rgee has two types of dependencies: strict dependencies that must be satisfied before the rgee installation and the credentials dependencies that unlock rgee import & export functions.
If the first group of dependencies is not fulfilled rgee just will not work. The dependencies that comprised this group are:
The activation of Earth Engine accounts depends on each user, check the oficial website of Google Earth Engine for more details. If you do not count with a Python environment or a version of the EarthEngine Python API, we strongly recommend you run ee_install(py_env = "rgee"). This function will realize the following six tasks:
reticulate::py_install.However, the use of ee_install() is not mandatory. You can count on with your own custom installation. This would be also allowed.
library(rgee) ee_install() # It is just neccessary once!
By the other hand, the credentials dependencies are only needed to move data from Google Drive and Google Cloud Storage to your local environment. These dependencies are not mandatory. However, they will help you to create a seamless connection between R and Earth Engine. The dependencies that comprised this group are shown below:
The batch import/export involves difficulties for most GEE users. In rgee, we are aware of it and we created several functions to help users to download and upload spatial data. If you are trying to export data from Earth Engine, rgee offers the following options:
| FROM | TO | RETURN | ||
|---|---|---|---|---|
| Image | ee_image_to_drive | EE server | Drive | Unstarted task |
| ee_image_to_gcs | EE server | Cloud Storage | Unstarted task | |
| ee_image_to_asset | EE server | EE asset | Unstarted task | |
| ee_as_raster | EE server | Local | RasterStack object | |
| ee_as_stars | EE server | Local | Proxy-stars object | |
| Table | ee_table_to_drive | EE server | Drive | Unstarted task |
| ee_table_to_gcs | EE server | Cloud Storage | Unstarted task | |
| ee_table_to_asset | EE server | EE asset | Unstarted task | |
| ee_as_sf | EE server | Local | sf object | |
| Generic | ee_drive_to_local | Drive | Local | object filename |
| ee_gcs_to_local | Cloud Storage | Local | GCS filename |
In rgee, all the functions from server to local side have the option to fetch data using an intermediate container (Google Drive or Google Cloud Storage) or through a REST call (“$getInfo”). Although the latter option performs a quick download, there is a request limit of 262144 pixels for ee$Image and 5000 elements for ee$FeatureCollection which makes it unsuitable for large objects. Other download functions, from server-side to others (see Table 1), are implemented to enable more customized download workflows. For example, using ee_image_to_drive and ee_drive_to_local users could create scripts which save results in the .TFRecord rather than the .GeoTIFF format. rgee to deal with Google Drive and Google Cloud Storage use the R package googledrive and googleCloudStorageR respectively, so you will need to install it before.
# please try as follow install.packages('googledrive') install.packages('googleCloudStorageR')
Google Drive is more friendly to novice Earth Engine users because the authentication process could be done without leaving R. However, if you are trying to move large images or vectors, is preferable use Google Cloud Storage instead. For using Google Cloud Storage, you will need to have your own Google Project with a credit card added to the service, charges will apply. See the GCS_AUTH_FILE tutorial to create your own service account key. If you want to understand why this is necessary, please have a look at Mark Edmondson tutorial.
Batch upload is a harder process, in rgee we try to make it simple. If you want to upload files in a batch way, firstly you must get authorization to read & write into a Google Cloud Storage (GCS) bucket. rgee implement the next functions:
| FROM | TO | RETURN | ||
|---|---|---|---|---|
| Image | gcs_to_ee_image | Cloud Storage | EE asset | EE Asset ID |
| raster_as_ee | Local | EE asset | EE Asset ID | |
| stars_as_ee | Local | EE asset | EE Asset ID | |
| Table | gcs_to_ee_table | Cloud Storage | EE asset | EE Asset ID |
| sf_as_ee | Local | EE asset | EE Asset ID | |
| Generic | local_to_gcs | Local | Cloud Storage | GCS filename |
The upload process follows the same logic (Table 2). rgee includes raster_as_ee and stars_as_ee for uploading images and sf_as_ee for uploading vector data.
As we have seen previously, rgee deal with three different Google API’s:
To authenticate either Google Drive or Google Cloud Storage, you just need to run as follow:
library(rgee) #ee_reattach() # reattach ee as a reserve word # Initialize just Earth Engine ee_Initialize() ee_Initialize(email = 'csaybar@gmail.com') # Use the argument email is not mandatory # Initialize Earth Engine and GD ee_Initialize(email = 'csaybar@gmail.com', drive = TRUE) # Initialize Earth Engine and GCS ee_Initialize(email = 'csaybar@gmail.com', gcs = TRUE) # Initialize Earth Engine, GD and GCS ee_Initialize(email = 'csaybar@gmail.com', drive = TRUE, gcs = TRUE)
If the Google account is verified and the permission is granted, you will be directed to an authentication token. Copy this token and paste it in the emerging GUI. Unlike Earth Engine and Google Drive, Google Cloud Storage need to set up its credential manually (link1 and link2). In all cases, the users credentials always will be stored in:
Remember you only have to authorize once, for next sessions it will not be necessary.
The ee_check() function will help you for checking the sanity of your rgee installation and dependencies.
ee_check_python() - Python versionee_check_credentials() - Google Drive and GCS credentialsee_check_python_packages() - Python packages