Support 'single-record-per-directory' format

datalad_tabby/conftest.py

@@ -5,6 +5,8 @@
 from datalad_tabby.tests.fixtures import (
     # provides the tabby "demorecord" that is shipped with the sources
     tabby_tsv_record,
+    # same again, but in a single-record-per-dir layout
+    tabby_tsv_singledir_record,
     # elementary tabby record comprising the key TSV buildiung blocks
     tabby_record_basic_components,
     # no-LD elementary record with overrides

datalad_tabby/io/load_utils.py

@@ -62,13 +62,19 @@ def _build_overrides(src: Path, obj: Dict):
 
 
 def _get_corresponding_sheet_fpath(fpath: Path, sheet_name: str) -> Path:
-    return fpath.parent / \
-        f'{_get_tabby_prefix_from_sheet_fpath(fpath)}_{sheet_name}.tsv'
+    prefix = _get_tabby_prefix_from_sheet_fpath(fpath)
+    if prefix:
+        return fpath.parent / f'{prefix}_{sheet_name}.tsv'
+    else:
+        return fpath.parent / f'{sheet_name}.tsv'
 
 
 def _get_record_context_fpath(fpath: Path) -> Path:
     prefix = _get_tabby_prefix_from_sheet_fpath(fpath)
-    return fpath.parent / f'{prefix}.ctx.jsonld'
+    if prefix:
+        return fpath.parent / f'{prefix}.ctx.jsonld'
+    else:
+        return fpath.parent / f'ctx.jsonld'
 
 
 def _get_corresponding_context_fpath(fpath: Path) -> Path:
@@ -85,8 +91,11 @@ def _get_corresponding_override_fpath(fpath: Path) -> Path:
 
 def _get_tabby_prefix_from_sheet_fpath(fpath: Path) -> str:
     stem = fpath.stem
-    # stem up to, but not including, the last '_'
-    return stem[:(-1) * stem[::-1].index('_') - 1]
+    if '_' not in stem:
+        return ''
+    else:
+        # stem up to, but not including, the last '_'
+        return stem[:(-1) * stem[::-1].index('_') - 1]
 
 
 def _get_index_after_last_nonempty(val: List) -> int:

datalad_tabby/io/tests/test_load.py

@@ -57,3 +57,9 @@ def test_load_almost_tabby_import(tmp_path):
 
     rec = load_tabby(src)
     assert rec['dummy'] == '@tabby-murks'
+
+
+def test_load_singldir_format(tabby_tsv_record, tabby_tsv_singledir_record):
+    rec = load_tabby(tabby_tsv_record['root_sheet'])
+    srec = load_tabby(tabby_tsv_singledir_record['root_sheet'])
+    assert rec == srec

datalad_tabby/tests/fixtures.py

@@ -1,6 +1,7 @@
 import json
 from pathlib import Path
 import pytest
+from shutil import copyfile
 
 from datalad_next.tests.utils import md5sum
 
@@ -23,6 +24,36 @@ def tabby_tsv_record():
     )
 
 
+@pytest.fixture(autouse=False, scope="session")
+def tabby_tsv_singledir_record(tmp_path_factory):
+    srcdir = Path(dttests.__file__).parent / 'data' / 'demorecord'
+    rfiles = list(srcdir.glob('tabbydemo*'))
+    rdir = tmp_path_factory.mktemp("demorecord_single") / 'tabbydemo'
+    rdir.mkdir()
+
+    # we copy all the files, windows does not play with symlinks in general
+    dst_files = [
+        # strip the prefix from the filename, clean `.ctx.jsonld`
+        rdir / ('ctx.jsonld'
+                if fpath.name == 'tabbydemo.ctx.jsonld'
+                else fpath.name.split('_')[-1].lstrip('.'))
+        for fpath in rfiles
+    ]
+    for s, d in zip(rfiles, dst_files):
+        copyfile(s, d)
+
+    root_sheet = rdir / 'dataset.tsv'
+    assert root_sheet in dst_files
+    assert root_sheet.exists()
+
+    sheets = [f for f in dst_files if f.name.endswith('.tsv')]
+    yield dict(
+        root_sheet=root_sheet,
+        sheets=sheets,
+        md5={s.name: md5sum(s) for s in sheets},
+    )
+
+
 @pytest.fixture(scope="session")
 def tabby_record_basic_components(tmp_path_factory):
     rdir = tmp_path_factory.mktemp("rec")