From 8a906fcad075438b56d6c5aa34d28d8347ef0e0a Mon Sep 17 00:00:00 2001
From: UncleCode <unclecode@kidocode.com>
Date: Tue, 29 Jul 2025 19:56:27 +0800
Subject: [PATCH] fix(dependencies): Update and clean up package versions in
 pyproject.toml, the bundle size will be much smaller.

---
 pyproject.toml | 29 +++++++++++++----------------
 1 file changed, 13 insertions(+), 16 deletions(-)

diff --git a/pyproject.toml b/pyproject.toml
index a582d430..366caf65 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -30,8 +30,6 @@ dependencies = [
     "pydantic>=2.10",
     "pyOpenSSL>=24.3.0",
     "psutil>=6.1.1",
-    "nltk>=3.9.1",
-    "playwright",
     "rich>=13.9.4",
     "cssselect>=1.2.0",
     "httpx>=0.27.2",
@@ -44,7 +42,6 @@ dependencies = [
     "brotli>=1.1.0",
     "humanize>=4.10.0",
     "lark>=1.2.2",
-    "sentence-transformers>=2.2.0",
     "alphashape>=1.3.1",
     "shapely>=2.0.0"
 ]
@@ -60,20 +57,20 @@ classifiers = [
 ]
 
 [project.optional-dependencies]
-pdf = ["PyPDF2"]  
-torch = ["torch", "nltk", "scikit-learn"]
-transformer = ["transformers", "tokenizers"]
-cosine = ["torch", "transformers", "nltk"]
-sync = ["selenium"]
+pdf = ["pypdf>=3.0.0"]  # PyPDF2 is deprecated, use pypdf instead
+torch = ["torch>=2.0.0", "nltk>=3.9.1", "scikit-learn>=1.3.0"]
+transformer = ["transformers>=4.34.0", "tokenizers>=0.15.0", "sentence-transformers>=2.2.0"]
+cosine = ["torch>=2.0.0", "transformers>=4.34.0", "nltk>=3.9.1", "sentence-transformers>=2.2.0"]
+sync = ["selenium>=4.0.0"]
 all = [
-    "PyPDF2",
-    "torch",
-    "nltk",
-    "scikit-learn",
-    "transformers",
-    "tokenizers",
-    "selenium",
-    "PyPDF2"  
+    "pypdf>=3.0.0",
+    "torch>=2.0.0",
+    "nltk>=3.9.1",
+    "scikit-learn>=1.3.0",
+    "transformers>=4.34.0",
+    "tokenizers>=0.15.0",
+    "sentence-transformers>=2.2.0",
+    "selenium>=4.0.0"
 ]
 
 [project.scripts]