feat: Disallow the usage of similarity scores and edit distances at t…

…he same time

docs/source/command-line-interface.md

@@ -33,8 +33,8 @@ The `samples` refer to a single or multiple alignment-files (or a directory cont
   - `--acc2tax`: Path to the (NCBI) `acc2tax.gz`. Mandatory for LCA.
   - `--min-similarity-score`: Normalised edit distance (read to reference similarity) minimum. Number between 0-1. Default: 0.95.
   - `--max-similarity-score`: Normalised edit distance (read to reference similarity) maximum. Number between 0-1 Default: 1.0.
-  - `--min-edit-dist`: Minimum edit distance (read to reference similarity). Number between 0-10. Default: 0.
-  - `--max-edit-dist`: Maximum edit distance (read to reference similarity). Number between 0-10. Default: 10.
+  - `--min-edit-dist`: Minimum edit distance (read to reference similarity). Positive integer. Note that edit distances scores cannot be set at the same time as similarity scores; choose one or the other.
+  - `--max-edit-dist`: Maximum edit distance (read to reference similarity). Positive integer. Note that edit distances scores cannot be set at the same time as similarity scores; choose one or the other.
   - `--min-mapping-quality`: Minimum mapping quality. Default: 0.
   - `--lca-rank`: The LCA rank used in ngsLCA. Can be either `family`, `genus`, `species` or `""` (everything). Default is `""`.
 

src/metaDMG/cli/cli.py

@@ -1,6 +1,6 @@
 #%%
 from pathlib import Path
-from typing import Optional
+from typing import Optional, Union
 
 import typer
 
@@ -52,52 +52,56 @@ def create_config(
         ...,
         help="Single or multiple alignment-files (or a directory containing them).",
     ),
-    # LCA parameters
+    # LCA parameter
     names: Optional[Path] = typer.Option(
         None,
         exists=True,
         file_okay=True,
         help="Path to the (NCBI) names-mdmg.dmp.",
     ),
+    # LCA parameter
     nodes: Optional[Path] = typer.Option(
         None,
         exists=True,
         file_okay=True,
         help="Path to the (NCBI) nodes-mdmg.dmp.",
     ),
+    # LCA parameter
     acc2tax: Optional[Path] = typer.Option(
         None,
         exists=True,
         file_okay=True,
         help="Path to the (NCBI) acc2tax.gz.",
     ),
-    min_similarity_score: float = typer.Option(
-        0.95,
+    min_similarity_score: Optional[float] = typer.Option(
+        None,
         "--min-similarity-score",
         "-s",
         help="Normalised edit distance (read to reference similarity) minimum. Number between 0-1.",
-        callback=lambda x: cli_utils.is_in_range(x, 0, 1),
+        callback=lambda x: cli_utils.is_in_range_or_None(x, 0, 1),
     ),
-    max_similarity_score: float = typer.Option(
-        1.0,
+    max_similarity_score: Optional[float] = typer.Option(
+        None,
         "--max-similarity-score",
         "-S",
         help="Normalised edit distance (read to reference similarity) maximum. Number between 0-1.",
-        callback=lambda x: cli_utils.is_in_range(x, 0, 1),
+        callback=lambda x: cli_utils.is_in_range_or_None(x, 0, 1),
     ),
-    min_edit_dist: int = typer.Option(
-        0,
+    min_edit_dist: Optional[int] = typer.Option(
+        None,
+        # 0,
         "--min-edit-dist",
         "-e",
-        help="Minimum edit distance (read to reference similarity). Number between 0-10.",
-        callback=lambda x: cli_utils.is_in_range(x, 0, 10),
+        help="Minimum edit distance (read to reference similarity). Positive integer.",
+        callback=lambda x: cli_utils.is_positive_int_or_None(x),
     ),
-    max_edit_dist: int = typer.Option(
-        10,
+    max_edit_dist: Optional[int] = typer.Option(
+        None,
+        # 10,
         "--max-edit-dist",
         "-E",
-        help="Maximum edit distance (read to reference similarity). Number between 0-10.",
-        callback=lambda x: cli_utils.is_in_range(x, 0, 10),
+        help="Maximum edit distance (read to reference similarity). Positive integer.",
+        callback=lambda x: cli_utils.is_positive_int_or_None(x),
     ),
     min_mapping_quality: int = typer.Option(
         0,
@@ -223,10 +227,9 @@ def create_config(
             "names": names,
             "nodes": nodes,
             "acc2tax": acc2tax,
-            "min_similarity_score": min_similarity_score,
-            "max_similarity_score": max_similarity_score,
-            "min_edit_dist": min_edit_dist,
-            "max_edit_dist": max_edit_dist,
+            **cli_utils.set_min_max_similarity_score_edit_dist(
+                min_similarity_score, max_similarity_score, min_edit_dist, max_edit_dist
+            ),
             "min_mapping_quality": min_mapping_quality,
             "lca_rank": lca_rank.value,  # important to get string
             "max_position": max_position,

src/metaDMG/cli/cli_utils.py

@@ -1,5 +1,5 @@
 from enum import Enum
-from typing import Iterable
+from typing import Iterable, Optional
 
 import typer
 from click import Context, Group
@@ -59,7 +59,9 @@ def version_callback(value: bool):
         raise typer.Exit()
 
