jnb_open_link to workflow ogc api record

Author: TejasMorbagal

deep_code/constants.py

@@ -32,3 +32,10 @@
 DEEPESDL_GIT_PULL_BASE = (
     "https://deep.earthsystemdatalab.net/hub/user-redirect/git-pull"
 )
+APPLICATION_TYPE_JUPYTER_SPEC = (
+    "https://raw.githubusercontent.com/EOEPCA/metadata"
+    "-profile/refs/heads/1.0/schemas/application-type-jupyter-notebook"
+)
+APPLICATION_STAC_EXTENSION_SPEC = (
+    "https://stac-extensions.github.io/application/v0.1.0/schema.json"
+)

deep_code/tools/publish.py

@@ -219,6 +219,7 @@ def publish_dataset(self, write_to_file: bool = False):
         osc_region = self.dataset_config.get("osc_region")
         osc_themes = self.dataset_config.get("osc_themes")
         cf_params = self.dataset_config.get("cf_parameter")
+        license_type = self.dataset_config.get("license_type")
 
         if not dataset_id or not self.collection_id:
             raise ValueError("Dataset ID or Collection ID missing in the config.")
@@ -230,6 +231,7 @@ def publish_dataset(self, write_to_file: bool = False):
             collection_id=self.collection_id,
             workflow_id=self.workflow_id,
             workflow_title=self.workflow_title,
+            license_type=license_type,
             documentation_link=documentation_link,
             access_link=access_link,
             osc_status=dataset_status,
@@ -314,7 +316,7 @@ def _update_base_catalog(
 
         return base_catalog
 
-    def publish_workflow_experiment(self, write_to_file: bool = False):
+    def generate_workflow_experiment_records(self, write_to_file: bool = False):
         """prepare workflow and experiment as ogc api record to publish it to the
         specified GitHub repository."""
         workflow_id = self._normalize_name(self.workflow_config.get("workflow_id"))
@@ -339,12 +341,16 @@ def publish_workflow_experiment(self, write_to_file: bool = False):
         application_link = link_builder.build_link_to_jnb(
             self.workflow_title, jupyter_notebook_url
         )
+        jnb_open_link = link_builder.make_related_link_for_opening_jnb_from_github(
+            jupyter_notebook_url=jupyter_notebook_url
+        )
+
         workflow_record = WorkflowAsOgcRecord(
             id=workflow_id,
             type="Feature",
             title=self.workflow_title,
             properties=wf_record_properties,
-            links=links + theme_links + application_link,
+            links=links + theme_links + application_link + jnb_open_link,
             jupyter_notebook_url=jupyter_notebook_url,
             themes=osc_themes,
         )
@@ -354,6 +360,7 @@ def publish_workflow_experiment(self, write_to_file: bool = False):
             del workflow_dict["jupyter_notebook_url"]
         if "osc_workflow" in workflow_dict["properties"]:
             del workflow_dict["properties"]["osc_workflow"]
+        # add workflow record to file_dict
         wf_file_path = f"workflows/{workflow_id}/record.json"
         file_dict = {wf_file_path: workflow_dict}
 
@@ -380,6 +387,7 @@ def publish_workflow_experiment(self, write_to_file: bool = False):
             del experiment_dict["collection_id"]
         if "osc:project" in experiment_dict["properties"]:
             del experiment_dict["properties"]["osc:project"]
+        # add experiment record to file_dict
         exp_file_path = f"experiments/{workflow_id}/record.json"
         file_dict[exp_file_path] = experiment_dict
 
@@ -406,7 +414,9 @@ def publish_all(self, write_to_file: bool = False):
         """Publish both dataset and workflow/experiment in a single PR."""
         # Get file dictionaries from both methods
         dataset_files = self.publish_dataset(write_to_file=write_to_file)
-        workflow_files = self.publish_workflow_experiment(write_to_file=write_to_file)
+        workflow_files = self.generate_workflow_experiment_records(
+            write_to_file=write_to_file
+        )
 
         # Combine the file dictionaries
         combined_files = {**dataset_files, **workflow_files}

deep_code/utils/ogc_api_record.py

@@ -1,10 +1,12 @@
-from typing import Any, Optional
-from urllib.parse import quote, urlencode
+from typing import Any, Optional, Tuple
+from urllib.parse import quote, urlencode, urlparse
 
 from xrlint.util.constructible import MappingConstructible
 from xrlint.util.serializable import JsonSerializable, JsonValue
 
 from deep_code.constants import (
+    APPLICATION_STAC_EXTENSION_SPEC,
+    APPLICATION_TYPE_JUPYTER_SPEC,
     BASE_URL_OSC,
     DEEPESDL_GIT_PULL_BASE,
     OGC_API_RECORD_SPEC,
@@ -141,36 +143,80 @@ def build_link_to_jnb(self, workflow_title, jupyter_nb_url):
             }
         ]
 
