658 lines
17 KiB
Markdown
Executable File
658 lines
17 KiB
Markdown
Executable File
# SIME - Structure-Informed Macrolide Expansion
|
||
|
||
SIME 是一个用于大环内酯类化合物结构扩展和抗菌活性预测的工具。
|
||
|
||
## 目录
|
||
|
||
- [原有功能](#原有功能)
|
||
- [MolE 抗菌活性预测](#mole-抗菌活性预测)
|
||
- [快速开始](#快速开始)
|
||
- [安装依赖](#安装依赖)
|
||
- [使用方法](#使用方法)
|
||
- [输出说明](#输出说明)
|
||
- [项目结构](#项目结构)
|
||
- [常见问题](#常见问题)
|
||
|
||
---
|
||
|
||
## 原有功能
|
||
|
||
SIME 提供大环内酯类化合物的结构设计和合成路径分析功能。
|
||
|
||
---
|
||
|
||
## MolE 抗菌活性预测
|
||
|
||
本工具集成了 MolE(Molecular Embeddings)模型,可以预测小分子的广谱抗菌活性。
|
||
|
||
### 快速开始
|
||
|
||
#### 使用 uv(推荐)
|
||
|
||
```bash
|
||
# 1. 创建虚拟环境(Python 3.12)
|
||
uv venv --python 3.12 --seed .venv
|
||
|
||
# 2. 激活环境
|
||
source .venv/bin/activate # Linux/Mac
|
||
# 或
|
||
.venv\Scripts\activate # Windows
|
||
|
||
# 3. 使用 uv 安装依赖
|
||
uv pip install -r requirements-mole.txt
|
||
|
||
# 4. 验证安装
|
||
python verify_setup.py
|
||
|
||
# 5. 运行预测
|
||
python utils/mole_predictor.py Data/fragment/Frags-Enamine-18M.csv
|
||
```
|
||
|
||
#### 使用 pyproject.toml 配置(uv 推荐)
|
||
|
||
项目提供了两个环境配置:
|
||
|
||
1. **SIME 原始环境** - 用于大环内酯结构设计
|
||
|
||
```bash
|
||
# 使用 uv 创建默认环境
|
||
uv sync
|
||
```
|
||
|
||
2. **MolE 预测环境** - 用于抗菌活性预测
|
||
|
||
```bash
|
||
# 使用 uv 创建 MolE 环境
|
||
uv sync --extra mole
|
||
```
|
||
|
||
#### 使用 pixi 配置(conda 用户推荐)
|
||
|
||
如果你使用 conda 或需要更好的包管理,可以使用 pixi:
|
||
|
||
```bash
|
||
# 安装 pixi(如果还没有)
|
||
curl -fsSL https://pixi.sh/install.sh | bash
|
||
|
||
# 创建 SIME 原始环境
|
||
pixi install
|
||
|
||
# 创建 MolE 预测环境
|
||
pixi install -e mole
|
||
|
||
# 激活 MolE 环境
|
||
pixi shell -e mole
|
||
|
||
# 在 pixi 环境中运行预测
|
||
pixi run -e mole predict Data/fragment/test_100.csv
|
||
```
|
||
|
||
### 安装依赖
|
||
|
||
#### 方法 1: 使用 uv(推荐)
|
||
|
||
```bash
|
||
# 创建虚拟环境
|
||
uv venv --python 3.12 .venv
|
||
source .venv/bin/activate
|
||
|
||
# 安装依赖
|
||
uv pip install -r requirements-mole.txt
|
||
```
|
||
|
||
#### 方法 2: 使用 pixi
|
||
|
||
```bash
|
||
# 创建虚拟环境
|
||
pixi init
|
||
|
||
# 基础环境
|
||
pixi add python=3.12
|
||
|
||
# nvidia cuda工具链
|
||
pixi workspace channel add nvidia
|
||
pixi add nvidia::cuda-toolkit=12.8
|
||
|
||
# 科学计算 安装 pandas 会自动安装上 numpy
|
||
pixi add
|
||
|
||
# torch-geometric
|
||
pixi add conda-forge::pandas conda-forge::torch-geometric conda-forge::xgboost conda-forge::pyyaml conda-forge::rdkit conda-forge::pip conda-forge::click conda-forge::openpyxl
|
||
|
||
# PyTorch相关(指定通道)
|
||
# 1. 添加 pytorch 频道 conda 太旧改为使用 pypi
|
||
# pixi workspace channel add pytorch
|
