BUG: Move handling of context arguments after handling of .depends keyword

skypy/pipeline/_pipeline.py

@@ -102,8 +102,6 @@ def __init__(self, configuration):
             self.dag.add_node(job, skip=False)
             if isinstance(settings, Item):
                 items[job] = settings
-                # infer additional item properties from context
-                settings.infer(context)
         for table, columns in self.table_config.items():
             table_complete = '.'.join((table, 'complete'))
             self.dag.add_node(table_complete)
@@ -115,8 +113,6 @@ def __init__(self, configuration):
                 self.dag.add_edge(job, table_complete)
                 if isinstance(settings, Item):
                     items[job] = settings
-                    # infer additional item properties from context
-                    settings.infer(context)
                 # DAG nodes for individual columns in multi-column assignment
                 names = [n.strip() for n in column.split(',')]
                 if len(names) > 1:
@@ -138,6 +134,8 @@ def __init__(self, configuration):
                 while c:
                     self.dag.add_edge(c, d)
                     c, d = c.rpartition('.')[0], c
+            # infer additional item properties from context
+            settings.infer(context)
 
     def execute(self, parameters={}):
         r'''Run a pipeline.

skypy/pipeline/tests/test_pipeline.py

@@ -2,7 +2,7 @@
 from astropy.cosmology.core import Cosmology
 from astropy.io import fits
 from astropy.io.misc.hdf5 import read_table_hdf5
-from astropy.table import Table
+from astropy.table import Table, vstack
 from astropy.table.column import Column
 from astropy.units import Quantity
 from astropy.utils.data import get_pkg_data_filename
@@ -267,3 +267,25 @@ def test_hdf5(tmp_path):
     pipeline.write(output_filename)
     hdf_table = read_table_hdf5(output_filename, 'tables/test_table', character_as_bytes=False)
     assert np.all(hdf_table == pipeline['test_table'])
+
+
+def test_depends():
+
+    # Regression test for GitHub Issue #464
+    # Previously the .depends keyword was also being passed to functions as a
+    # keyword argument. This was because Pipeline was executing Item.infer to
+    # handle additional function arguments from context before handling
+    # additional dependencies specified using the .depends keyword. The
+    # .depends keyword is now handled first.
+
+    config = {'tables': {
+                'table_1': {
+                  'column1': Call(np.random.uniform, [0, 1, 10])},
+                'table_2': {
+                    '.init': Call(vstack, [], {
+                      'tables': [Ref('table_1')],
+                      '.depends': ['table_1.complete']})}}}
+
+    pipeline = Pipeline(config)
+    pipeline.execute()
+    assert np.all(pipeline['table_1'] == pipeline['table_2'])