 
-def is_in_range(x: float, val_min: float, val_max: float) -> float:
+def is_in_range_or_None(
+    x: Optional[float], val_min: float, val_max: float
+) -> Optional[float]:
     """Confirms that x is val_min <= x <= val_max
 
     Parameters
@@ -81,13 +83,43 @@ def is_in_range(x: float, val_min: float, val_max: float) -> float:
         If x is outside bounds
     """
 
+    if x is None:
+        return x
+
     if x < val_min or val_max < x:
         raise typer.BadParameter(
             f"x has to be between {val_min} and {val_max}. Got: {x}"
         )
     return x
 
 
+def is_positive_int_or_None(x: Optional[int]) -> Optional[int]:
+    """Confirms that x is 0 <= x
+
+    Parameters
+    ----------
+    x
+        Value to check
+
+    Returns
+    -------
+        Confirmed value
+
+    Raises
+    ------
+    typer.BadParameter
+        If x is outside bounds
+    """
+
+    if x is None:
+        return x
+
+    if x < 0:
+        raise typer.BadParameter(f"x has to be positive. Got: {x}")
+
+    return x
+
+
 #%%
 class RANKS(str, Enum):
     "Ranks allowed in the LCA"
@@ -107,3 +139,59 @@ class DAMAGE_MODE(str, Enum):
 
 
 #%%
+
+
+def set_min_max_similarity_score_edit_dist(
+    min_similarity_score: Optional[float],
+    max_similarity_score: Optional[float],
+    min_edit_dist: Optional[int],
+    max_edit_dist: Optional[int],
+) -> dict[str, float]:
+
+    if any([min_similarity_score, max_similarity_score]) and any(
+        [min_edit_dist, max_edit_dist]
+    ):
+        raise typer.BadParameter(
+            f"You cannot use both similarity scores and edit distances at the same time."
+        )
+
+    # edit distances
+    if any([min_edit_dist, max_edit_dist]):
+
+        if all([min_edit_dist, max_edit_dist]):
+
+            if min_edit_dist > max_edit_dist:
+                raise typer.BadParameter(
+                    f"min-edit-dist ({min_edit_dist}) "
+                    f"has to be lower than max-edit-dist ({max_edit_dist})"
+                )
+
+            return {
+                "min_edit_dist": min_edit_dist,
+                "max_edit_dist": max_edit_dist,
+            }
+
+        else:
+            raise typer.BadParameter(
+                f"If using (absolute) edit distances, you have to set "
+                "both `min_edit_dist` and `max_edit_dist`."
+            )
+
+    # similarity scores
+
+    if min_similarity_score is None:
+        min_similarity_score = 0.95
+
+    if max_similarity_score is None:
+        max_similarity_score = 1.0
+
+    if min_similarity_score > max_similarity_score:
+        raise typer.BadParameter(
+            f"min-similarity-score ({min_similarity_score}) "
+            f"has to be lower than max-similarity-score ({max_similarity_score})"
+        )
+
+    return {
+        "min_similarity_score": min_similarity_score,
+        "max_similarity_score": max_similarity_score,
+    }

src/metaDMG/fit/serial.py

@@ -57,9 +57,17 @@ def data_dir(config: Config, name, suffix="parquet"):
 #%%
 
 
+def _exists_in_config(s: str, key: str, config: Config) -> str:
+    return f"-{s} {config[key]}" if key in config else ""
+
+
 def get_LCA_command(config: Config) -> str:
     outnames = config["path_tmp"] / config["sample"]
     lca_rank = f"-lca_rank {config['lca_rank']}" if config["lca_rank"] != "" else ""
+    simscorelow = _exists_in_config("simscorelow", "min_similarity_score", config)
+    simscorehigh = _exists_in_config("simscorehigh", "max_similarity_score", config)
+    editdistmin = _exists_in_config("editdistmin", "min_edit_dist", config)
+    editdistmax = _exists_in_config("editdistmax", "max_edit_dist", config)
 
     command = (
         f"{config['metaDMG_cpp']} lca "
@@ -68,10 +76,10 @@ def get_LCA_command(config: Config) -> str:
         f"-names {config['names']} "
         f"-nodes {config['nodes']} "
         f"-acc2tax {config['acc2tax']} "
-        f"-simscorelow {config['min_similarity_score']} "
-        f"-simscorehigh {config['max_similarity_score']} "
-        f"-editdistmin {config['min_edit_dist']} "
-        f"-editdistmax {config['max_edit_dist']} "
+        f"{simscorelow} "
+        f"{simscorehigh} "
+        f"{editdistmin} "
+        f"{editdistmax} "
         f"{lca_rank} "
         f"-minmapq {config['min_mapping_quality']} "
         f"-howmany {config['max_position']} "