||
# pixi add pytorch::pytorch=2.6 pytorch::pytorch-cuda=12.4
|
||
pixi add --pypi torch==2.8.0 torchvision==0.23.0 torchaudio==2.8.0
|
||
|
||
# 然后在 pixi.toml 中手动编辑为:
|
||
[pypi-dependencies]
|
||
torch = { version = "==2.8.0", index = "https://download.pytorch.org/whl/cu128" }
|
||
torchvision = { version = "==0.23.0", index = "https://download.pytorch.org/whl/cu128" }
|
||
torchaudio = { version = "==2.8.0", index = "https://download.pytorch.org/whl/cu128" }
|
||
|
||
# 安装依赖
|
||
pixi install
|
||
|
||
# 激活
|
||
pixi shell
|
||
```
|
||
|
||
不同机器使用配置方式:
|
||
|
||
```bash
|
||
# 在 Linux GPU 机器上安装和运行
|
||
pixi install --environment gpu
|
||
pixi run --environment gpu <your-task>
|
||
|
||
# 在 macOS 或 CPU 机器上安装和运行
|
||
pixi install --environment cpu
|
||
pixi run --environment cpu <your-task>
|
||
|
||
# 或使用默认环境(CPU)
|
||
pixi install
|
||
pixi run <your-task>
|
||
```
|
||
|
||
#### RDKit 安装建议
|
||
|
||
RDKit 推荐使用 conda 安装:
|
||
|
||
```bash
|
||
conda install -c conda-forge rdkit
|
||
```
|
||
|
||
### 使用方法
|
||
|
||
#### 1. 命令行使用
|
||
|
||
**基本用法:**
|
||
|
||
```bash
|
||
# 预测 CSV 文件(仅聚合结果)
|
||
python utils/mole_predictor.py input.csv
|
||
|
||
# 包含 40 种菌株的详细预测数据
|
||
python utils/mole_predictor.py input.csv --include-strain-predictions
|
||
|
||
# 指定输出路径
|
||
python utils/mole_predictor.py input.csv output.csv
|
||
|
||
# 自定义列名
|
||
python utils/mole_predictor.py input.csv output.csv \
|
||
--smiles-column SMILES \
|
||
--id-column compound_id
|
||
|
||
# 使用 GPU 加速
|
||
python utils/mole_predictor.py input.csv --device cuda:0
|
||
|
||
# 调整批次大小和工作进程
|
||
python utils/mole_predictor.py input.csv \
|
||
--batch-size 200 \
|
||
--n-workers 8
|
||
|
||
# 完整示例:包含菌株预测 + GPU 加速
|
||
python utils/mole_predictor.py input.csv output.csv \
|
||
--include-strain-predictions \
|
||
--device cuda:0 \
|
||
--batch-size 200
|
||
```
|
||
|
||
**查看所有选项:**
|
||
|
||
```bash
|
||
python utils/mole_predictor.py --help
|
||
```
|
||
|
||
**预测项目数据:**
|
||
|
||
```bash
|
||
# 预测 Frags-Enamine-18M.csv
|
||
# 创建测试文件(前 1001 行,包含表头)
|
||
head -1001 Data/fragment/Frags-Enamine-18M.csv > Data/fragment/test_1000.csv
|
||
|
||
# 测试命令 - 保守参数
|
||
nohup pixi run python utils/mole_predictor.py \
|
||
Data/fragment/test_1000.csv \
|
||
Data/fragment/test_1000_predicted.csv \
|
||
--device cuda:0 \
|
||
--batch-size 100 \
|
||
--n-workers 8 \
|
||
> Data/fragment/mole_test_1000.log 2>&1 &
|
||
|
||
# 查看日志
|
||
tail -f Data/fragment/mole_test_1000.log
|
||
|
||
# 预测 GDB11-27M.csv
|
