Merge branch 'main' into data_europa_eu

J535D165 · Sep 29, 2024 · a0e8101 · a0e8101
2 parents 42a8d13 + 6ca501b
commit a0e8101
Show file tree

Hide file tree

Showing 4 changed files with 27 additions and 1 deletion.
diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
@@ -10,7 +10,7 @@ repos:
       "mkdocs.yml"
   - id: check-added-large-files
 - repo: https://github.com/astral-sh/ruff-pre-commit
-  rev: v0.4.9
+  rev: v0.6.7
   hooks:
     # Run the linter.
     - id: ruff

diff --git a/datahugger/config.py b/datahugger/config.py
@@ -11,6 +11,7 @@
 from datahugger.services import MendeleyDataset
 from datahugger.services import OSFDataset
 from datahugger.services import PangaeaDataset
+from datahugger.services import SeaNoeDataset
 from datahugger.services import ZenodoDataset
 
 # fast lookup
@@ -113,6 +114,7 @@
     "researchdata.ntu.edu.sg": DataverseDataset,
     "rin.lipi.go.id": DataverseDataset,
     "ssri.is": DataverseDataset,
+    "www.seanoe.org": SeaNoeDataset,
     "trolling.uit.no": DataverseDataset,
     "www.sodha.be": DataverseDataset,
     "www.uni-hildesheim.de": DataverseDataset,

diff --git a/datahugger/services.py b/datahugger/services.py
@@ -405,3 +405,23 @@ class DataEuropaDataset(DatasetDownloader):
     ATTR_FILE_LINK_JSONPATH = "'dcat:accessURL'.@id"
     ATTR_NAME_JSONPATH = "'dct:title'"
     ATTR_SIZE_JSONPATH = "'dcat:byteSize'.@value"
+
+
+class SeaNoeDataset(DatasetDownloader):
+    """Downloader for SeaNoe publication."""
+
+    REGEXP_ID = r"https://www.seanoe\.org/data/[0-9]+/(?P<record_id>.*)/"
+
+    # the base entry point of the REST API
+    API_URL = "https://www.seanoe.org/api/"
+
+    # the files and metadata about the dataset
+    API_URL_META = "{api_url}find-by-id/{record_id}"
+    META_FILES_JSONPATH = "files[*]"
+
+    # paths to file attributes
+    ATTR_NAME_JSONPATH = "fileName"
+    ATTR_FILE_LINK_JSONPATH = "fileUrl"
+    ATTR_SIZE_JSONPATH = "size"
+    ATTR_HASH_JSONPATH = "checksum"
+    ATTR_HASH_TYPE_VALUE = "sha256"
diff --git a/tests/test_repositories.toml b/tests/test_repositories.toml
@@ -113,3 +113,7 @@ files = "cbsodata-main/README.md"
 [[dataeuropa]]
 location = "https://data.europa.eu/data/datasets/65e092e4009f18f050b14216"
 files = "consolidation-wattzhub-schema-irve-dynamic-20240918-033000.csv"
+
+[[seanoe]]
+location = "https://doi.org/10.17882/101042"
+files = "111609.xlsx"