-    def build_deepesdl_notebook_href(
-        repo_url: str,
-        notebook_path: str,
-        branch: str = "main",
+    @staticmethod
+    def _parse_github_notebook_url(url: str) -> Tuple[str, str, str, str]:
+        """
+        Returns (repo_url, repo_name, branch, file_path_in_repo) from a GitHub URL.
+
+        Supports:
+          - https://github.com/<owner>/<repo>/blob/<branch>/<path/to/notebook.ipynb>
+          - https://raw.githubusercontent.com/<owner>/<repo>/<branch>/<path/to/notebook.ipynb>
+        """
+        p = urlparse(url)
+        parts = p.path.strip("/").split("/")
+
+        if p.netloc == "github.com":
+            if len(parts) >= 5 and parts[2] in ("blob", "tree"):
+                owner, repo, _blob_or_tree, branch = parts[:4]
+                file_path = "/".join(parts[4:])
+            else:
+                raise ValueError(f"Unexpected GitHub URL format: {url}")
+            repo_url = f"https://github.com/{owner}/{repo}"
+            repo_name = repo
+
+        elif p.netloc == "raw.githubusercontent.com":
+            if len(parts) >= 4:
+                owner, repo, branch = parts[:3]
+                file_path = "/".join(parts[3:])
+            else:
+                raise ValueError(f"Unexpected raw.githubusercontent URL format: {url}")
+            repo_url = f"https://github.com/{owner}/{repo}"
+            repo_name = repo
+
+        else:
+            raise ValueError(f"Only GitHub URLs are supported: {url}")
+
+        return repo_url, repo_name, branch, file_path
+
+    def build_deepesdl_notebook_href_from_github(
+        self,
+        jupyter_notebook_url: str,
         base_redirect: str = DEEPESDL_GIT_PULL_BASE,
+        branch_override: str | None = None,
     ) -> str:
         """
-        Build a DeepESDL git-pull redirect URL:
-        {base}?repo=<encoded_repo>&urlpath=<encoded_lab_tree_path>&branch=<encoded_branch>
+        Build DeepESDL git-pull redirect from a full GitHub notebook URL.
+        {base}?repo=<repo_url>&urlpath=lab/tree/<repo_name>/<path>&branch=<branch>
         """
+        repo_url, repo_name, branch, file_path = self._parse_github_notebook_url(
+            jupyter_notebook_url
+        )
+        if branch_override:
+            branch = branch_override
+
         params = {
             "repo": repo_url,
-            "urlpath": f"lab/tree/{notebook_path.lstrip('/')}",
+            "urlpath": f"lab/tree/{repo_name}/{file_path}",
             "branch": branch,
         }
         return f"{base_redirect}?{urlencode(params, quote_via=quote)}"
 
-    def make_related_link_for_opening_jnb(
+    def make_related_link_for_opening_jnb_from_github(
         self,
-        repo_url: str,
-        notebook_path: str,
-        branch: str = "main",
+        jupyter_notebook_url: str,
         title: str = "Open notebook on the DeepESDL platform",
+        branch_override: str | None = None,
     ) -> dict[str, str]:
-        return {
-            "rel": "related",
-            "href": self.build_deepesdl_notebook_href(repo_url, notebook_path, branch),
-            "type": "text/html",
-            "title": title,
-        }
+        return [
+            {
+                "rel": "related",
+                "href": self.build_deepesdl_notebook_href_from_github(
+                    jupyter_notebook_url, branch_override=branch_override
+                ),
+                "type": "text/html",
+                "title": title,
+            }
+        ]
 
 
 class WorkflowAsOgcRecord(MappingConstructible["OgcRecord"], JsonSerializable):
@@ -188,7 +234,11 @@ def __init__(
         themes: Optional[Any] = None,
     ):
         if conformsTo is None:
-            conformsTo = [OGC_API_RECORD_SPEC]
+            conformsTo = [
+                OGC_API_RECORD_SPEC,
+                APPLICATION_TYPE_JUPYTER_SPEC,
+                APPLICATION_STAC_EXTENSION_SPEC,
+            ]
         self.id = id
         self.type = type
         self.title = title
@@ -227,6 +277,14 @@ def _generate_static_links(self):
                 "title": "Jupyter Notebook",
                 "href": f"{self.jupyter_notebook_url}",
             },
+            {
+                "rel": "application-originating-platform",
+                "title": "DeepESDL platform",
+                "href": "https://deep.earthsystemdatalab.net/",
+                "type": "text/html",
+                "application:platform_supports": ["jupyter-notebook"],
+                "application:preferred_app": "JupyterLab",
+            },
             {
                 "rel": "related",
                 "href": f"../../projects/{PROJECT_COLLECTION_NAME}/collection.json",
@@ -258,7 +316,11 @@ def __init__(
         if linkTemplates is None:
             linkTemplates = []
         if conformsTo is None:
-            conformsTo = [OGC_API_RECORD_SPEC]
+            conformsTo = [
+                OGC_API_RECORD_SPEC,
+                APPLICATION_TYPE_JUPYTER_SPEC,
+                APPLICATION_STAC_EXTENSION_SPEC,
+            ]
         self.id = id
         self.title = title
         self.type = type
@@ -291,18 +353,20 @@ def _generate_static_links(self):
                 "type": "application/json",
                 "title": f"Workflow: {self.title}",
             },
-            # {
-            #     "rel": "child",
-            #     "href": f"../../products/{self.collection_id}/collection.json",
-            #     "type": "application/json",
-            #     "title": f"{self.collection_id}",
-            # },
             {
                 "rel": "related",
                 "href": f"../../projects/{PROJECT_COLLECTION_NAME}/collection.json",
                 "type": "application/json",
                 "title": "Project: DeepESDL",
             },
+            {
+                "rel": "application-originating-platform",
+                "title": "DeepESDL platform",
+                "href": "https://deep.earthsystemdatalab.net/",
+                "type": "text/html",
+                "application:platform_supports": ["jupyter-notebook"],
+                "application:preferred_app": "JupyterLab",
+            },
             {
                 "rel": "input",
                 "href": "./input.yaml",