Merge pull request #16 from cthoyt/simplify-getting-started

update constructor of base ensemble to run w/o arguments (using same default as CLI), expose BaseEnsemble

Author: sfluegel05

README.md

@@ -75,15 +75,11 @@ python -m chebifier predict --help
 You can also use the package programmatically:
 
 ```python
-from chebifier.ensemble.base_ensemble import BaseEnsemble
-import yaml
+from chebifier import BaseEnsemble
 
-# Load configuration from YAML file
-with open('configs/example_config.yml', 'r') as f:
-    config = yaml.safe_load(f)
-
-# Instantiate ensemble model
-ensemble = BaseEnsemble(config)
+# Instantiate ensemble model. If desired, can pass
+# a path to a configuration, like 'configs/example_config.yml'
+ensemble = BaseEnsemble()
 
 # Make predictions
 smiles_list = ["CC(=O)OC1=CC=CC=C1C(=O)O", "C1=CC=C(C=C1)C(=O)O"]

chebifier/__init__.py

@@ -1,6 +1,11 @@
 # Note: The top-level package __init__.py runs only once,
 # even if multiple subpackages are imported later.
 
-from ._custom_cache import PerSmilesPerModelLRUCache
+from ._custom_cache import PerSmilesPerModelLRUCache, modelwise_smiles_lru_cache
+from .ensemble.base_ensemble import BaseEnsemble
 
-modelwise_smiles_lru_cache = PerSmilesPerModelLRUCache(max_size=100)
+__all__ = [
+    "BaseEnsemble",
+    "PerSmilesPerModelLRUCache",
+    "modelwise_smiles_lru_cache",
+]

chebifier/_custom_cache.py

@@ -6,6 +6,11 @@
 from functools import wraps
 from typing import Any, Callable
 
+__all__ = [
+    "PerSmilesPerModelLRUCache",
+    "modelwise_smiles_lru_cache",
+]
+
 
 class PerSmilesPerModelLRUCache:
     """
@@ -206,3 +211,6 @@ def _load_cache(self) -> None:
                         self._cache = loaded
             except Exception as e:
                 print(f"[Cache Load Error] {e}")
+
+
+modelwise_smiles_lru_cache = PerSmilesPerModelLRUCache(max_size=100)

chebifier/cli.py

@@ -1,7 +1,4 @@
-import importlib.resources
-
 import click
-import yaml
 
 from chebifier.model_registry import ENSEMBLES
 
@@ -72,43 +69,10 @@ def predict(
     resolve_inconsistencies=True,
 ):
     """Predict ChEBI classes for SMILES strings using an ensemble model."""
-    # Load configuration from YAML file
-    if not ensemble_config:
-        print("Using default ensemble configuration")
-        with (
-            importlib.resources.files("chebifier")
-            .joinpath("ensemble.yml")
-            .open("r") as f
-        ):
-            config = yaml.safe_load(f)
-    else:
-        print(f"Loading ensemble configuration from {ensemble_config}")
-        with open(ensemble_config, "r") as f:
-            config = yaml.safe_load(f)
-
-    with (
-        importlib.resources.files("chebifier")
-        .joinpath("model_registry.yml")
-        .open("r") as f
-    ):
-        model_registry = yaml.safe_load(f)
-
-    new_config = {}
-    for model_name, entry in config.items():
-        if "load_model" in entry:
-            if entry["load_model"] not in model_registry:
-                raise ValueError(
-                    f"Model {entry['load_model']} not found in model registry. "
-                    f"Available models are: {','.join(model_registry.keys())}."
-                )
-            new_config[model_name] = {**model_registry[entry["load_model"]], **entry}
-        else:
-            new_config[model_name] = entry
-    config = new_config
 
     # Instantiate ensemble model
     ensemble = ENSEMBLES[ensemble_type](
-        config,
+        ensemble_config,
         chebi_version=chebi_version,
         resolve_inconsistencies=resolve_inconsistencies,
     )

chebifier/ensemble/base_ensemble.py

@@ -1,32 +1,60 @@
+import importlib
 import os
 import time
+from pathlib import Path
+from typing import Union
 
 import torch
 import tqdm
+import yaml
 
 from chebifier.check_env import check_package_installed
 from chebifier.hugging_face import download_model_files
 from chebifier.inconsistency_resolution import PredictionSmoother
 from chebifier.prediction_models.base_predictor import BasePredictor
-from chebifier.utils import get_disjoint_files, load_chebi_graph
+from chebifier.utils import (
+    get_default_configs,
+    get_disjoint_files,
+    load_chebi_graph,
+    process_config,
+)
 
 
 class BaseEnsemble:
