  <dataset_name>:
    description: <short description of the dataset>
    driver: opendap
    args:
      auth: null
      urlpath: <url_to_dataset>
      chunks: {}
      engine: netcdf4

(see 3. of via command line for more details

Save changes and create pull request
Check if the automatic tests that are run on your edits succeed. You can access the tests e.g. via the Action Tab ( ).
A reviewer will look at the pull request and will discuss any additional steps necessary to merge the changes to the main repository.

How-To add new datasets

Contents

How-To add new datasets#

via the command line#

via GitHub web interface#