Change direct access handling to support JupyterHub

CHANGELOG.md

@@ -7,6 +7,22 @@ and this project uses [Semantic Versioning](https://semver.org/spec/v2.0.0.html)
 
 ## [Unreleased]
 
+### Removed
+- Remove `earthaccess.__store__.in_region` member variable. There is no accurate way to determine if what region
+  a client is running in. Better to try and raise an exception or attempt HTTPS access
+- Remove in_region specific unit tests
+
+### Added
+- add `access` argument to `api.download` and `store.get` functions. Allows you to skip direct access and immediately
+  attempt HTTPS access
+- Add `out_of_region_handling` argument to `api.download` and `store.get` functions. Defaults to `raise`, which raises
+  Exceptions if direct access fails. If `handle` is passed and direct access fails, a warning is raised that includes
+  the encountered exception, and then HTTPS access is attempted. ([#444](https://github.com/nsidc/earthaccess/issues/444))
+
+### Changed
+- Update `test_download_deferred_failure` and `test_download_immediate_failure` integration tests to test both direct an external access
+- raise exception in `conftest.py` if EarthData environment variables not found
+
 ## [v0.12.0] - 2024-11-13
 
 ### Changed

earthaccess/api.py

@@ -206,6 +206,8 @@ def download(
     local_path: Optional[Union[Path, str]] = None,
     provider: Optional[str] = None,
     threads: int = 8,
+    access: Optional[str] = None,
+    out_of_region_handling: Optional[str] = "raise",
     *,
     pqdm_kwargs: Optional[Mapping[str, Any]] = None,
 ) -> List[str]:
@@ -224,6 +226,9 @@ def download(
             of a UUID4 value.
         provider: if we download a list of URLs, we need to specify the provider.
         threads: parallel number of threads to use to download the files, adjust as necessary, default = 8
+        access: "direct" to attempt direct S3 access, "external" for HTTPS
+        out_of_region_handling: "raise" to raise an Exception if attempting out-of-region access or
+            "handle" (or anything else) to attempt using HTTPS upon faliure
         pqdm_kwargs: Additional keyword arguments to pass to pqdm, a parallel processing library.
             See pqdm documentation for available options. Default is to use immediate exception behavior
             and the number of jobs specified by the `threads` parameter.
@@ -243,7 +248,12 @@ def download(
 
     try:
         return earthaccess.__store__.get(
-            granules, local_path, provider, threads, pqdm_kwargs=pqdm_kwargs
+            granules,
+            local_path,
+            provider,
+            threads,
+            access=access,
+            pqdm_kwargs=pqdm_kwargs,
         )
     except AttributeError as err:
         logger.error(

earthaccess/auth.py

@@ -178,6 +178,11 @@ def get_s3_credentials(
                 auth_url = self._get_cloud_auth_url(
                     daac_shortname=daac, provider=provider
                 )
+                if not auth_url:
+                    # Display possible typos in a helpfull error
+                    raise Exception(
+                        f'auth_url not found using daac: "{daac}" and provider: "{provider}"'
+                    )
             else:
                 auth_url = endpoint
             if auth_url.startswith("https://"):

earthaccess/results.py

@@ -303,50 +303,37 @@ def _derive_s3_link(self, links: List[str]) -> List[str]:
                 s3_links.append(f's3://{links[0].split("nasa.gov/")[1]}')
         return s3_links
 
-    def data_links(
-        self, access: Optional[str] = None, in_region: bool = False
-    ) -> List[str]:
+    def data_links(self, access: Optional[str] = None) -> List[str]:
         """Placeholder.
 
         Returns the data links from a granule.
 
         Parameters:
             access: direct or external.
                 Direct means in-region access for cloud-hosted collections.
-            in_region: True if we are running in us-west-2.
-                It is meant for the store class.
 
         Returns:
             The data links for the requested access type.
         """
         https_links = self._filter_related_links("GET DATA")
         s3_links = self._filter_related_links("GET DATA VIA DIRECT ACCESS")
-        if in_region:
-            # we are in us-west-2
-            if self.cloud_hosted and access in (None, "direct"):
-                # this is a cloud collection, and we didn't specify the access type
-                # default to S3 links
-                if len(s3_links) == 0 and len(https_links) > 0:
-                    # This is guessing the S3 links for some cloud collections that for
-                    # some reason only offered HTTPS links
-                    return self._derive_s3_link(https_links)
-                else:
-                    # we have the s3 links so we return those
-                    return s3_links
+
+        # we assume, perhaps incorrectly, that we are in us-west-2
+        if self.cloud_hosted and access == "direct":
+            # this is a cloud collection, and we didn't specify the access type
+            # default to S3 links
+            if len(s3_links) == 0 and len(https_links) > 0:
+                # This is guessing the S3 links for some cloud collections that for
+                # some reason only offered HTTPS links
+                return self._derive_s3_link(https_links)
             else:
-                # Even though we are in us-west-2, the user wants the HTTPS links used in-region.
-                # They are S3 signed links from TEA.
-                # <https://github.com/asfadmin/thin-egress-app>
-                return https_links
-        else:
-            # we are not in-region
-            if access == "direct":
-                # maybe the user wants to collect S3 links and use them later
-                # from the cloud
+                # we have the s3 links so we return those
                 return s3_links
-            else:
-                # we are not in us-west-2, even cloud collections have HTTPS links
-                return https_links
+        else:
+            # Even though we are in us-west-2, the user wants the HTTPS links used in-region.
+            # They are S3 signed links from TEA.
+            # <https://github.com/asfadmin/thin-egress-app>
+            return https_links
 
     def dataviz_links(self) -> List[str]:
         """Placeholder.