Main

README.md

@@ -106,16 +106,19 @@ python -m rrparser --dataset rhea --diameters '<=8' -o rules.csv
 python -m rrparser --dataset rhea --diameters '>=8' -o rules.csv
 ```
 
-`--radii` supports the same expressions over integer radii 0–10:
+`-r` / `--radii` supports the same expressions over integer radii 0–10:
 `0,1,2`, `0-4`, `<4`, `<=4`, `>4`, `>=4`, and combinations such as `0-2,>=8`.
 For example, `--radii '<=4'` is equivalent to `--diameters '<=8'`.
 The two options are mutually exclusive. Both are expanded into normalized
 integer radii by the argument parser, before calling `parse_rules()`.
 Without either option, all radii 0–10 (diameters 0,2,…,20) are selected.
 
 For templates, a row is retained when at least one source radius matches
-`diameter / 2`, once per template. The output `RADII` column retains all source
-radii, separated by semicolons. Explicit odd diameters or values above 20 are
+`diameter / 2`, once per template. The output `RADII` column retains only the
+selected radii, separated by semicolons; `RADIUS_MIN` and `RADIUS_MAX` are
+updated to match. For example, `-r '>4'` exports only radii 5–10 even if a
+matching source template also covers radius 4. Cached sources are unchanged.
+Explicit odd diameters or values above 20 are
 rejected before parsing, including for rr02.
 
 rr02 supplies diameters 2,4,…,16. The extended default therefore selects all
@@ -161,7 +164,7 @@ archive cache.
 | `--list-releases` | Print available release/dataset pairs and exit. |
 | `--cache-dir`, `--rules-dir` | Source cache directory; default package-local `rrparser/cache`. |
 | `-d`, `--diameters` | List, range, or comparison; default all even diameters 0–20. |
-| `--radii` | List, range, or comparison over radii 0–10, as an alternative to `--diameters`. |
+| `-r`, `--radii` | List, range, or comparison over radii 0–10, as an alternative to `--diameters`. |
 | `-rt`, `--rule-type` | `all` (default), `retro`, or `forward`; templates require `all`. |
 | `-ec`, `--ec` | File of comma-separated EC prefixes to keep. |
 | `-ecx`, `--ecx` | File of comma-separated EC prefixes to exclude; mutually exclusive with `--ec`. |

rrparser/Args.py

@@ -142,6 +142,7 @@ def _add_arguments(parser):
         "For templates, diameter = 2 * radius; rr02 provides diameters 2 to 16",
     )
     sizes.add_argument(
+        "-r",
         "--radii",
         type=parse_radii,
         help='Radii: list (0,1,2), inclusive range (0-4), "<4", "<=4", ">4" or ">=4". '

rrparser/Parser.py

@@ -14,7 +14,7 @@
 from re import split
 from tarfile import open as open_tar
 
 from pandas import DataFrame, concat, read_csv
 
 from .Args import (
     DEFAULT_RADII,
@@ -97,7 +97,7 @@
     }
     for column in ("REACTIONS", "ECS", "DATASETS", "RADII"):
         if column in aggregations:
             aggregations[column] = union
     aggregations["SCORE"] = "max"
     if "VALID" in aggregations:
         aggregations["VALID"] = "all"
@@ -115,7 +115,7 @@
         for column, position in (("RADIUS_MIN", 0), ("RADIUS_MAX", -1)):
             if column in merged.columns:
                 merged[column] = radii.map(
                     lambda values, position=position: (
                         values[position] if values else None
                     )
                 )
@@ -208,7 +208,7 @@
                 "SCORE": "Score",
             }
         )
     results = filter_(rf, rule_type, radii, ecx, ec, logger=logger)
 
     # If 'rules_scores_file' is set, then read it and merge with 'results'
     if scores_file:
@@ -299,7 +299,7 @@
     ecx = ecx if ecx is not None else []
     ec = ec if ec is not None else []
     ec_filter = ecx if ecx else ec
     filter = "in" if ec else "not in"
     logger.info("Filtering rules:")
     logger.info(f" -> EC numbers ({filter}): {ec_filter}")
     logger.info(f" -> Radii: {radii}")
@@ -320,16 +320,18 @@
             raise ValueError(
                 "RetroRules templates have no retro/forward annotation; use --rule-type all"
             )
-        radius_idx = df[
+        selection = set(radii)
+        selected_radii = (
             df["RADII"]
             .fillna("")
             .astype(str)
-            .apply(
-                lambda values: any(
-                    int(radius) in radii for radius in values.split(";") if radius
+            .map(
+                lambda values: sorted(
+                    {int(radius) for radius in values.split(";") if radius} & selection
                 )
             )
-        ].index
+        )
+        radius_idx = df[selected_radii.map(bool)].index
         query_vars["radius_idx"] = radius_idx
         query += " & index in @radius_idx"
     else:
@@ -340,6 +342,13 @@
     logger.debug(f"Query: {query}")
     try:
         _df = df.query(query, local_dict=query_vars)
+        if "RADII" in _df.columns:
+            _df = _df.copy()
+            retained_radii = selected_radii.loc[_df.index]
+            _df["RADII"] = retained_radii.map(lambda values: ";".join(map(str, values)))
+            for column, bound in (("RADIUS_MIN", min), ("RADIUS_MAX", max)):
+                if column in _df.columns:
+                    _df[column] = retained_radii.map(bound)
         logger.info("")
         logger.info(f"=> {len(_df)} rules found")
         logger.info("")
@@ -349,7 +358,7 @@
 
 
 def filter_rules_by_ec(
     df: DataFrame, ec_filter: list[str], logger: Logger = LOGGER
 ) -> list[int]:
     """
     Filter out rules that contain at least one EC number that starts with one in 'ecx'.