     def __init__(
         self,
-        model_configs: dict,
+        model_configs: Union[str, Path, dict, None] = None,
         chebi_version: int = 241,
         resolve_inconsistencies: bool = True,
     ):
         # Deferred Import: To avoid circular import error
         from chebifier.model_registry import MODEL_TYPES
 
+        # Load configuration from YAML file
+        if not model_configs:
+            config = get_default_configs()
+        elif isinstance(model_configs, dict):
+            config = model_configs
+        else:
+            print(f"Loading ensemble configuration from {model_configs}")
+            with open(model_configs, "r") as f:
+                config = yaml.safe_load(f)
+
+        with (
+            importlib.resources.files("chebifier")
+            .joinpath("model_registry.yml")
+            .open("r") as f
+        ):
+            model_registry = yaml.safe_load(f)
+
+        processed_configs = process_config(config, model_registry)
+
         self.chebi_graph = load_chebi_graph()
         self.disjoint_files = get_disjoint_files()
 
         self.models = []
         self.positive_prediction_threshold = 0.5
-        for model_name, model_config in model_configs.items():
+        for model_name, model_config in processed_configs.items():
             model_cls = MODEL_TYPES[model_config["type"]]
             if "hugging_face" in model_config:
                 hugging_face_kwargs = download_model_files(model_config["hugging_face"])

chebifier/inconsistency_resolution.py

@@ -1,8 +1,9 @@
 import csv
 import os
-import torch
 from pathlib import Path
 
+import torch
+
 
 def get_disjoint_groups(disjoint_files):
     if disjoint_files is None:

chebifier/model_registry.py

@@ -4,15 +4,15 @@
     WMVwithPPVNPVEnsemble,
 )
 from chebifier.prediction_models import (
+    ChEBILookupPredictor,
     ChemlogPeptidesPredictor,
     ElectraPredictor,
     ResGatedPredictor,
-    ChEBILookupPredictor,
 )
 from chebifier.prediction_models.c3p_predictor import C3PPredictor
 from chebifier.prediction_models.chemlog_predictor import (
-    ChemlogXMolecularEntityPredictor,
     ChemlogOrganoXCompoundPredictor,
+    ChemlogXMolecularEntityPredictor,
 )
 
 ENSEMBLES = {

chebifier/prediction_models/base_predictor.py

@@ -1,7 +1,7 @@
 import json
 from abc import ABC
 
-from chebifier import modelwise_smiles_lru_cache
+from .._custom_cache import modelwise_smiles_lru_cache
 
 
 class BasePredictor(ABC):

chebifier/prediction_models/chemlog_predictor.py

@@ -2,8 +2,8 @@
 
 import tqdm
 
-from .base_predictor import BasePredictor
 from .. import modelwise_smiles_lru_cache
+from .base_predictor import BasePredictor
 
 AA_DICT = {
     "A": "L-alanine",

chebifier/utils.py

@@ -1,10 +1,13 @@
+import importlib.resources
 import os
+import pickle
 
+import fastobo
 import networkx as nx
 import requests
-import fastobo
+import yaml
+
 from chebifier.hugging_face import download_model_files
-import pickle
 
 
 def load_chebi_graph(filename=None):
@@ -123,9 +126,27 @@ def get_disjoint_files():
     return disjoint_files
 
 
-if __name__ == "__main__":
-    # chebi_graph = build_chebi_graph(chebi_version=241)
-    # save the graph to a file
-    # pickle.dump(chebi_graph, open("chebi_graph.pkl", "wb"))
-    chebi_graph = load_chebi_graph()
-    print(chebi_graph)
+def get_default_configs():
+    default_config_name = "ensemble.yml"
+    print(f"Using default ensemble configuration from {default_config_name}")
+    with (
+        importlib.resources.files("chebifier")
+        .joinpath(default_config_name)
+        .open("r") as f
+    ):
+        return yaml.safe_load(f)
+
+
+def process_config(config, model_registry):
+    new_config = {}
+    for model_name, entry in config.items():
+        if "load_model" in entry:
+            if entry["load_model"] not in model_registry:
+                raise ValueError(
+                    f"Model {entry['load_model']} not found in model registry. "
+                    f"Available models are: {','.join(model_registry.keys())}."
+                )
+            new_config[model_name] = {**model_registry[entry["load_model"]], **entry}
+        else:
+            new_config[model_name] = entry
+    return new_config