Feature/settings

pyproject.toml

@@ -9,6 +9,7 @@ dependencies = [
     "loguru>=0.7.3",
     "mdanalysis>=2.9.0",
     "pydantic>=2.11.3",
+    "pydantic-settings>=2.12.0",
     "streamlit>=1.50.0",
 ]
 

src/grodecoder/cli/__init__.py

@@ -3,6 +3,7 @@
 from ..main import main as grodecoder_main
 from .args import Arguments as CliArgs
 from .args import CoordinatesFile, StructureFile
+from ..settings import get_settings
 from ..logging import setup_logging
 
 
@@ -11,7 +12,6 @@
 @click.argument("coordinates_file", type=CoordinatesFile, required=False)
 @click.option(
     "--bond-threshold",
-    default=5.0,
     type=float,
     help="Threshold for interchain bond detection (default: 5 Å)",
 )
@@ -29,12 +29,15 @@ def cli(**kwargs):
     args = CliArgs(
         structure_file=kwargs["structure_file"],
         coordinates_file=kwargs["coordinates_file"],
+        bond_threshold=kwargs["bond_threshold"],
         no_atom_ids=kwargs["no_atom_ids"],
         print_to_stdout=kwargs["stdout"],
     )
 
+    get_settings().debug = kwargs["verbose"]
+
     logfile = args.get_log_filename()
-    setup_logging(logfile, kwargs["verbose"])
+    setup_logging(logfile)
     grodecoder_main(args)
 
 

src/grodecoder/cli/args.py

@@ -59,14 +59,14 @@ class Arguments:
     Attrs:
         structure_file (Path): Path to the structure file.
         coordinates_file (Path): Path to the coordinates file.
-        bond_threshold (float): Threshold for interchain bond detection.
+        bond_threshold (float | None): Threshold for interchain bond detection.
         no_atom_ids (bool): If True, use compact serialization (no atom indices).
         print_to_stdout (bool): Whether to output results to stdout.
     """
 
     structure_file: StructureFile
     coordinates_file: CoordinatesFile | None = None
-    bond_threshold: float = 5.0
+    bond_threshold: float | None = None
     no_atom_ids: bool = True
     print_to_stdout: bool = False
 

src/grodecoder/core.py

@@ -7,26 +7,43 @@
 from .io import read_universe
 from .models import Decoded
 from .toputils import guess_resolution
+from .settings import get_settings
 
 
 def _now() -> str:
     """Returns the current date and time formatted string."""
     return datetime.now().strftime("%Y-%m-%d %H:%M:%S")
 
 
-def decode(universe: UniverseLike, bond_threshold: float = 5.0) -> Decoded:
+def decode(universe: UniverseLike) -> Decoded:
     """Decodes the universe into an inventory of segments."""
+
+    settings = get_settings()
+
+    resolution = guess_resolution(universe, cutoff_distance=settings.resolution_detection.distance_cutoff)
+    logger.info(f"Guessed resolution: {resolution}")
+
+    # Guesses the chain dection distance cutoff if not provided by the user.
+    chain_detection_settings = get_settings().chain_detection
+
+    if chain_detection_settings.distance_cutoff.is_set():
+        value = chain_detection_settings.distance_cutoff.get()
+        logger.debug(f"chain detection: using user-defined value: {value:.2f}")
+    else:
+        logger.debug("chain detection: guessing distance cutoff based on resolution")
+        chain_detection_settings.distance_cutoff.guess(resolution)
+
+    distance_cutoff = chain_detection_settings.distance_cutoff.get()
+
     return Decoded(
-        inventory=identify(universe, bond_threshold=bond_threshold),
-        resolution=guess_resolution(universe),
+        inventory=identify(universe, bond_threshold=distance_cutoff),
+        resolution=resolution,
     )
 
 
-def decode_structure(
-    structure_path: PathLike, coordinates_path: PathLike | None = None, bond_threshold: float = 5.0
-) -> Decoded:
+def decode_structure(structure_path: PathLike, coordinates_path: PathLike | None = None) -> Decoded:
     """Reads a structure file and decodes it into an inventory of segments."""
     universe = read_universe(structure_path, coordinates_path)
     assert universe.atoms is not None  # required by type checker for some reason
     logger.debug(f"Universe has {len(universe.atoms):,d} atoms")
-    return decode(universe, bond_threshold=bond_threshold)
+    return decode(universe)

src/grodecoder/identifier.py

@@ -28,6 +28,9 @@ def identify_small_molecule(
         residue = SmallMolecule(
             atoms=selection, description=definition.description, molecular_type=molecular_type
         )
+        logger.debug(
+            f"identified small molecule {residue.description}: {len(selection.residues)} residues, {len(selection.atoms)} atoms"
+        )
         counts.append(residue)
     return counts
 
@@ -66,18 +69,25 @@ def _select_protein(universe: UniverseLike) -> AtomGroup:
     selection_str = f"resname {' '.join(protein_residue_names)}"
 
     # Exclude methanol residues from the selection.
+    logger.debug("excluding possible methanol residues (MET) from protein")
     methanol = _find_methanol(universe)
     if methanol:
         selection_str += f" and not index {' '.join(map(str, methanol))}"
 
