Guidelines and processes to run OGC API - Processes / CWL / EOAP

CHANGELOG.md

@@ -6,6 +6,12 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## Unreleased / Draft
 
+### Added
+
+- `run_ogcapi`
+- `run_ogcapi_processes`
+- Implementation guide for implementing OGC API - Processes in openEO
+
 ### Changed
 
 - `clip`: Throw an exception if min > max [#472](https://github.com/Open-EO/openeo-processes/issues/472)

dev/.words

@@ -48,3 +48,4 @@ Hyndman
 date1
 date2
 favor
+Undeploy

meta/implementation.md

@@ -230,3 +230,64 @@ We have found some libraries that can be used for an implementation:
 - Julia: [Statistics.quantile](https://docs.julialang.org/en/v1/stdlib/Statistics/#Statistics.quantile!), type 7 is the default.
 - Python: [numpy](https://numpy.org/doc/stable/reference/generated/numpy.quantile.html), [pandas](https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.quantile.html), [xarray](http://xarray.pydata.org/en/stable/generated/xarray.DataArray.quantile.html) - type 7 (called 'linear' for the interpolation parameter) is the default for all of them.
 - R: [quantile](https://stat.ethz.ch/R-manual/R-patched/library/stats/html/quantile.html) - type 7 is the default.
+
+## OGC API - Processes
+
+OGC API - Processes and OGC EO Application Packages (AP) can generally be utilized in openEO in three different ways:
+
+1. **openEO process**
+
+   As a pre-defined process that exposes itself as a normal openEO process.
+   It is not exposed to the user that in the background an AP is invoked.
+2. **Pre-deployment through *OGC API - Processes - Part 2: Deploy, Replace, Undeploy***
+
+   In addition to the openEO API, a provider can offer access to an instance of *OGC API - Processes - Part 2: Deploy, Replace, Undeploy* (OGC DRU).
+   The OGC DRU instance is likely external to the openEO API tree due to the conflicting `GET /processes` endpoint.
+   As such the OGC DRU instance exposes itself in the `GET /` endpoint of the openEO API instance through a link.
+   The link must have the relation type `http://www.opengis.net/def/rel/ogc/1.0/processes`, which points to the `/processes` of the OGC API - Processes endpoint.
+   Users can deploy APs through the OGC DRU instance and use them through the process `run_ogcapi`.
+
+   If the provider doesn't offer an OGC DRU instance itself, users could also deploy their AP with another provider.
+   In this case use the process `run_ogcapi_externally` instead.
+
+   Example process node:
+
+   ```json
+   {
+     "process_id": "run_ogcapi",
+     "arguments": {
+       "data": ..., // Data, e.g. subtypes datacube or stac
+       "id": "my-ap", // Identifier of the application package in OGC API - Processes
+       "context": { // Parameters as defined in the CWL file
+         "cwl_param1": true,
+         "param2": 99
+       }
+     }
+   }
+   ```
+3. **CWL provided at runtime (UDF runtime)**
+
+   Providers can also provide a UDF runtime for the language CWL (instead of e.g. Python or R).
+   The runtime is exposed through the endpoint `GET /udf_runtimes`.
+
+   Example process node:
+
+   ```json
+   {
+     "process_id": "run_udf",
+     "arguments": {
+       "data": ..., // Data, e.g. subtypes datacube or stac
+       "udf": "...", // CWL as YAML string/JSON object, URL, or file on the API user workspace
+       "runtime": "cwl", // Assuming the UDF runtime is named "cwl"
+       "context": { // Parameters as defined in the CWL file
+         "cwl_param1": true,
+         "param2": 99
+       }
+     }
+   }
+   ```
+
+Generally, we recommend to use the following types and formats for the CWL inputs and outputs:
+
+- `type`: `File` or `File[]` depending on the capabilities of the CWL workflow
+- `format`: For STAC inputs and outputs either `stac:item`, `stac:collection`, `stac:catalog`, `stac:itemcollection`, or `stac:stac` (i.e. and of the types before).

proposals/run_ogcapi.json

@@ -0,0 +1,47 @@
+{
+    "id": "run_ogcapi",
+    "summary": "Run a OGC API process",
+    "description": "Runs an OGC API - Processes process that the service provider offers through an instance of OGC API - Processes - Part 2 (Deploy, Replace, Undeploy).",
+    "categories": [
+        "cubes",
+        "import",
+        "udf"
+    ],
+    "experimental": true,
+    "parameters": [
+        {
+            "name": "data",
+            "description": "The data to be passed to the OGC API process.\n\nThe data must be given in a way that the external process can understand it.",
+            "schema": {
+                "description": "A value of any data type."
+            }
+        },
+        {
+            "name": "id",
+            "description": "The identifier of the OGC API process.",
+            "schema": {
+                "type": "string"
+            }
+        },
+        {
+            "name": "context",
+            "description": "Additional parameters as defined by the OGC API process.",
+            "schema": {
+                "type": "object"
+            },
+            "default": {},
+            "optional": true
+        }
+    ],
+    "exceptions": {
+        "InvalidOgcApiProcess": {
+            "message": "The specified OGC API process identifier does not exist."
+        }
+    },
+    "returns": {
+        "description": "The data processed by the OGC API process. The returned value can be of any data type and is exactly what the OGC API process returns.",
+        "schema": {
+            "description": "Any data type."
+        }
+    }
+}

proposals/run_ogcapi_externally.json

@@ -0,0 +1,57 @@
+{
+    "id": "run_ogcapi_externally",
+    "summary": "Run an externally hosted OGC API process",
+    "description": "Runs an OGC API - Processes process that is either externally hosted by a service provider or running on a local machine of the user.",
+    "categories": [
+        "cubes",
+        "import",
+        "udf"
+    ],
+    "experimental": true,
+    "parameters": [
+        {
+            "name": "data",
+            "description": "The data to be passed to the external process.\n\nThe data must be given in a way that the external process can understand it.",
+            "schema": {
+                "description": "A value of any data type."
+            }
+        },
+        {
+            "name": "url",
+            "description": "Absolute URL to the OGC API - Processes landing page.",
+            "schema": {
+                "type": "string",
+                "format": "uri",
+                "subtype": "uri",
+                "pattern": "^https?://"
+            }
+        },
+        {
+            "name": "id",
+            "description": "The identifier of the OGC API process.",
+            "schema": {
+                "type": "string"
+            }
+        },
+        {
+            "name": "context",
+            "description": "Additional parameters as defined by the OGC API process.",
+            "schema": {
+                "type": "object"
+            },
+            "default": {},
+            "optional": true
+        }
+    ],
+    "exceptions": {
+        "InvalidOgcApiProcess": {
+            "message": "The specified OGC API process does not exist."
+        }
+    },
+    "returns": {
+        "description": "The data processed by the OGC API process. The returned value can be of any data type and is exactly what the OGC API process returns.",
+        "schema": {
+            "description": "Any data type."
+        }
+    }
+}