||
python utils/mole_predictor.py Data/fragment/GDB11-27M.csv
|
||
```
|
||
|
||
#### 2. Python API 使用
|
||
|
||
**预测单个文件:**
|
||
|
||
```python
|
||
from utils.mole_predictor import predict_csv_file
|
||
|
||
# 基本使用(仅聚合结果)
|
||
df_result = predict_csv_file(
|
||
input_path="Data/fragment/Frags-Enamine-18M.csv",
|
||
output_path="results/predictions.csv",
|
||
smiles_column="smiles",
|
||
batch_size=100,
|
||
device="auto"
|
||
)
|
||
|
||
# 包含 40 种菌株的详细预测数据
|
||
df_result_with_strains = predict_csv_file(
|
||
input_path="Data/fragment/Frags-Enamine-18M.csv",
|
||
output_path="results/predictions_with_strains.csv",
|
||
smiles_column="smiles",
|
||
batch_size=100,
|
||
device="auto",
|
||
include_strain_predictions=True # 启用菌株级别预测
|
||
)
|
||
|
||
# 查看结果
|
||
print(f"总分子数: {len(df_result)}")
|
||
print(f"广谱分子数: {df_result['broad_spectrum'].sum()}")
|
||
|
||
# 如果包含菌株预测,数据行数会增加(每个分子 40 行)
|
||
if 'strain_name' in df_result_with_strains.columns:
|
||
print(f"包含菌株预测的总行数: {len(df_result_with_strains)}")
|
||
print(f"菌株数: {df_result_with_strains['strain_name'].nunique()}")
|
||
```
|
||
|
||
**批量预测多个文件:**
|
||
|
||
```python
|
||
from utils.mole_predictor import predict_multiple_files
|
||
|
||
input_files = [
|
||
"Data/fragment/Frags-Enamine-18M.csv",
|
||
"Data/fragment/GDB11-27M.csv"
|
||
]
|
||
|
||
results = predict_multiple_files(
|
||
input_paths=input_files,
|
||
output_dir="results/",
|
||
smiles_column="smiles",
|
||
batch_size=100,
|
||
device="auto"
|
||
)
|
||
```
|
||
|
||
**直接使用预测器:**
|
||
|
||
```python
|
||
from models import (
|
||
ParallelBroadSpectrumPredictor,
|
||
PredictionConfig,
|
||
MoleculeInput
|
||
)
|
||
|
||
# 创建配置
|
||
config = PredictionConfig(
|
||
batch_size=100,
|
||
device="auto" # 或 "cpu", "cuda:0"
|
||
)
|
||
|
||
# 创建预测器
|
||
predictor = ParallelBroadSpectrumPredictor(config)
|
||
|
||
# 预测单个分子(仅聚合结果)
|
||
molecule = MoleculeInput(smiles="CCO", chem_id="ethanol")
|
||
result = predictor.predict_single(molecule)
|
||
|
||
print(f"化合物ID: {result.chem_id}")
|
||
print(f"广谱抗菌: {result.broad_spectrum}")
|
||
print(f"抗菌得分: {result.apscore_total:.3f}")
|
||
print(f"抑制菌株数: {result.ginhib_total}")
|
||
|
||
# 批量预测(包含 40 种菌株的详细预测)
|
||
smiles_list = ["CCO", "c1ccccc1", "CC(=O)O"]
|
||
chem_ids = ["ethanol", "benzene", "acetic_acid"]
|
||
|
||
results = predictor.predict_batch(
|
||
[MoleculeInput(smiles=s, chem_id=c) for s, c in zip(smiles_list, chem_ids)],
|
||
include_strain_predictions=True # 启用菌株级别预测
|
||
)
|
||
|
||
