validate_process
Argument
if not required, which
affects mostly the “context” parameter of reducer functionsArray
type if there is more
than one sub type allowedget_sample()
), which will modify a process graph by
replacing the original spatial extent of the request with a smaller
subset of 0.0003 degrees extent #94metadata-filter
in order
to filter collections based on a list of functions mostly applied in
load_collection
#102httr2
package until the httr2
package changes
are made available on CRAN
toJSON
for Process
and
Graph
processToJSON
to easily convert a process to JSONprocessToJSON
and graphToJSON
are marked
as deprecatedprocess_viewer
for user defined processes #110compute_result()
will set or override existing
save_result
node entries when parameter format
is provided #62compute_result()
will return a stars
object if the parameter as_stars
is set TRUE #39download_results()
uses httr2 functions instead of
download.file()
#108,
also parameter job
is allowed to be a ResultList or
AssetListtimeout
parameter in logs()
is now
optional. If omitted 60s timeout is used for active batch processes and
enabled services #109httr
by httr2
to include
the device_code and PKCE authentication methodsconnect
no longer carries the login parameters
separately, but uses ...
to pass on those informationlogin_type
removed. type will be
deduced based on the other parameters.EPSGCode
like EPSG:4326 #99 and
wrote a test case for thatOpenEOClient
Boolean
argument validationProcessNode
was passed as a value|
,
&
and xor
where an old signature of the
corresponding openEO process was usedBugfixes and preparation for CRAN release, ‘openEO API’ version 1.1.0
as(x,"Process")
to get a valid process from the result node #54as(x,"Graph")
#68connect()
in
terms of the parameter version
In general this release will cover the changes between API versions 0.4.2 and 1.0.0 without backward compatibility. Changes in the data models due to the new API are not listed here.
Note: similarly named functions that work on different objects are
abbreviated by a *
, e.g. describe_*()
for
describe_job()
, describe_service()
, …
Parameter
into a
descriptive formGraph
(e.g. as(x,"Graph")
, where x is a R
function returning a ProcessNode
)privacy_policy()
and
terms_of_service()
to view the privacy policy and the terms
of serviceconformance()
operated on the new
corresponding API endpointProcessCollection
in order to divide the
Graph object and the graph building. It is created upon connection and
obtainable via processes()
.logs()
, log_job()
and
log_service()
to request the back-end logs of a job or a
service. The logging blocks the console and requests every second an
update of the log at the back-end. Also when queuing a job via
start_job()
an immediate logging can be triggered.FileFormatList
(as.data.frame.FileFormaList
)FileFormat
(print.FileFormat
)ServiceType
(print.ServiceType
)ServiceList
(as.data.frame.ServiceList
,
print.ServiceList
)Service
(print.Service
)CollectionList
Collection
(print.Collection
)Log
(print.Log
)Job
(print.Job
)JobList
(print.JobList
)ResultList
(print.ResultList
)AssetList
(as.data.frame.AssetList
)ProcessInfo
(as.Graph.ProcessInfo
)UdfRuntime
(as.character.UdfRuntime
)UdfRuntimeList
CubeDimension
(as.character.CubeDimension
)CubeDimensions
(print.CubeDimensions
)ProcessInfo
(as.Process.ProcessInfo
)UdfRuntimeArgument
,
UdfRuntimeVersionArgument
,
UdfCodeArgument
ArgumentList
to
Process
in order to allow an easier value assignment,
e.g. node1$parameter$x = ...
