该包的并行在windows上不适用

[qibi777]

您好！
首先非常感谢您的团队开发出这个功能强大的注释工具。我最近在Windows系统上使用这个工具时遇到报错：“AttributeError: Can't pickle local object 'STCAT..process_CD4'”，我认为这条报错说明该包的并行功能不适用于Windows系统。
恳请您和您的团队为Windows系统发布该包的正式版本，另外我也通过对 STCAT\lib\site-packages\STCAT\STCAT.py 进行修改从而顺利在Windows系统上运行了该包。
具体操作包括：
1.找到STCAT函数内部定义的process_CD4和process_CD8这两个嵌套函数。
2.将这两个函数从STCAT函数内部移到外面，让它们成为文件顶层的全局函数。

初次之外还有另外一处错误出现：
AttributeError: 'Categorical' object has no attribute 'empty'
同样进行了修改：
原代码：
if 'over_clustering_CD4' in adata_T_sub_CD4.obs and not adata_T_sub_CD4.obs['over_clustering_CD4'].unique().empty:

修改后：
if 'over_clustering_CD4' in adata_T_sub_CD4.obs and adata_T_sub_CD4.obs['over_clustering_CD4'].unique().size > 0:

原代码：
if 'over_clustering_CD8' in adata_T_sub_CD8.obs and not adata_T_sub_CD8.obs['over_clustering_CD8'].unique().empty:

修改后：
if 'over_clustering_CD8' in adata_T_sub_CD8.obs and adata_T_sub_CD8.obs['over_clustering_CD8'].unique().size > 0:

[1147230641]

您好，我遇到了同样的问题，但是由于我的编程基础比较薄弱，目前仅达到运用的阶段，根据您提示的”具体操作方案“，我仍不得要领。
能否请您分享修改后的函数包，非常感谢！
如您愿意分享，我的邮箱为754638545@qq.com，如有叨扰，不慎惭愧。

[Panipenem]

Ran into the same problem on macOS. I changed multiprocessing to sequential execution to resolve this problem.

in STCAT.py file:

1. change worker function to return values

before:

def process_CD4(q1):
    ...
    q1.put(adata_T_sub_CD4)

# and

def process_CD8(q2):
    ...
    q2.put(adata_T_sub_CD8)

after:

def process_CD4():
    ...
    return adata_T_sub_CD4

def process_CD8():
    ...
    return adata_T_sub_CD8

2. Change the multiprocessing block

remove:

def start_processes():
    q1 = Queue()
    q2 = Queue()
    p1 = Process(target=process_CD4, args=(q1,))
    p2 = Process(target=process_CD8, args=(q2,))
    p1.start()
    p2.start()
    adata_T_sub_CD4 = q1.get()
    adata_T_sub_CD8 = q2.get()
    p1.join()
    p2.join()
    return adata_T_sub_CD4, adata_T_sub_CD8

adata_T_sub_CD4, adata_T_sub_CD8 = start_processes()

add:

adata_T_sub_CD4 = process_CD4()
adata_T_sub_CD8 = process_CD8()

3. restart kernel and it should work.

[1147230641]

您好，我已收到您的邮件，会尽快查看，谢谢！

[1147230641]

Thanks for you reply. I will try this solution as soon.  蒋朦朦 ***@***.***  

…

------------------ 原始邮件 ------------------ 发件人: ***@***.***>; 发送时间: 2025年11月17日(星期一) 下午5:10 收件人: ***@***.***>; 抄送: ***@***.***>; ***@***.***>; 主题: Re: [GuoBioinfoLab/STCAT] 该包的并行在windows上不适用 (Issue #7) Panipenem left a comment (GuoBioinfoLab/STCAT#7) Ran into the same problem on macOS. I changed multiprocessing to sequential execution to resolve this problem. in STCAT.py file: change worker function to return values before: def process_CD4(q1): ... q1.put(adata_T_sub_CD4) # and def process_CD8(q2): ... q2.put(adata_T_sub_CD8) after: def process_CD4(): ... return adata_T_sub_CD4 def process_CD8(): ... return adata_T_sub_CD8 Change the multiprocessing block remove: def start_processes(): q1 = Queue() q2 = Queue() p1 = Process(target=process_CD4, args=(q1,)) p2 = Process(target=process_CD8, args=(q2,)) p1.start() p2.start() adata_T_sub_CD4 = q1.get() adata_T_sub_CD8 = q2.get() p1.join() p2.join() return adata_T_sub_CD4, adata_T_sub_CD8 adata_T_sub_CD4, adata_T_sub_CD8 = start_processes() add: adata_T_sub_CD4 = process_CD4() adata_T_sub_CD8 = process_CD8() restart kernel and it should work. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***>