for r in results:
|
||
print(f"\n{r.chem_id}:")
|
||
print(f" 广谱抗菌: {r.broad_spectrum}")
|
||
print(f" 抗菌得分: {r.apscore_total:.3f}")
|
||
print(f" 抑制菌株数: {r.ginhib_total}")
|
||
|
||
# 访问菌株级别预测数据(DataFrame 格式)
|
||
if r.strain_predictions is not None:
|
||
print(f" 菌株预测数据形状: {r.strain_predictions.shape}")
|
||
print(f" 示例菌株预测:")
|
||
print(r.strain_predictions.head(3))
|
||
|
||
# 强化学习场景:提取特定菌株的预测概率
|
||
strain_probs = r.strain_predictions['antimicrobial_predictive_probability'].values
|
||
print(f" 预测概率向量形状: {strain_probs.shape}") # (40,)
|
||
|
||
# 或转换为类型安全的列表
|
||
strain_list = r.to_strain_predictions_list()
|
||
print(f" 第一个菌株: {strain_list[0].strain_name}")
|
||
print(f" 预测概率: {strain_list[0].antimicrobial_predictive_probability:.6f}")
|
||
```
|
||
|
||
### 输出说明
|
||
|
||
#### 1. 聚合结果(默认输出)
|
||
|
||
预测结果会添加以下 7 个新列:
|
||
|
||
| 列名 | 类型 | 说明 |
|
||
|------|------|------|
|
||
| `apscore_total` | float | 总体抗菌潜力分数(对数尺度,值越大抗菌活性越强) |
|
||
| `apscore_gnegative` | float | 革兰阴性菌抗菌潜力分数 |
|
||
| `apscore_gpositive` | float | 革兰阳性菌抗菌潜力分数 |
|
||
| `ginhib_total` | int | 被抑制的菌株总数(0-40) |
|
||
| `ginhib_gnegative` | int | 被抑制的革兰阴性菌株数 |
|
||
| `ginhib_gpositive` | int | 被抑制的革兰阳性菌株数 |
|
||
| `broad_spectrum` | int | 是否为广谱抗菌(1=是,0=否) |
|
||
|
||
**输出示例**(每个分子一行):
|
||
|
||
```csv
|
||
smiles,chem_id,apscore_total,apscore_gnegative,apscore_gpositive,ginhib_total,ginhib_gnegative,ginhib_gpositive,broad_spectrum
|
||
CCO,mol1,-9.93,-10.17,-9.74,0,0,0,0
|
||
```
|
||
|
||
#### 2. 菌株级别预测(使用 `--include-strain-predictions` 时)
|
||
|
||
启用菌株级别预测后,输出会包含以下额外列(每个分子 40 行):
|
||
|
||
| 列名 | 类型 | 说明 |
|
||
|------|------|------|
|
||
| `pred_id` | str | 预测ID,格式为 `chem_id:strain_name` |
|
||
| `strain_name` | str | 菌株名称(40 种菌株之一) |
|
||
| `antimicrobial_predictive_probability` | float | XGBoost 预测的抗菌概率(0-1) |
|
||
| `no_growth_probability` | float | 预测不抑制的概率(1 - antimicrobial_predictive_probability) |
|
||
| `growth_inhibition` | int | 二值化抑制结果(0=不抑制,1=抑制) |
|
||
| `gram_stain` | str | 革兰染色类型(negative 或 positive) |
|
||
|
||
**输出示例**(每个分子 40 行,对应 40 个菌株):
|
||
|
||
```csv
|
||
smiles,chem_id,apscore_total,...,pred_id,strain_name,antimicrobial_predictive_probability,no_growth_probability,growth_inhibition,gram_stain
|
||
CCO,mol1,-9.93,...,mol1:Akkermansia muciniphila (NT5021),Akkermansia muciniphila (NT5021),0.000102,0.999898,0,negative
|
||
CCO,mol1,-9.93,...,mol1:Bacteroides caccae (NT5050),Bacteroides caccae (NT5050),0.000155,0.999845,0,negative
|
||
...(共 40 行,对应 40 个菌株)
|
||
```
|