, which comes in handy, when
the node is assigned to a variable and a parameter has to changed
laterUserProcessCollection
and user_processes()
)
like ProcessCollection
for predefined processes, which
offers a users stored process graph as a usable function that creates a
ProcessNode
status
for the connection, jobs and
services to immediate get the status information. It also always queries
the back-end for an update on calltoJSON
(e.g. especially when
stating the number of digits to use during this serialization)Parameter
parent class
in order to manipulate parameter representations later, when defining
Parameters in user defined processesProcessGraph
as a parameter requires a function
wrapping - meaning you should not pass an openEO process like
p$min
directly, because with the elaborated
ProcessGraphParameter
we cannot do a reliable automatic
variable matching, e.g. often the ProcessGraphParameter
“context” is misinterpreted*_viewer()
functions to visualize collection and processes following the new api
definitionanyOf
is still be used in the
client.Callback
was renamed into
ProcessGraphArgument
(Graph is still a different Concept),
which is used in reducer or aggregation functionscallback-value
was renamed into
ProcessGraphParameter
ProcessGraphParameter
are derived automatically
from R functions that are passed as ProcessGraphArgument
(former Callback
) and they become parameters that shall be
set by a user upon use of this user defined graphprint
functionsupdate_*
and create_*
functions return
their respective object (e.g. update_job
–>
Job
) instead of their ID (previous behavior)list_*
functions return a named list of those
particular exchange objects (e.g. list_jobs
–> named
list of Job
)Graph
object does not hold a reference about the
available data sets any longerupdate_process_graph
and
merged with the new information given by the function parameterlist_file_formats()
to
list_file_formats()
list_process_graphs()
to
list_user_processes()
describe_process_graph()
to
describe_user_process()
delete_process_graph()
to
delete_user_process()
create_process_graph()
to
create_user_process()
update_process_graph()
to
update_user_process()
validate_process_graph()
to
validate_process()
FileFormat
object can passed to
OutputFormat
parameterlist_*
and describe_*
can be passed as parameter*Info
class names are reduced to the respective
shorter class name without “Info”, except for
ProcessInfo
list_*
and describe_*
are labeled with the same class and printed
with the information available (either overview or detailed view)GeoJson
argumentlog
at
start_job()
follow_job()
(replaced by
log_job()
)connection
which linked to the used
openEO client from the process graph objectprocess_graph_builder()
callback()
, because those Process
Graphs are either created from an R function or an resulting
ProcessNode
user_id
as a parameter from the files request,
because the user information is send with the bearer tokensend_udf
can now handle user_contextsend_udf
has now a parameter for legacy support (which
calls the legacy endpoint of a UDF service with API version 0.1.0)run_udf
openEO process) [#48]commonmark
process_viewer
spawns now a viewer
in RStudio and renders a beautiful website containing the process
informationprocesses
that will create an object that has
the processes that are available on the connected openEO service. It is
quite similar to the return value of process_graph_builder
which will create a Graph
object with basically the same
functionality plus the graph related functionssend_udf
and
added documentation to the functionGraph
object
explicitly, passing the end node of the defined workflow is now
possible.active_connection
now gets and sets the latest connected
service into a package variablepi
and e
(in R
pi
and exp(1)
) are parsed from the data and
translated into the corresponding openEO processes in order to avoid
errors due to rounded valuesProcessNode
and
callback-value
classesmarkdownViewer
function in favor of
process_viewer
and collection_viewer
band_arithmetic
, because callbacks
can now be functionsconnect
with new parameters
exchange_token
and external
. With the latter
we can specify ‘google’ as an external identity provider and with
exchange_token
we can specify if we send the default
access_token
or the id_token
as Bearer token
in the authentication headerOpenEOClient$isLoggedIn()
) and made it
optional at the following endpoints to sent a Bearer token as
Authorization-header depending on whether the user is logged in:GET /output_formats
GET /udf_runtimes
GET /service_types
GET /collections
GET /collections/{collection_id}
GET /processes
POST /validation
gee-evi-example
now that the GEE driver
supports all relevant openEO processes, it now creates the graph with
apply, linear_scale and saves it as PNG; added also an example for the
band_arithmetic use caseclient_version
was renamed into
api_version
and returns the openEO API version which the
client usescreate_job
now refers to the HTTP response header
“openEO-Identifier” rather than parsing the “Location” headerband_arithmetic
where the parameter
in array_element were falsely assigneddownload_results
list_file_types
compute_result
)