From 42da9317b2bdad37fa08c7f0984a494ab65edf89 Mon Sep 17 00:00:00 2001
From: J08nY
Date: Fri, 10 Feb 2023 20:09:12 +0100
Subject: Drop inheritance from sklearn objects.

It is unused and causes big memory usag spikes on their
import due to their inefficient import practice:
https://github.com/scikit-learn/scikit-learn/issues/25590

We don't actually use their API anywhere (afaik).
---
 src/sec_certs/model/cpe_matching.py    | 5 ++---
 src/sec_certs/model/sar_transformer.py | 9 +++++----
 2 files changed, 7 insertions(+), 7 deletions(-)

(limited to 'src')

diff --git a/src/sec_certs/model/cpe_matching.py b/src/sec_certs/model/cpe_matching.py
index 2c6b3f07..5d08d7af 100644
--- a/src/sec_certs/model/cpe_matching.py
+++ b/src/sec_certs/model/cpe_matching.py
@@ -8,7 +8,6 @@ from typing import Pattern
 
 import spacy
 from rapidfuzz import fuzz
-from sklearn.base import BaseEstimator
 
 from sec_certs import cert_rules, constants
 from sec_certs.sample.cpe import CPE
@@ -17,10 +16,10 @@ from sec_certs.utils.tqdm import tqdm
 logger = logging.getLogger(__name__)
 
 
-class CPEClassifier(BaseEstimator):
+class CPEClassifier:
     """
     Class that can predict CPE matches for certificate instances.
-    Adheres to sklearn BaseEstimator interface.
+    Adheres to sklearn `sklearn.base.BaseEstimator` interface.
     Fit method is called on list of CPEs and build two look-up dictionaries, see description of attributes.
     """
 
diff --git a/src/sec_certs/model/sar_transformer.py b/src/sec_certs/model/sar_transformer.py
index 20ce9dec..a60f7495 100644
--- a/src/sec_certs/model/sar_transformer.py
+++ b/src/sec_certs/model/sar_transformer.py
@@ -3,8 +3,6 @@ from __future__ import annotations
 import logging
 from typing import Dict, Iterable, cast
 
-from sklearn.base import BaseEstimator, TransformerMixin
-
 from sec_certs.sample.cc import CCCertificate
 from sec_certs.sample.sar import SAR, SAR_DICT_KEY
 
@@ -12,10 +10,10 @@ logger = logging.getLogger(__name__)
 
 
 # TODO: Right now we ignore number of ocurrences for final SAR selection. If we keep it this way, we can discard that variable
-class SARTransformer(BaseEstimator, TransformerMixin):
+class SARTransformer:
     """
     Class for transforming SARs defined in st_keywords and report_keywords dictionaries into SAR objects.
-    This class implements sklearn transformer interface, so fit_transform() can be called on it.
+    This class implements `sklearn.base.Transformer` interface, so fit_transform() can be called on it.
     """
 
     def fit(self, certificates: Iterable[CCCertificate]) -> SARTransformer:
@@ -27,6 +25,9 @@ class SARTransformer(BaseEstimator, TransformerMixin):
         """
         return self
 
+    def fit_transform(self, X, y=None, **fit_params):
+        return self.fit(X).transform(X)
+
     def transform(self, certificates: Iterable[CCCertificate]) -> list[set[SAR] | None]:
         """
         Just a wrapper around transform_single_cert() called on an iterable of CCCertificate.
-- 
cgit v1.3.1