refactoring/linting

Author: florian-huber

ms2query/database/spec_to_compound_mapper.py

@@ -27,6 +27,7 @@ class SpecToCompoundMap:
     sqlite_path: str
     table: str = "spec_to_comp"
     compound_table: str = "compounds"
+
     _conn: sqlite3.Connection = field(init=False, repr=False)
 
     def __post_init__(self):
@@ -84,6 +85,8 @@ def link_many(self, pairs: Iterable[Tuple[int, str]]):
             cur.execute("ROLLBACK")
             raise
 
+    # ---- getters: spec_id -> comp_id ----
+
     def get_comp_id_for_specs(self, spec_ids: List[str]) -> pd.DataFrame:
         """Return a DataFrame with columns [spec_id, comp_id] for the provided spec_ids."""
         if not spec_ids:

ms2query/database/spectral_database.py

@@ -390,16 +390,6 @@ def get_embeddings(
             X = X / n
         return np.asarray(sids, dtype=object), X
 
-    def get_embedding_for_id(
-        self,
-        spec_id: str,
-        *,
-        embeddings_table: str = "embeddings",
-        normalized: bool = True,
-    ) -> Optional[np.ndarray]:
-        ids, X = self.get_embeddings([spec_id], embeddings_table=embeddings_table, normalized=normalized)
-        return X[0] if X.shape[0] else None
-
     # expose raw connection for ANN builders
     @property
     def connection(self) -> sqlite3.Connection:

ms2query/ms2query_library.py

@@ -26,8 +26,6 @@ class MS2QueryLibrary:
 
     Notes
     -----
-    * `model_path` is optional; provide it if you want on-the-fly embedding of ad-hoc spectra.
-      If omitted, you can still query using precomputed embeddings fetched from SQLite.
     * EmbeddingIndex must be built/loaded elsewhere (creation handled by setup workflow).
     """
     db: MS2QueryDatabase
@@ -221,7 +219,16 @@ def query_compounds_by_compounds(
             ]
             results_all.append(one)
 
-        return results_all
+        # Convert to DataFrame
+        results_df = pd.DataFrame(
+            [
+                {"query_ix": qi, **item}
+                for qi, lst in enumerate(results_all)
+                for item in lst
+            ],
+            columns=["query_ix", "rank", "comp_id", "score"],
+        )
+        return results_df
 
     def query_compounds_by_spectra(
         self,