Propagate implications

Author: znicholls

src/pandas_openscm/db/backends.py

@@ -148,7 +148,7 @@ def get_instance(self, option: str) -> OpenSCMDBIndexBackend:
         )
         raise KeyError(msg)
 
-    def guess_backend(self, index_file_name: str) -> OpenSCMDBDataBackend:
+    def guess_backend(self, index_file_name: str) -> OpenSCMDBIndexBackend:
         """
         Guess backend from a file name
 

src/pandas_openscm/db/openscm_db.py

@@ -338,6 +338,14 @@ def from_gzipped_tar_archive(
                 ):
                     backend_data = DATA_BACKENDS.guess_backend(member.name)
 
+        if backend_data is None:  # pragma: noqa
+            # Should be impossible to get here
+            raise TypeError(backend_data)
+
+        if backend_index is None:  # pragma: noqa
+            # Should be impossible to get here
+            raise TypeError(backend_index)
+
         res = cls(backend_data=backend_data, backend_index=backend_index, db_dir=db_dir)
 
         return res
@@ -700,6 +708,7 @@ def save(  # noqa: PLR0913
                     file_map_start=file_map_db,
                     data_to_write=data,
                     get_new_data_file_path=self.get_new_data_file_path,
+                    db_dir=self.db_dir,
                 )
 
                 # As needed, re-write files without deleting the old files
@@ -782,7 +791,7 @@ def to_gzipped_tar_archive(self, out_file: Path, mode: str = "w:gz") -> Path:
 
             This is the same as `out_file`, but is returned for convenience.
         """
-        with tarfile.open(out_file, mode) as tar:
+        with tarfile.open(out_file, mode=mode) as tar:
             tar.add(self.db_dir, arcname="db")
 
         return out_file

src/pandas_openscm/db/rewriting.py

@@ -17,6 +17,7 @@
 from attrs import define
 
 from pandas_openscm.db.interfaces import OpenSCMDBDataBackend
+from pandas_openscm.db.path_handling import DBPath
 from pandas_openscm.index_manipulation import (
     unify_index_levels_check_index_types,
     update_index_from_candidates,
@@ -180,7 +181,8 @@ def make_move_plan(
     index_start: pd.DataFrame,
     file_map_start: pd.Series[Path],  # type: ignore # pandas confused about ability to support Path
     data_to_write: pd.DataFrame,
-    get_new_data_file_path: Callable[[int], Path],
+    get_new_data_file_path: Callable[[int], DBPath],
+    db_dir: Path,
 ) -> MovePlan:
     """
     Make a plan for moving data around to make room for new data
@@ -196,6 +198,12 @@ def make_move_plan(
     data_to_write
         Data that is going to be written in the database
 
+    get_new_data_file_path
+        Callable which, given an integer, returns the path info for the new data file
+
+    db_dir
+        Database directory
+
     Returns
     -------
     :
@@ -229,7 +237,7 @@ def make_move_plan(
         # (would be even more efficient to just update the file IDs,
         # but that would create a coupling I can't get my head around right now).
         delete_file_ids = full_overwrite.index[full_overwrite]
-        delete_paths = file_map_start.loc[delete_file_ids]
+        delete_paths = (db_dir / v for v in file_map_start.loc[delete_file_ids])
         moved_index = index_start[~index_start["file_id"].isin(delete_file_ids)]
         file_map_out = file_map_start.loc[moved_index["file_id"].unique()]
 
@@ -246,7 +254,7 @@ def make_move_plan(
     full_overwrite_file_ids = full_overwrite.index[full_overwrite]
     partial_overwrite_file_ids = partial_overwrite.index[partial_overwrite]
     file_ids_to_delete = np.union1d(full_overwrite_file_ids, partial_overwrite_file_ids)
-    delete_paths = file_map_start.loc[file_ids_to_delete]
+    delete_paths = (db_dir / v for v in file_map_start.loc[file_ids_to_delete])
 
     file_id_map = {}
     max_file_id_start = file_map_start.index.max()
@@ -259,12 +267,13 @@ def make_move_plan(
     ):
         new_file_id = max_file_id_start + 1 + increment
 
-        file_map_out.loc[new_file_id] = get_new_data_file_path(new_file_id)
+        new_db_path = get_new_data_file_path(new_file_id)
+        file_map_out.loc[new_file_id] = new_db_path.rel_db
 
         rewrite_actions_l.append(
             ReWriteAction(
-                from_file=file_map_start.loc[file_id_old],
-                to_file=file_map_out.loc[new_file_id],
+                from_file=db_dir / file_map_start.loc[file_id_old],
+                to_file=new_db_path.abs,
                 locator=fiddf.index.droplevel("file_id"),
             )
         )