SURE-tools 2.1.6__tar.gz → 2.1.7__tar.gz

{sure_tools-2.1.6 → sure_tools-2.1.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: SURE-tools
-Version: 2.1.6
+Version: 2.1.7
 Summary: Succinct Representation of Single Cells
 Home-page: https://github.com/ZengFLab/SURE
 Author: Feng Zeng

{sure_tools-2.1.6 → sure_tools-2.1.7}/SURE/__init__.py

@@ -7,5 +7,6 @@ from . import SURE
 from . import PerturbFlow
 from . import atac
 from . import flow 
+from . import perturb
 
-__all__ = ['SURE', 'PerturbFlow', 'flow', 'atac', 'utils', 'codebook']
+__all__ = ['SURE', 'PerturbFlow', 'flow', 'perturb', 'atac', 'utils', 'codebook']

sure_tools-2.1.7/SURE/perturb/__init__.py

@@ -0,0 +1 @@
+from .perturb import label_to_matrix, matrix_to_labels, vectorized_label_to_matrix, parallel_label_to_matrix

sure_tools-2.1.7/SURE/perturb/perturb.py

@@ -0,0 +1,68 @@
+import re
+import numpy as np
+from itertools import chain
+from joblib import Parallel, delayed
+
+def label_to_matrix(labels, sep_pattern=r'[;_\-\s]'):
+    """
+    将混合分隔符的多标签数据转换为 0-1 矩阵
+    
+    Args:
+        labels: 原始标签列表，如 ["cat", "dog", "cat;dog", "bird_dog"]
+        sep_pattern: 多标签分隔符的正则模式（默认匹配 ; _ - 和空格）
+    
+    Returns:
+        one_hot_matrix: 0-1 矩阵
+        unique_labels: 唯一标签列表
+    """
+    # 统一分隔符
+    labels_unified = [re.sub(sep_pattern, ';', label) for label in labels]
+    
+    # 获取所有唯一标签
+    all_unique_labels = sorted(set(chain(*[label.split(';') for label in labels_unified])))
+    
+    # 生成 0-1 矩阵
+    matrix = np.zeros((len(labels), len(all_unique_labels)), dtype=int)
+    label_to_idx = {label: i for i, label in enumerate(all_unique_labels)}
+    
+    for i, label in enumerate(labels_unified):
+        for sub_label in label.split(';'):
+            if sub_label in label_to_idx:
+                matrix[i, label_to_idx[sub_label]] = 1
+    
+    return matrix, all_unique_labels
+
+
+def vectorized_label_to_matrix(labels, sep_pattern=r'[;_\-\s]'):
+    labels_unified = [re.sub(sep_pattern, ';', label) for label in labels]
+    unique_labels = sorted(set(chain(*[label.split(';') for label in labels_unified])))
+    
+    # 向量化操作
+    label_matrix = np.array([label.split(';') for label in labels_unified], dtype=object)
+    matrix = np.zeros((len(labels), len(unique_labels)), dtype=int)
+    
+    for i, label in enumerate(unique_labels):
+        matrix[:, i] = np.array([label in lst for lst in label_matrix], dtype=int)
+    
+    return matrix, unique_labels
+
+
+
+def parallel_label_to_matrix(labels, n_jobs=4):
+    labels_unified = [re.sub(r'[;_\-\s]', ';', label) for label in labels]
+    unique_labels = sorted(set(chain(*[label.split(';') for label in labels_unified])))
+    
+    def process_row(row_labels, unique_labels):
+        return [1 if label in row_labels else 0 for label in unique_labels]
+    
+    label_lists = [label.split(';') for label in labels_unified]
+    matrix = Parallel(n_jobs=n_jobs)(
+        delayed(process_row)(row, unique_labels) for row in label_lists
+    )
+    
+    return np.array(matrix, dtype=int), unique_labels
+
+
+def matrix_to_labels(matrix, unique_labels):
+    return [';'.join([unique_labels[i] for i in np.where(row)[0]]) 
+            for row in matrix]

{sure_tools-2.1.6 → sure_tools-2.1.7}/SURE_tools.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: SURE-tools
-Version: 2.1.6
+Version: 2.1.7
 Summary: Succinct Representation of Single Cells
 Home-page: https://github.com/ZengFLab/SURE
 Author: Feng Zeng

{sure_tools-2.1.6 → sure_tools-2.1.7}/SURE_tools.egg-info/SOURCES.txt

@@ -14,6 +14,8 @@ SURE/codebook/codebook.py
 SURE/flow/__init__.py
 SURE/flow/flow_stats.py
 SURE/flow/plot_quiver.py
+SURE/perturb/__init__.py
+SURE/perturb/perturb.py
 SURE/utils/__init__.py
 SURE/utils/custom_mlp.py
 SURE/utils/queue.py

{sure_tools-2.1.6 → sure_tools-2.1.7}/setup.py

@@ -5,7 +5,7 @@ with open("README.md", "r") as fh:
 
 setup(
     name='SURE-tools',
-    version='2.1.6',
+    version='2.1.7',
     description='Succinct Representation of Single Cells',
     long_description=long_description,
     long_description_content_type="text/markdown",