sciv 0.0.102__tar.gz → 0.0.104__tar.gz

{sciv-0.0.102 → sciv-0.0.104}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sciv
-Version: 0.0.102
+Version: 0.0.104
 Summary: Unveiling the pivotal cell types involved in variant function regulation at a single-cell resolution
 Project-URL: github, https://github.com/YuZhengM/sciv
 Author-email: Zheng-Min Yu <yuzmbio@163.com>

{sciv-0.0.102 → sciv-0.0.104}/pyproject.toml

@@ -3,7 +3,7 @@ requires = ["hatchling"]
 build-backend = "hatchling.build"
 [project]
 name = "sciv"
-version = "0.0.102"
+version = "0.0.104"
 authors = [
     { name = "Zheng-Min Yu", email = "yuzmbio@163.com" },
 ]

{sciv-0.0.102 → sciv-0.0.104}/src/sciv/model/_core_.py

@@ -76,8 +76,8 @@ def core(
     early_stopping_patience: int = 50,
     batch_key: Optional[str] = None,
     resolution: float = 0.5,
-    k: int = 30,
-    or_k: int = 1,
+    k: int = 50,
+    or_k: int = 25,
     weight: float = 0.1,
     kernel: Literal["laplacian", "gaussian"] = "gaussian",
     local_k: int = 10,

{sciv-0.0.102 → sciv-0.0.104}/src/sciv/plot/_scatter_.py

@@ -174,6 +174,8 @@ def scatter_3d(
     edge_color: str = None,
     size: Union[float, collection] = 0.1,
     legend_name: str = None,
+    is_add_max_label: bool = False,
+    text_left_offset: float = 0.5,
     output: path = None,
     show: bool = True,
     close: bool = False,
@@ -186,13 +188,14 @@ def scatter_3d(
     fig = plt.figure(figsize=(width, height))
     ax = fig.add_subplot(projection='3d')
 
-    hue_cat = pd.Categorical(df[hue])
+    if hue is not None:
+        hue_cat = pd.Categorical(df[hue])
 
     scatter = ax.scatter(
         df[x],
         df[y],
         df[z],
-        c=hue_cat.codes,
+        c=hue_cat.codes if hue is not None else None,
         cmap=cmap,
         s=size,
         edgecolors=edge_color,
@@ -214,7 +217,7 @@ def scatter_3d(
     if title is not None:
         ax.set_title(title, fontsize=font_size)
 
-    if is_add_legend:
+    if is_add_legend and hue is not None:
         unique_types = hue_cat.categories
         legend_elements = [
             plt.Line2D(
@@ -226,6 +229,24 @@ def scatter_3d(
 
         ax.legend(handles=legend_elements, title=legend_name, loc='upper left')
 
+    if is_add_max_label:
+
+        max_idx = df[z].idxmax()
+        max_x = df.loc[max_idx, x]
+        max_y = df.loc[max_idx, y]
+        max_value = df.loc[max_idx, z]
+
+        # 在最大值点的位置添加文本标签
+        ax.text(
+            max_x - text_left_offset,
+            max_y,
+            max_value,
+            f'({max_x}, {max_y}): {max_value:.3f}',
+            fontsize=font_size - 2,
+            color='red',
+            ha='left'
+        )
+
     plot_end(fig, None, None, None, output, show, close)
 
 

{sciv-0.0.102 → sciv-0.0.104}/src/sciv/preprocessing/_scvi_.py

@@ -137,7 +137,7 @@ def poisson_vi(
                 ul.log(__name__).warning(f"GPU failed to run, try to switch to CPU running.\n {ome}")
 
                 try:
-                    model = scvi.external.POISSONVI.load(model_dir, adata=adata, accelerator="cpu", device="cpu")
+                    model = scvi.external.POISSONVI.load(model_dir, adata=adata, accelerator="cpu")
                 except Exception as e:
                     ul.log(__name__).error(f"File `model.pt` failed to load, you can execute `Poisson VI` again by deleting file `model.pt` ({model_dir}/model.pt).\n {e}")
                     raise ValueError(f"File `model.pt` failed to load, you can execute `Poisson VI` again by deleting file `model.pt` ({model_dir}/model.pt).")

{sciv-0.0.102 → sciv-0.0.104}/src/sciv/tool/__init__.py

@@ -36,7 +36,7 @@ from ._algorithm_ import (
     calculate_init_score_weight,
     obtain_cell_cell_network,
     perturb_data,
-    add_noise
+    add_bernoulli_fluctuation_noise
 )
 
 from ._random_walk_ import (
@@ -98,7 +98,7 @@ __all__ = [
     "calculate_init_score_weight",
     "obtain_cell_cell_network",
     "perturb_data",
-    "add_noise",
+    "add_bernoulli_fluctuation_noise",
     "split_matrix",
     "merge_matrix",
     "down_sampling_data",

