From b051d60c99ff4481ac22c36c106e69b7530416ce Mon Sep 17 00:00:00 2001
From: what_in_the_nim <chatcharinsang@gmail.com>
Date: Sat, 10 Jan 2026 11:42:10 +0700
Subject: [PATCH 1/4] refactor: remove numpy dependency

---
 pythainlp/tools/misspell.py | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

diff --git a/pythainlp/tools/misspell.py b/pythainlp/tools/misspell.py
index c940d2929..a5fa84208 100644
--- a/pythainlp/tools/misspell.py
+++ b/pythainlp/tools/misspell.py
@@ -4,7 +4,8 @@
 # SPDX-License-Identifier: Apache-2.0
 from typing import List
 
-import numpy as np
+import math
+import random
 
 THAI_CHARACTERS_WITHOUT_SHIFT = [
     "ผปแอิืทมใฝ",
@@ -126,10 +127,8 @@ def misspell(sentence: str, ratio: float = 0.05):
         # output:
         ภาษาไทยปรากฏครั้งแรกในกุทธศักราช 1727
     """
-    num_misspells = np.floor(len(sentence) * ratio).astype(int)
-    positions = np.random.choice(
-        len(sentence), size=num_misspells, replace=False
-    )
+    num_misspells = math.floor(len(sentence) * ratio)
+    positions = random.sample(range(len(sentence)), k=num_misspells)
 
     # convert strings to array of characters
     misspelled = list(sentence)
@@ -138,7 +137,7 @@ def misspell(sentence: str, ratio: float = 0.05):
         if potential_candidates is None:
             continue
 
-        candidate = np.random.choice(potential_candidates)
+        candidate = random.choice(potential_candidates)
 
         misspelled[pos] = candidate
 

From 901d859d641cf6d4dd73afad218ac50cb38d86ff Mon Sep 17 00:00:00 2001
From: what_in_the_nim <chatcharinsang@gmail.com>
Date: Sat, 10 Jan 2026 11:43:11 +0700
Subject: [PATCH 2/4] refactor: make benchmarks lazy import

---
 pythainlp/cli/benchmark.py | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/pythainlp/cli/benchmark.py b/pythainlp/cli/benchmark.py
index b1b2cb7c4..e242a6162 100644
--- a/pythainlp/cli/benchmark.py
+++ b/pythainlp/cli/benchmark.py
@@ -8,10 +8,8 @@
 import json
 import os
 
-import yaml
 
 from pythainlp import cli
-from pythainlp.benchmarks import word_tokenization
 from pythainlp.tools import safe_print
 
 
@@ -88,6 +86,14 @@ def __init__(self, name, argv):
             % (args.input_file, args.test_file, len(actual))
         )
 
+        try:
+            import yaml
+            from pythainlp.benchmarks import word_tokenization
+        except ImportError:
+            raise ImportError(
+                "Please install the extra dependencies `benchmarks` to use the command by `pip install pythainlp[benchmarks]`"
+            )
+
         df_raw = word_tokenization.benchmark(expected, actual)
 
         columns = [

From c6c178af9606cfb1473b50da26dda02ee7b378c2 Mon Sep 17 00:00:00 2001
From: what_in_the_nim <chatcharinsang@gmail.com>
Date: Sat, 10 Jan 2026 11:53:10 +0700
Subject: [PATCH 3/4] chore: remove unnecessary base dependencies

---
 setup.py | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/setup.py b/setup.py
index 86dc13857..0b3951cd1 100644
--- a/setup.py
+++ b/setup.py
@@ -45,9 +45,6 @@
 requirements = [
     "backports.zoneinfo; python_version<'3.9'",
     "requests>=2.31",
-    PYYAML,
-    PANDAS,
-    NUMPY,
     "tzdata; sys_platform == 'win32'",
 ]
 

From 33ba93bd5de7773657ee4409446d0dabda820cc2 Mon Sep 17 00:00:00 2001
From: Chatcharin Sangbutsarakum
 <67754293+what-in-the-nim@users.noreply.github.com>
Date: Sat, 10 Jan 2026 12:22:14 +0700
Subject: [PATCH 4/4] chore: format message as copilot suggested

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
---
 pythainlp/cli/benchmark.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/pythainlp/cli/benchmark.py b/pythainlp/cli/benchmark.py
index e242a6162..68d59c10d 100644
--- a/pythainlp/cli/benchmark.py
+++ b/pythainlp/cli/benchmark.py
@@ -91,7 +91,7 @@ def __init__(self, name, argv):
             from pythainlp.benchmarks import word_tokenization
         except ImportError:
             raise ImportError(
-                "Please install the extra dependencies `benchmarks` to use the command by `pip install pythainlp[benchmarks]`"
+                "Please install the extra dependencies `benchmarks` to use this command by running `pip install pythainlp[benchmarks]`"
             )
 
         df_raw = word_tokenization.benchmark(expected, actual)