From 93856d58add0d2ea40e01524d352f850312e46ca Mon Sep 17 00:00:00 2001
From: Derrick Chambers <d-chambers@users.noreply.github.com>
Date: Tue, 13 Sep 2022 20:23:21 -0600
Subject: [PATCH] add cable/instrument ids to min_size for directory spool
 (#49)

---
 dascore/utils/hdf5.py               |  2 ++
 tests/conftest.py                   |  2 +-
 tests/test_clients/test_dirspool.py | 21 ++++++++++++++++-----
 3 files changed, 19 insertions(+), 6 deletions(-)

diff --git a/dascore/utils/hdf5.py b/dascore/utils/hdf5.py
index 6edd41a6..ef71eed8 100644
--- a/dascore/utils/hdf5.py
+++ b/dascore/utils/hdf5.py
@@ -123,6 +123,8 @@ class HDFPatchIndexManager:
         "station": 8,
         "dims": 40,
         "file_version": 9,
+        "cable_id": 40,
+        "instrument_id": 40,
     }
     # columns which should be indexed for fast querying
     _query_columns = ("time_min", "time_max", "distance_min", "distance_max")
diff --git a/tests/conftest.py b/tests/conftest.py
index 0aeb1f3c..4ddd34ad 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -229,7 +229,7 @@ def diverse_spool_directory(diverse_spool):
 
 
 @pytest.fixture(scope="class")
-def diverse_file_spool(diverse_spool_directory):
+def diverse_directory_spool(diverse_spool_directory):
     """Save the diverse spool contents to a directory."""
     out = dascore.spool(diverse_spool_directory).update()
     return out
diff --git a/tests/test_clients/test_dirspool.py b/tests/test_clients/test_dirspool.py
index ce721876..150f9548 100644
--- a/tests/test_clients/test_dirspool.py
+++ b/tests/test_clients/test_dirspool.py
@@ -10,6 +10,7 @@
 import dascore as dc
 from dascore.constants import ONE_SECOND
 from dascore.core.schema import PatchFileSummary
+from dascore.utils.hdf5 import HDFPatchIndexManager
 from dascore.utils.misc import register_func
 
 FILE_SPOOLS = []
@@ -113,9 +114,9 @@ def test_is_in_tag(self, basic_file_spool, spool_tag):
         out = basic_file_spool.select(tag=tag_collection).get_contents()
         assert out["tag"].isin(tag_collection).all()
 
-    def test_multiple_selects(self, diverse_file_spool):
+    def test_multiple_selects(self, diverse_directory_spool):
         """Ensure selects can be stacked."""
-        spool = diverse_file_spool
+        spool = diverse_directory_spool
         contents = spool.get_contents()
         duration = contents["time_max"] - contents["time_min"]
         new_max = (contents["time_min"] + duration.mean() / 2).median()
@@ -145,15 +146,15 @@ def test_select_time_tuple_with_string(self, basic_file_spool):
         for pa1, pa2 in zip(spool1, spool2):
             assert pa1.attrs["time_max"] == pa2.attrs["time_max"]
 
-    def test_select_non_zero_index(self, diverse_file_spool):
+    def test_select_non_zero_index(self, diverse_directory_spool):
         """
         A Bug caused the contents of the source dataframe to have
         non-zero based indices, thus spools didnt work. This fixes
         the issue.
         """
-        contents = diverse_file_spool.get_contents()
+        contents = diverse_directory_spool.get_contents()
         end_time = contents["time_max"].min()
-        sub = diverse_file_spool.select(
+        sub = diverse_directory_spool.select(
             time=(None, end_time),
             distance=(100, 200),
         )
@@ -194,6 +195,16 @@ def test_sub_chunk(self, one_file_file_spool):
             assert isinstance(patch, dc.Patch)
 
 
+class TestGetContents:
+    """Tests for getting the contents of the spool."""
+
+    def test_str_columns_in_dataframe(self, diverse_directory_spool):
+        """Ensure all the string columns are in index."""
+        df = diverse_directory_spool.get_contents()
+        expected = HDFPatchIndexManager._min_itemsize
+        assert set(df.columns).issuperset(set(expected))
+
+
 class TestFileSpoolIntegrations:
     """Small integration tests for the file spool."""