ENH: added parse_blocking_func function that accepts strings as input

If input string gets provided, check if included in blocking map.

emm/helper/blocking_functions.py

@@ -23,6 +23,10 @@
 Please don't modify the function names.
 """
 
+from __future__ import annotations
+
+from typing import Callable
+
 
 def first(x: str) -> str:
     """First character blocking function."""
@@ -37,3 +41,27 @@ def first2(x: str) -> str:
 def first3(x: str) -> str:
     """First two characters blocking function."""
     return x.strip().lower()[:3]
+
+
+BLOCKING_MAP = {"first": first, "first2": first2, "first3": first3}
+
+
+def _parse_blocking_func(input: Callable[[str], str] | str | None = None) -> Callable[[str], str] | None:
+    """Helper function to get blocking function
+
+    Args:
+        input: blocking function or name of existing blocking function
+
+    Returns:
+        blocking function or None
+    """
+    if input is None or callable(input):
+        return input
+    if isinstance(input, str):
+        if input not in BLOCKING_MAP:
+            msg = f"Input {input} is not a recognized blocking function."
+            raise ValueError(msg)
+        return BLOCKING_MAP[input]
+
+    msg = "Input is not None, no string and not callable."
+    raise TypeError(msg)

emm/indexing/pandas_cos_sim_matcher.py

@@ -30,6 +30,7 @@
 from sklearn.base import TransformerMixin
 from sparse_dot_topn import sp_matmul_topn
 
+from emm.helper.blocking_functions import _parse_blocking_func
 from emm.helper.util import groupby
 from emm.indexing.base_indexer import CosSimBaseIndexer
 from emm.indexing.pandas_normalized_tfidf import PandasNormalizedTfidfVectorizer
@@ -52,7 +53,7 @@ def __init__(
         max_features: int | None = None,
         n_jobs: int = 1,
         spark_session: Any | None = None,
-        blocking_func: Callable[[str], str] | None = None,
+        blocking_func: Callable[[str], str] | str | None = None,
         dtype: type[float] = np.float32,
         indexer_id: int | None = None,
     ) -> None:
@@ -99,7 +100,7 @@ def __init__(
         self.dtype = dtype
         self.cos_sim_lower_bound = cos_sim_lower_bound
         self.partition_size = partition_size
-        self.blocking_func = blocking_func
+        self.blocking_func = _parse_blocking_func(blocking_func)
         self.n_jobs = n_jobs if n_jobs != -1 else multiprocessing.cpu_count()
         self.spark_session = spark_session
         # attributes below are set during fit

emm/indexing/spark_cos_sim_matcher.py

@@ -37,6 +37,7 @@
 from pyspark.sql.types import ArrayType, FloatType, LongType, StringType, StructField, StructType
 from sparse_dot_topn import awesome_cossim_topn
 
+from emm.helper.blocking_functions import _parse_blocking_func
 from emm.helper.spark_custom_reader_writer import SparkReadable, SparkWriteable
 from emm.helper.spark_utils import set_spark_job_group
 from emm.indexing.base_indexer import BaseIndexer, CosSimBaseIndexer
@@ -74,7 +75,7 @@ def __init__(
         num_candidates: int = 2,
         cos_sim_lower_bound: float = 0.5,
         max_features: int = 2**25,
-        blocking_func: Callable[[str], str] | None = None,
+        blocking_func: Callable[[str], str] | str | None = None,
         streaming: bool = False,
         indexer_id: int | None = None,
         keep_all_cols: bool = False,
@@ -146,7 +147,7 @@ def __init__(
                 num_candidates=parameters.get("num_candidates", 10),
                 cos_sim_lower_bound=parameters["cos_sim_lower_bound"],
                 streaming=parameters["streaming"],
-                blocking_func=parameters["blocking_func"],
+                blocking_func=_parse_blocking_func(parameters["blocking_func"]),
                 indexer_id=parameters["indexer_id"],
                 n_threads=parameters.get("n_threads", 1),
             )