@@ -358,7 +367,7 @@
     ----------
     df: DataFrame
         Pandas dataframe
     ec: List[str]
         List of EC numbers to filter
     logger : Logger
         The logger object.
@@ -367,7 +376,7 @@
     ec_numbers = df["EC number"].fillna("").str.split(";")
 
     # Find indices of rows where any EC number starts with one in 'ec_filter'
     def filter_ec_numbers(ec_list, ec_filter):
         return any(any(ec.startswith(ec_f) for ec_f in ec_filter) for ec in ec_list)
 
     # Apply filtering function and get indices
@@ -396,9 +405,9 @@
 
 def fetch_retrorules(
     rules_dir: str = DEFAULT_RULES_DIR,
     logger: Logger = LOGGER,
     release: str = DEFAULT_RELEASE,
     dataset: str = "metanetx",
 ) -> str:
     """
     Fetch RetroRules over Internet if not already on disk.

tests/test_aggregation.py

@@ -80,6 +80,9 @@ def test_cached_datasets_are_merged_before_filtering(tmp_path, selection):
     assert shared["RADIUS_MIN"] == 0
     assert shared["RADIUS_MAX"] == 10
     assert pd.read_csv(StringIO(filtered))["Rule ID"].tolist() == ["shared"]
+    filtered_row = pd.read_csv(StringIO(filtered), dtype={"RADII": str}).iloc[0]
+    assert filtered_row["RADII"] == "10"
+    assert filtered_row["RADIUS_MIN"] == filtered_row["RADIUS_MAX"] == 10
     assert [source.read_bytes() for source in sources] == originals
 
 

tests/test_radii.py

@@ -50,6 +50,7 @@ def test_radius_expressions_select_templates(tmp_path, expression, expected):
     [
         ("--radii", "0-4", [0, 1, 2, 3, 4]),
         ("--radii", "<=4", [0, 1, 2, 3, 4]),
+        ("-r", ">4", [5, 6, 7, 8, 9, 10]),
         ("--diameters", "0-8", [0, 1, 2, 3, 4]),
         ("-d", "<=8", [0, 1, 2, 3, 4]),
         ("--radii", ">=8", [8, 9, 10]),
@@ -108,3 +109,43 @@ def test_defaults_and_empty_selection():
     args = build_args_parser().parse_args([])
     assert list(args.radii) == list(range(11))
     assert parse_diameters("<0") == parse_radii("<0") == []
+
+
+@pytest.mark.parametrize(
+    "option,expression,expected",
+    [
+        ("-r", ">4", "5;6;7;8;9;10"),
+        ("--radii", ">=4", "4;5;6;7;8;9;10"),
+        ("-r", "4-6", "4;5;6"),
+        ("-r", "0,10", "10"),
+        ("-d", ">8", "5;6;7;8;9;10"),
+        ("-r", ">10", ""),
+    ],
+)
+def test_exported_radius_metadata_matches_selection(
+    tmp_path, option, expression, expected
+):
+    source = tmp_path / "templates.csv"
+    pd.DataFrame(
+        {
+            "TEMPLATE_ID": ["RR:shared"],
+            "TEMPLATE": ["[C]>>[O]"],
+            "SCORE": [1.0],
+            "ECS": ["1.1.1.1"],
+            "RADII": ["4;5;6;7;8;9;10"],
+            "RADIUS_MIN": [4],
+            "RADIUS_MAX": [10],
+        }
+    ).to_csv(source, index=False)
+    original = source.read_bytes()
+    args = build_args_parser().parse_args([option, expression])
+    result = parse_rules(None, rules_file=str(source), radii=args.radii)
+    output = pd.read_csv(StringIO(result), dtype={"RADII": str})
+    if expected:
+        assert output["RADII"].tolist() == [expected]
+        radii = list(map(int, expected.split(";")))
+        assert output["RADIUS_MIN"].tolist() == [min(radii)]
+        assert output["RADIUS_MAX"].tolist() == [max(radii)]
+    else:
+        assert output.empty
+    assert source.read_bytes() == original