Remove duckdb and cli index selection

nomenklatura/cli.py

@@ -10,7 +10,7 @@
 from followthemoney.cli.aggregate import sorted_aggregate
 
 from nomenklatura.cache import Cache
-from nomenklatura.index import MEMORY_INDEX_PATH
+from nomenklatura.index import Index
 from nomenklatura.matching import train_v2_matcher, train_v1_matcher
 from nomenklatura.store import load_entity_file_store
 from nomenklatura.resolver import Resolver
@@ -64,7 +64,6 @@ def cli() -> None:
 @click.option("-l", "--limit", type=click.INT, default=5000)
 @click.option("--algorithm", default=DefaultAlgorithm.NAME)
 @click.option("--scored/--unscored", is_flag=True, type=click.BOOL, default=True)
-@click.option("-i", "--index", type=click.STRING)
 @click.option(
     "-c",
     "--clear",
@@ -79,7 +78,7 @@ def xref_file(
     algorithm: str = DefaultAlgorithm.NAME,
     limit: int = 5000,
     scored: bool = True,
-    index: str = MEMORY_INDEX_PATH,
+    index: str = Index.name,
     clear: bool = False,
 ) -> None:
     resolver_ = _get_resolver(path, resolver)
@@ -103,7 +102,6 @@ def xref_file(
         algorithm=algorithm_type,
         scored=scored,
         limit=limit,
-        index_path=index,
     )
     resolver_.save()
     log.info("Xref complete in: %s", resolver_.path)

nomenklatura/index/__init__.py

@@ -1,40 +1,5 @@
-from importlib import import_module
-import logging
-from pathlib import Path
-from typing import Type, Optional
-
 from nomenklatura.index.index import Index
 from nomenklatura.index.common import BaseIndex
-from nomenklatura.store import View
-from nomenklatura.dataset import DS
-from nomenklatura.entity import CE
-
-log = logging.getLogger(__name__)
-
-
-MEMORY_INDEX_PATH = "nomenklatura.index.index.Index"
-
-
-def get_index(
-    view: View[DS, CE], path: Path, class_path: Optional[str]
-) -> BaseIndex[DS, CE]:
-    """Get the best available index class to use."""
-    clazz: Type[BaseIndex[DS, CE]] = Index[DS, CE]
-    if class_path is not None:
-        try:
-            module_path, class_name = class_path.rsplit(".", 1)
-            module = import_module(module_path)
-            clazz_ref = getattr(module, class_name)
-
-            clazz = clazz_ref[DS, CE]
-        except ImportError:
-            log.warning(
-                "f`{class_path}` is not available, falling back to in-memory index."
-            )
-
-    index = clazz(view, path)
-    index.build()
-    return index
 
 
-__all__ = ["BaseIndex", "Index", "MEMORY_INDEX_PATH", "get_index"]
+__all__ = ["BaseIndex", "Index"]

nomenklatura/xref.py

@@ -3,12 +3,13 @@
 from followthemoney.schema import Schema
 from pathlib import Path
 
+from nomenklatura import Index
 from nomenklatura.dataset import DS
 from nomenklatura.entity import CE
 from nomenklatura.store import Store
 from nomenklatura.judgement import Judgement
 from nomenklatura.resolver import Resolver
-from nomenklatura.index import get_index
+from nomenklatura.index import BaseIndex
 from nomenklatura.matching import DefaultAlgorithm, ScoringAlgorithm
 from nomenklatura.conflicting_match import ConflictingMatchReporter
 
@@ -31,6 +32,7 @@ def xref(
     resolver: Resolver[CE],
     store: Store[DS, CE],
     index_dir: Path,
+    index_type: Type[BaseIndex[DS, CE]] = Index,
     limit: int = 5000,
     limit_factor: int = 10,
     scored: bool = True,
@@ -41,12 +43,12 @@ def xref(
     conflicting_match_threshold: Optional[float] = None,
     focus_dataset: Optional[str] = None,
     algorithm: Type[ScoringAlgorithm] = DefaultAlgorithm,
-    index_path: Optional[str] = None,
     user: Optional[str] = None,
 ) -> None:
     log.info("Begin xref: %r, resolver: %s", store, resolver)
     view = store.default_view(external=external)
-    index = get_index(view, index_dir, index_path)
+    index = index_type(view, index_dir)
+    index.build()
     conflict_reporter = None
     if conflicting_match_threshold is not None:
         conflict_reporter = ConflictingMatchReporter(

tests/conftest.py

@@ -6,7 +6,6 @@
 from tempfile import mkdtemp
 
 from nomenklatura import settings
-from nomenklatura.index.duckdb_index import DuckDBIndex
 from nomenklatura.index.tantivy_index import TantivyIndex
 from nomenklatura.store import load_entity_file_store, SimpleMemoryStore
 from nomenklatura.kv import get_redis
@@ -82,13 +81,6 @@ def tantivy_index(index_path: Path, dstore: SimpleMemoryStore):
     yield index
 
 
-@pytest.fixture(scope="function")
-def duckdb_index(index_path: Path, dstore: SimpleMemoryStore):
-    index = DuckDBIndex(dstore.default_view(), index_path)
-    index.build()
-    yield index
-
-
 @pytest.fixture(scope="function")
 def index_path():
     index_path = Path(mkdtemp()) / "index-dir"