||
|
||
**数据使用场景**:
|
||
|
||
1. **强化学习状态表示**:
|
||
```python
|
||
# 提取预测概率作为状态向量
|
||
state_vector = result.strain_predictions['antimicrobial_predictive_probability'].values
|
||
# 形状: (40,) - 可直接用于 RL 环境
|
||
```
|
||
|
||
2. **筛选特定菌株**:
|
||
```python
|
||
# 筛选革兰阴性菌
|
||
gram_negative = result.strain_predictions[
|
||
result.strain_predictions['gram_stain'] == 'negative'
|
||
]
|
||
```
|
||
|
||
3. **可视化分析**:
|
||
```python
|
||
import matplotlib.pyplot as plt
|
||
df = result.strain_predictions
|
||
df.plot(x='strain_name', y='antimicrobial_predictive_probability', kind='bar')
|
||
```
|
||
|
||
#### 40 种测试菌株列表
|
||
|
||
预测涵盖以下 40 种人类肠道菌株:
|
||
|
||
**革兰阴性菌(23 种)**:
|
||
- Akkermansia muciniphila (NT5021)
|
||
- Bacteroides 属: caccae, fragilis (ET/NT), ovatus, thetaiotaomicron, uniformis, vulgatus, xylanisolvens (8 种)
|
||
- Escherichia coli 各亚型 (4 种)
|
||
- Klebsiella pneumoniae (NT5049)
|
||
- 其他肠道革兰阴性菌
|
||
|
||
**革兰阳性菌(17 种)**:
|
||
- Bifidobacterium 属 (3 种)
|
||
- Clostridium 属 (5 种)
|
||
- Enterococcus 属 (2 种)
|
||
- Lactobacillus 属 (3 种)
|
||
- Streptococcus 属 (2 种)
|
||
- 其他肠道革兰阳性菌
|
||
|
||
完整菌株列表详见 `Data/mole/README.md`。
|
||
|
||
#### 广谱抗菌判断标准
|
||
|
||
默认情况下,如果一个分子能抑制 **10 个或更多菌株** (`ginhib_total >= 10`),则被认为是广谱抗菌分子。
|
||
|
||
#### 输出文件位置
|
||
|
||
默认情况下,输出文件会添加 `_predicted` 后缀:
|
||
|
||
- 输入: `Data/fragment/Frags-Enamine-18M.csv`
|
||
- 输出: `Data/fragment/Frags-Enamine-18M_predicted.csv`
|
||
- 输出(含菌株): `Data/fragment/Frags-Enamine-18M_predicted.csv`(每个分子 40 行)
|
||
|
||
#### 数据量说明
|
||
|
||
- **仅聚合结果**:输出行数 = 输入分子数
|
||
- **包含菌株预测**:输出行数 = 输入分子数 × 40
|
||
|
||
---
|
||
|
||
## 项目结构
|
||
|
||
```
|
||
SIME/
|
||
├── models/ # MolE 预测模型
|
||
│ ├── __init__.py
|
||
│ ├── broad_spectrum_predictor.py # 核心预测器
|
||
│ ├── dataset_representation.py # 数据集表示
|
||
│ ├── ginet_concat.py # GIN 神经网络
|
||
│ └── mole_representation.py # MolE 表示生成
|
||
│
|
||
├── utils/
|
||
│ ├── mole_predictor.py # 预测工具脚本
|
||
│ └── ... (其他工具)
|
||
│
|
||
├── Data/
|
||
│ └── fragment/ # 待预测数据
|
||
│ ├── Frags-Enamine-18M.csv
|
||
│ └── GDB11-27M.csv
|
||
│
|
||
├── pyproject.toml # uv 项目配置
|
||
├── requirements.txt # SIME 原始依赖
|
||
├── requirements-mole.txt # MolE 预测依赖
|
||
│
|
||
├── verify_setup.py # 设置验证工具
|
||
├── check_mole_dependencies.py # 依赖检查工具
|
||
└── test_mole_predictor.py # 功能测试
|
||
```
|
||
|
||
---
|
||
|
||
## 依赖说明
|
||
|
||
### SIME 原始依赖 (requirements.txt)
|
||
|
||
用于大环内酯结构设计功能。
|
||
|
||
### MolE 预测依赖 (requirements-mole.txt)
|
||
|
||
用于抗菌活性预测,主要包括:
