Migrate from datapackage to frictionless

Re #491

pyproject.toml

@@ -15,7 +15,7 @@ requires-python = ">=3.9"
 dependencies = [
     "SQLAlchemy >=1.4, <1.5",
     "alembic >=1.7",
-    "datapackage >=1.15.2",
+    "frictionless >=5.18",
     "python-dateutil >=2.8.1",
     "numpy >=1.20.2",
     "scipy >=1.7.1",

spinedb_api/spine_io/importers/datapackage_reader.py

@@ -11,10 +11,9 @@
 ######################################################################################################################
 
 """ Contains DataPackageConnector class. """
-from itertools import chain
+from itertools import chain, islice
 import threading
-from datapackage import Package
-import tabulator.exceptions
+import frictionless
 from ...exception import ConnectorError
 from .reader import SourceConnection
 
@@ -63,7 +62,7 @@ def connect_to_source(self, source, **extras):
             **extras: ignored
         """
         if source:
-            self._datapackage = Package(source)
+            self._datapackage = frictionless.Package(source)
             self._filename = source
 
     def disconnect(self):
@@ -95,11 +94,12 @@ def get_data_iterator(self, table, options, max_rows=-1):
         """
         if not self._datapackage:
             return iter([]), []
+        max_rows = max_rows if max_rows >= 0 else None
 
-        def iterator(r):
+        def iterator(i):
             try:
-                yield from (item for row, item in enumerate(r.iter(cast=False)) if row != max_rows)
-            except tabulator.exceptions.TabulatorException as error:
+                yield from i
+            except frictionless.exception.FrictionlessException as error:
                 raise ConnectorError(str(error)) from error
 
         has_header = options.get("has_header", True)
@@ -108,9 +108,11 @@ def iterator(r):
                 if resource.name is None:
                     resource.infer()
             if table == resource.name:
-                if has_header:
-                    header = resource.schema.field_names
-                    return iterator(resource), header
-                return chain([resource.headers], iterator(resource)), None
+                with resource:
+                    i = iterator(islice((row.to_list(json=True) for row in resource.row_stream), max_rows))
+                    if has_header:
+                        header = resource.header
+                        return i, header
+                    return chain([resource.header.labels], i), None
         # table not found
         return iter([]), []

tests/spine_io/importers/test_datapackage_reader.py

@@ -9,18 +9,13 @@
 # Public License for more details. You should have received a copy of the GNU Lesser General Public License along with
 # this program. If not, see <http://www.gnu.org/licenses/>.
 ######################################################################################################################
-
-"""
-Contains unit tests for DatapackageConnector.
-
-"""
 from contextlib import contextmanager
 import csv
 from pathlib import Path
 import pickle
 from tempfile import TemporaryDirectory
 import unittest
-from datapackage import Package
+from frictionless import Package, Resource
 from spinedb_api.exception import ConnectorError
 from spinedb_api.spine_io.importers.datapackage_reader import DataPackageConnector
 
@@ -67,10 +62,10 @@ def test_wrong_datapackage_encoding_raises_connector_error(self):
             with open(csv_file_path, "wb") as csv_file:
                 for row in data:
                     csv_file.write(row)
-            package = Package(base_path=temp_dir)
-            package.add_resource({"path": str(csv_file_path.relative_to(temp_dir))})
+            package = Package(basepath=temp_dir)
+            package.add_resource(Resource(path=str(csv_file_path.relative_to(temp_dir))))
             package_path = Path(temp_dir, "datapackage.json")
-            package.save(package_path)
+            package.to_json(package_path)
             reader = DataPackageConnector(None)
             reader.connect_to_source(str(package_path))
             data_iterator, header = reader.get_data_iterator("test_data", {"has_header": False})
@@ -85,10 +80,10 @@ def test_datapackage(rows):
         with open(csv_file_path, "w", newline="") as csv_file:
             csv_writer = csv.writer(csv_file)
             csv_writer.writerows(rows)
-        package = Package(base_path=temp_dir)
-        package.add_resource({"path": str(csv_file_path.relative_to(temp_dir))})
+        package = Package(basepath=temp_dir)
+        package.add_resource(Resource(path=str(csv_file_path.relative_to(temp_dir))))
         package_path = Path(temp_dir, "datapackage.json")
-        package.save(package_path)
+        package.to_json(package_path)
         yield package_path