SURE-tools 1.0.7__py3-none-any.whl → 1.0.9__py3-none-any.whl

SURE/SURE.py

@@ -22,6 +22,9 @@ import datatable as dt
 from tqdm import tqdm
 from scipy import sparse
 
+import scanpy as sc
+from .atac import binarize
+
 from typing import Literal
 
 import warnings

SURE/__init__.py

@@ -3,5 +3,6 @@ from .SURE import SURE
 from . import utils 
 from . import codebook
 from . import SURE
+from . import atac
 
-__all__ = ['SURE', 'utils', 'codebook']
+__all__ = ['SURE', 'atac', 'utils', 'codebook']

SURE/atac/__init__.py

@@ -0,0 +1 @@
+from .utils import tfidf, binarize

SURE/atac/utils.py

@@ -0,0 +1,151 @@
+from typing import Union, Optional
+from warnings import warn
+
+import numpy as np
+from scipy.sparse import csr_matrix, dia_matrix, issparse
+from sklearn.preprocessing import MinMaxScaler
+
+from anndata import AnnData
+
+import scanpy as sc
+from scanpy._utils import view_to_actual
+
+
+# Computational methods for preprocessing
+
+
+def tfidf(
+    data: AnnData,
+    log_tf: bool = True,
+    log_idf: bool = True,
+    log_tfidf: bool = False,
+    scale_factor: Union[int, float] = 1e4,
+    inplace: bool = True,
+    copy: bool = False,
+    from_layer: Optional[str] = None,
+    to_layer: Optional[str] = None,
+):
+    """
+    Transform peak counts with TF-IDF (Term Frequency - Inverse Document Frequency).
+
+    TF: peak counts are normalised by total number of counts per cell
+    DF: total number of counts for each peak
+    IDF: number of cells divided by DF
+
+    By default, log(TF) * log(IDF) is returned.
+
+    Parameters
+    ----------
+    data
+            AnnData object with peak counts.
+    log_idf
+            Log-transform IDF term (True by default).
+    log_tf
+            Log-transform TF term (True by default).
+    log_tfidf
+            Log-transform TF*IDF term (False by default).
+            Can only be used when log_tf and log_idf are False.
+    scale_factor
+            Scale factor to multiply the TF-IDF matrix by (1e4 by default).
+    inplace
+            If to modify counts in the AnnData object (True by default).
+    copy
+            If to return a copy of the AnnData object or the 'atac' modality (False by default).
+            Not compatible with inplace=False.
+    from_layer
+            Layer to use counts (AnnData.layers[from_layer])
+            instead of AnnData.X used by default.
+    to_layer
+            Layer to save transformed counts to (AnnData.layers[to_layer])
+            instead of AnnData.X used by default.
+            Not compatible with inplace=False.
+    """
+    if isinstance(data, AnnData):
+        adata = data
+    else:
+        raise TypeError("Expected AnnData object")
+
+    if log_tfidf and (log_tf or log_idf):
+        raise AttributeError(
+            "When returning log(TF*IDF), \
+            applying neither log(TF) nor log(IDF) is possible."
+        )
+
+    if copy and not inplace:
+        raise ValueError("`copy=True` cannot be used with `inplace=False`.")
+
+    if to_layer is not None and not inplace:
+        raise ValueError(f"`to_layer='{str(to_layer)}'` cannot be used with `inplace=False`.")
+
+    if copy:
+        adata = adata.copy()
+
+    view_to_actual(adata)
+
+    counts = adata.X if from_layer is None else adata.layers[from_layer]
+
+    # Check before the computation
+    if to_layer is not None and to_layer in adata.layers:
+        warn(f"Existing layer '{str(to_layer)}' will be overwritten")
+
+    if issparse(counts):
+        n_peaks = np.asarray(counts.sum(axis=1)).reshape(-1)
+        n_peaks = dia_matrix((1.0 / n_peaks, 0), shape=(n_peaks.size, n_peaks.size))
+        # This prevents making TF dense
+        tf = np.dot(n_peaks, counts)
+    else:
+        n_peaks = np.asarray(counts.sum(axis=1)).reshape(-1, 1)
+        tf = counts / n_peaks
+
+    if scale_factor is not None and scale_factor != 0 and scale_factor != 1:
+        tf = tf * scale_factor
+    if log_tf:
+        tf = np.log1p(tf)
+
+    idf = np.asarray(adata.shape[0] / counts.sum(axis=0)).reshape(-1)
+    if log_idf:
+        idf = np.log1p(idf)
+
+    if issparse(tf):
+        idf = dia_matrix((idf, 0), shape=(idf.size, idf.size))
+        tf_idf = np.dot(tf, idf)
+    else:
+        tf_idf = np.dot(csr_matrix(tf), csr_matrix(np.diag(idf)))
+
+    if log_tfidf:
+        tf_idf = np.log1p(tf_idf)
+
+    res = np.nan_to_num(tf_idf, nan=0.0)
+    if not inplace:
+        return res
+
+    if to_layer is not None:
+        adata.layers[to_layer] = res
+    else:
+        adata.X = res
+
+    if copy:
+        return adata
+
+
+def binarize(data: AnnData, threshold: np.float32=0):
+    """
+    Transform peak counts to the binary matrix (all the non-zero values become 1).
+
+    Parameters
+    ----------
+    data
+            AnnData object with peak counts.
+    """
+    if isinstance(data, AnnData):
+        adata = data
+    else:
+        raise TypeError("Expected AnnData object")
+
+    if issparse(adata.X):
+        # Sparse matrix
+        adata.X.data[adata.X.data > threshold] = 1
+    else:
+        adata.X[adata.X > threshold] = 1
+
+