|
||
|
||
- **深度学习**: torch, torch-geometric
|
||
- **科学计算**: numpy, pandas, scipy
|
||
- **机器学习**: scikit-learn, xgboost
|
||
- **化学信息**: rdkit
|
||
- **其他**: openpyxl, pyyaml, click
|
||
|
||
---
|
||
|
||
## 验证和测试
|
||
|
||
### 验证安装
|
||
|
||
```bash
|
||
# 检查 Python 依赖
|
||
python verify_setup.py
|
||
|
||
# 检查模型文件
|
||
python check_mole_dependencies.py
|
||
```
|
||
|
||
### 运行测试
|
||
|
||
```bash
|
||
# 功能测试(使用小规模测试数据)
|
||
python test_mole_predictor.py
|
||
```
|
||
|
||
---
|
||
|
||
## 常见问题
|
||
|
||
### Q1: 如何处理大文件?
|
||
|
||
**方案 1:** 增加批次大小和工作进程数
|
||
|
||
```bash
|
||
python utils/mole_predictor.py large_file.csv \
|
||
--batch-size 500 \
|
||
--n-workers 8
|
||
```
|
||
|
||
**方案 2:** 先提取部分数据测试
|
||
|
||
```bash
|
||
# 提取前 1000 行
|
||
head -1001 large_file.csv > test_1000.csv
|
||
python utils/mole_predictor.py test_1000.csv
|
||
```
|
||
|
||
### Q2: 如何只使用 CPU?
|
||
|
||
```bash
|
||
python utils/mole_predictor.py input.csv --device cpu
|
||
```
|
||
|
||
### Q3: 列名大小写问题?
|
||
|
||
工具会自动进行大小写不敏感的列名匹配,所以 `SMILES`、`smiles`、`Smiles` 都可以识别。
|
||
|
||
### Q4: ModuleNotFoundError 错误?
|
||
|
||
确保已安装依赖:
|
||
|
||
```bash
|
||
uv pip install -r requirements-mole.txt
|
||
```
|
||
|
||
对于 RDKit,推荐使用 conda:
|
||
|
||
```bash
|
||
conda install -c conda-forge rdkit
|
||
```
|
||
|
||
### Q5: 如何自定义模型路径?
|
||
|
||
```python
|
||
from models import PredictionConfig, ParallelBroadSpectrumPredictor
|
||
|
||
config = PredictionConfig(
|
||
xgboost_model_path="/path/to/model.pkl",
|
||
mole_model_path="/path/to/mole_model",
|
||
strain_categories_path="/path/to/strain_data.tsv.gz",
|
||
gram_info_path="/path/to/gram_info.xlsx",
|
||
app_threshold=0.044,
|
||
min_nkill=10,
|
||
batch_size=100,
|
||
device="auto"
|
||
)
|
||
|
||
predictor = ParallelBroadSpectrumPredictor(config)
|
||
```
|
||
|
||
### Q6: GPU 内存不足?
|
||
|
||
减小批次大小:
|
||
|
||
```bash
|
||
python utils/mole_predictor.py input.csv --batch-size 50
|
||
```
|
||
|
||
### Q7: 模型文件在哪里?
|
||
|
||
模型文件位于相邻的 `mole_broad_spectrum_parallel` 项目中:
|
||
|
||
```
|
||
../mole_broad_spectrum_parallel/
|
||
├── pretrained_model/model_ginconcat_btwin_100k_d8000_l0.0001/
|
||
│ ├── config.yaml
|
||
│ └── model.pth
|
||
├── data/03.model_evaluation/MolE-XGBoost-08.03.2024_14.20.pkl
|
||
└── ...
|
||
```
|
||
|
||
运行 `python check_mole_dependencies.py` 检查文件是否存在。
|
||
|
||
---
|
||
|
||
## 性能建议
|
||
|
||
- **使用 GPU**: 设置 `--device cuda:0` 可大幅加速(需要 CUDA)
|
||
- **调整批次**: 较大的批次(100-500)通常更快
|
||
- **多进程**: 使用 `--n-workers` 指定工作进程数
|
||
- **首次加载**: 首次运行需要加载模型(~30秒),后续会更快
|
||
|
||
### 性能参考
|
||
|
||
| 分子数量 | CPU (8核) | GPU (CUDA) |
|
||
|---------|----------|------------|
|
||
| 100 | ~30秒 | ~10秒 |
|
||
| 1,000 | ~5分钟 | ~1分钟 |
|
||
| 10,000 | ~50分钟 | ~8分钟 |
|
||
|
||
---
|
||
|
||
## 系统要求
|
||
|
||
- **Python**: 3.7 或更高版本(推荐 3.12)
|
||
- **内存**: 最低 8 GB RAM
|
||
- **存储**: 至少 2 GB 可用空间
|
||
- **GPU**: 可选,但强烈推荐(需要 CUDA 支持)
|
||
|
||
---
|
||
|
||
## 技术支持
|
||
|
||
如有问题:
|
||
|
||
1. 查看验证结果: `python verify_setup.py`
|
||
2. 检查模型文件: `python check_mole_dependencies.py`
|
||
3. 运行功能测试: `python test_mole_predictor.py`
|
||
|
||
---
|
||
|
||
## 许可
|
||
|
||
详见 LICENSE 文件。
|
||
|
||
## 引用
|
||
|
||
如果使用本工具,请引用相关论文。
|
||
|
||
---
|
||
|
||
**更新日期**: 2025-10-16
|
||
**版本**: 1.0.0
|