{sciv-0.0.102 → sciv-0.0.104}/src/sciv/tool/_algorithm_.py

@@ -8,7 +8,6 @@ from scipy import sparse
 from scipy.stats import norm
 from tqdm import tqdm
 from joblib import Parallel, delayed
-import multiprocessing
 
 import numpy as np
 from anndata import AnnData
@@ -1005,15 +1004,15 @@ def overlap_sum(regions: AnnData, variants: dict, trait_info: DataFrame, n_jobs:
     total = sum(len(ld) for ld, _, _ in results)
     row_indices = np.empty(total, dtype=np.int32)
     col_indices = np.empty(total, dtype=np.int32)
-    data_vals  = np.empty(total, dtype=np.float32)
+    data_vals = np.empty(total, dtype=np.float32)
 
     ptr = 0
 
     for local_data, local_rows, local_cols in results:
         n = len(local_data)
-        row_indices[ptr:ptr+n] = local_rows
-        col_indices[ptr:ptr+n] = local_cols
-        data_vals[ptr:ptr+n] = local_data
+        row_indices[ptr:ptr + n] = local_rows
+        col_indices[ptr:ptr + n] = local_cols
+        data_vals[ptr:ptr + n] = local_data
         ptr += n
 
     # Build sparse matrix, then convert to csr format
@@ -1404,39 +1403,36 @@ def perturb_data(data: collection, percentage: float) -> collection:
     return new_data
 
 
-def add_noise(data: matrix_data, rate: float) -> matrix_data:
-    """
-    Add peak percentage noise to each cell
+def add_bernoulli_fluctuation_noise(
+    counts_matrix: matrix_data,
+    noise_level: float = 0.1
+) -> matrix_data:
     """
+    Add Bernoulli fluctuation noise to the counts matrix (add 1 with probability noise_level)
 
-    if rate <= 0 or rate >= 1:
-        raise ValueError("The value of the `rate` parameter must be greater than 0 and less than 1.")
+    Parameters
+    ----------
+    counts_matrix : matrix_data
+        Input counts matrix
+    noise_level : float, default 0.1
+        Noise level, i.e., the probability of randomly adding 1 (range: 0.0 - 1.0)
 
-    shape = data.shape
-    noise = to_dense(data.copy())
-
-    for i in tqdm(range(shape[0])):
-        count_i = np.array(noise[i, :]).flatten()
-        # Add noise to the accessibility of unopened chromatin
-        count0_i = count_i[count_i == 0]
-        max_i = np.max(count_i)
-        count0 = int(count0_i.size * rate)
-        noise0_i = np.random.randint(low=1, high=2 if max_i < 2 else max_i, size=count0)
-        random_index0 = np.random.choice(np.arange(0, count0_i.size), size=count0, replace=False)
-        count0_i[random_index0] = noise0_i
-        count_i_value = count_i.copy()
-        count_i_value[count_i_value == 0] = count0_i
-        noise[i, :] = count_i_value
+    Returns
+    -------
+    matrix_data
+        Matrix after adding noise
+    """
+    if noise_level < 0 or noise_level > 1:
+        ul.log(__name__).error("The value of the `noise_level` parameter must be greater than 0 and less than 1.")
+        raise ValueError("The value of the `noise_level` parameter must be greater than 0 and less than 1.")
 
-        # Close open chromatin accessibility
-        count1_i = count_i[count_i == 1]
-        count1 = int(count1_i.size * rate)
-        random_index1 = np.random.choice(np.arange(0, count1_i.size), size=count1, replace=False)
-        count1_i[random_index1] = 0
-        count_i[count_i == 1] = count1_i
-        noise[i, :] = count_i
+    if noise_level == 0:
+        return counts_matrix.copy()
 
-        # disturbance
-        noise[i, :] = perturb_data(noise[i, :], rate)
+    noise = np.random.binomial(
+        n=1,
+        p=noise_level,
+        size=counts_matrix.shape
+    ).astype(counts_matrix.dtype)
 
-    return noise
+    return counts_matrix + noise

{sciv-0.0.102 → sciv-0.0.104}/src/sciv/util/_core_.py

@@ -39,14 +39,12 @@ def log(name: str = None) -> Logger:
     return Logger(name, log_path=os.path.join(ul.log_file_path, name), is_form_file=ul.is_form_log_file)
 
 
-def track_with_memory(is_monitor: bool = False , interval: float = 60) -> Callable:
+def track_with_memory(interval: float = 60) -> Callable:
     """
     Decorator: Records memory usage at fixed intervals during function execution and returns the result, elapsed time, and memory list.
 
     Parameters
     ----------
-    is_monitor : bool, optional
-        Whether to enable memory monitoring, default is False.
     interval : float, optional
         Sampling interval (seconds), default is 60 seconds.
 
@@ -63,9 +61,6 @@ def track_with_memory(is_monitor: bool = False , interval: float = 60) -> Callab
         @wraps(func)
         def wrapper(*args, **kwargs) -> Union[Any, dict]:
 
-            if not is_monitor:
-                return func(*args, **kwargs)
-
             process = psutil.Process(os.getpid())
 
             stop_monitor = False