{sure_tools-1.0.7.dist-info → sure_tools-1.0.9.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: SURE-tools
-Version: 1.0.7
+Version: 1.0.9
 Summary: Succinct Representation of Single Cells
 Home-page: https://github.com/ZengFLab/SURE
 Author: Feng Zeng

{sure_tools-1.0.7.dist-info → sure_tools-1.0.9.dist-info}/RECORD

@@ -1,17 +1,19 @@
-SURE/SURE.py,sha256=3k28h-IMb2poRjoM5yq_u7FWO1Z2ixr_k6o05fF9eEE,48333
-SURE/__init__.py,sha256=SbIRwAVBnNhza9vbsUH4N04atr0q_Abp04pCUTBhNio,127
+SURE/SURE.py,sha256=mEvzoIvjZUSO0UYSrb-CyGJtkFEweT92_XMQor_lsk0,48381
+SURE/__init__.py,sha256=UyoyPqpGllsoVLMAO8gX0U6RJUMla6rX86UoAueiKEc,154
 SURE/assembly/__init__.py,sha256=jxZLURXKPzXe21LhrZ09LgZr33iqdjlQy4oSEj5gR2Q,172
 SURE/assembly/assembly.py,sha256=6IMdelPOiRO4mUb4dC7gVCoF1Uvfw86-Map8P_jnUag,21477
 SURE/assembly/atlas.py,sha256=ALjmVWutm_tOHTcT1aqOxmuCEQw-XzrtDoMCV_8oXLk,21794
+SURE/atac/__init__.py,sha256=3smP8IKHfwNCd1G_sZH3pKHXuLkLpFuLtjUTUSy7_As,34
+SURE/atac/utils.py,sha256=m4NYwpy9O5T1pXTzgCOCcmlwrC6GTi-cQ5sm2wZu2O8,4354
 SURE/codebook/__init__.py,sha256=2T5gjp8JIaBayrXAnOJYSebQHsWprOs87difpR1OPNw,243
 SURE/codebook/codebook.py,sha256=ZlN6gRX9Gj2D2u3P5KeOsbZri0MoMAiJo9lNeL-MK-I,17117
 SURE/utils/__init__.py,sha256=Htqv4KqVKcRiaaTBsR-6yZ4LSlbhbzutjNKXGD9-uds,660
 SURE/utils/custom_mlp.py,sha256=07TYX1HgxfEjb_3i5MpiZfNhOhx3dKntuwGkrpteWiM,7036
 SURE/utils/queue.py,sha256=E_5PA5EWcBoGAZj8BkKQnkCK0p4C-4-xcTPqdIXaPXU,1892
 SURE/utils/utils.py,sha256=IUHjDDtYaAYllCWsZyIzqQwaLul6fJRvHRH4vIYcR-c,8462
-sure_tools-1.0.7.dist-info/licenses/LICENSE,sha256=TFHKwmrAViXQbSX5W-NDItkWFjm45HWOeUniDrqmnu0,1065
-sure_tools-1.0.7.dist-info/METADATA,sha256=jS2KOJhrVk3boa1Yws4cQyIQykMCk9LaCkozyIn5Uj8,2650
-sure_tools-1.0.7.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-sure_tools-1.0.7.dist-info/entry_points.txt,sha256=u12payZYgCBy5FCwRHP6AlSQhKCiWSEDwj68r1DVdn8,40
-sure_tools-1.0.7.dist-info/top_level.txt,sha256=BtFTebdiJeqra4r6mm-uEtwVRFLZ_IjYsQ7OnalrOvY,5
-sure_tools-1.0.7.dist-info/RECORD,,
+sure_tools-1.0.9.dist-info/licenses/LICENSE,sha256=TFHKwmrAViXQbSX5W-NDItkWFjm45HWOeUniDrqmnu0,1065
+sure_tools-1.0.9.dist-info/METADATA,sha256=BrvN2nIVYq-z39MdOb3nGNnpYbPifOHOmDBJF2zphVo,2650
+sure_tools-1.0.9.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+sure_tools-1.0.9.dist-info/entry_points.txt,sha256=u12payZYgCBy5FCwRHP6AlSQhKCiWSEDwj68r1DVdn8,40
+sure_tools-1.0.9.dist-info/top_level.txt,sha256=BtFTebdiJeqra4r6mm-uEtwVRFLZ_IjYsQ7OnalrOvY,5
+sure_tools-1.0.9.dist-info/RECORD,,