-    return universe.select_atoms(selection_str)
+    logger.debug("selecting protein")
+    protein = universe.select_atoms(selection_str)
+    logger.debug("selecting protein - done")
+    return protein
 
 
 def _select_nucleic(universe: UniverseLike) -> AtomGroup:
     """Selects the nucleic acid atoms from the universe."""
     nucleic_acid_residue_names = DB.get_nucleotide_names()
     selection_str = f"resname {' '.join(nucleic_acid_residue_names)}"
-    return universe.select_atoms(selection_str)
+    logger.debug("selecting nucleic")
+    nucleic = universe.select_atoms(selection_str)
+    logger.debug("selecting nucleic - done")
+    return nucleic
 
 
 def _iter_chains(atoms: AtomGroup, bond_threshold: float = 5.0) -> Iterator[AtomGroup]:
@@ -87,8 +97,14 @@ def _iter_chains(atoms: AtomGroup, bond_threshold: float = 5.0) -> Iterator[Atom
     """
     if len(atoms) == 0:
         return
-    segments = toputils.detect_chains(atoms, cutoff=bond_threshold)
+    logger.debug(f"detecting segments using cutoff distance {bond_threshold:.2f}")
+    segments = toputils.detect_chains(atoms, cutoff_distance=bond_threshold)
+
+    n_seg_str = f"{len(segments)} segment" + "s" if len(segments) > 1 else ""
+    logger.debug(f"detecting segments - done: found {n_seg_str}")
+
     for start, end in segments:
+        logger.debug(f"yielding segment containing residues {start} to {end}")
         yield atoms.residues[start : end + 1].atoms
 
 
@@ -156,6 +172,7 @@ def _log_identified_molecules(molecules: list[SmallMolecule], label: str) -> Non
 
 def _identify(universe: UniverseLike, bond_threshold: float = 5.0) -> Inventory:
     """Identifies the molecules in the universe."""
+    logger.debug("Residu identification: start")
 
     # Ensure the universe is an AtomGroup.
     universe = universe.select_atoms("all")
@@ -211,8 +228,12 @@ def _identify(universe: UniverseLike, bond_threshold: float = 5.0) -> Inventory:
             )
             unknown_molecules.append(molecule)
 
-    return Inventory(
+    logger.debug("Creating inventory")
+    inventory = Inventory(
         segments=protein + nucleic,
         small_molecules=ions + solvents + lipids + others + unknown_molecules,
         total_number_of_atoms=total_number_of_atoms,
     )
+
+    logger.debug("Residu identification: end")
+    return inventory

src/grodecoder/logging.py

@@ -6,13 +6,22 @@
 
 from loguru import logger
 
+from .settings import get_settings
 
-def setup_logging(logfile: Path, debug: bool = False):
+
+def setup_logging(logfile: Path):
     """Sets up logging configuration."""
+    debug = get_settings().debug
+
     fmt = "<green>{time:YYYY-MM-DD HH:mm:ss}</green> <level>{level}: {message}</level>"
     level = "DEBUG" if debug else "INFO"
+
     logger.remove()
+
+    # Screen logger.
     logger.add(sys.stderr, level=level, format=fmt, colorize=True)
+
+    # File logger
     logger.add(logfile, level=level, format=fmt, colorize=False, mode="w")
 
     # Sets up loguru to capture warnings (typically MDAnalysis warnings)
@@ -24,6 +33,7 @@ def showwarning(message, *args, **kwargs):
 
 def is_logging_debug() -> bool:
     """Returns True if at least one logging handler is set to level DEBUG."""
+    print("COUCOU", get_logging_level())
     return "DEBUG" in get_logging_level()
 
 

src/grodecoder/main.py

@@ -10,6 +10,7 @@
 from .databases import get_database_version
 from .models import GrodecoderRunOutput
 from .version import get_version
+from .settings import get_settings
 
 if TYPE_CHECKING:
     from .cli.args import Arguments as CliArgs
@@ -27,22 +28,27 @@ def main(args: "CliArgs"):
     structure_path = args.structure_file.path
     coordinates_path = args.coordinates_file.path if args.coordinates_file else None
 
+    # Storing cli arguments into settings.
+    settings = get_settings()
+    settings.chain_detection.distance_cutoff = args.bond_threshold
+    settings.output.atom_ids = not args.no_atom_ids
+
     logger.info(f"Processing structure file: {structure_path}")
 
     # Decoding.
-    decoded = decode_structure(
-        structure_path, coordinates_path=coordinates_path, bond_threshold=args.bond_threshold
-    )
+    decoded = decode_structure(structure_path, coordinates_path=coordinates_path)
 
     output = GrodecoderRunOutput(
         decoded=decoded,
         structure_file_checksum=_get_checksum(structure_path),
         database_version=get_database_version(),
         grodecoder_version=get_version(),
+        input_settings=settings,
     )
 
     # Serialization.
-    serialization_mode = "compact" if args.no_atom_ids else "full"
+    logger.debug("Creating json output")
+    serialization_mode = "full" if settings.output.atom_ids else "compact"
 
     # Updates run time as late as possible.
     output_json = output.model_dump(context={"serialization_mode": serialization_mode})

src/grodecoder/models.py

@@ -1,5 +1,4 @@
 from __future__ import annotations
-from pydantic import model_validator
 
 from enum import StrEnum
 from typing import Protocol
@@ -14,9 +13,11 @@
     computed_field,
     field_serializer,
     model_serializer,
+    model_validator,
 )
 
 from . import toputils
+from .settings import Settings
 
 
 class MolecularResolution(StrEnum):
@@ -194,6 +195,7 @@ class GrodecoderRunOutput(BaseModel):
     structure_file_checksum: str
     database_version: str
     grodecoder_version: str
+    input_settings: Settings
 
 
 # =========================================================================================================