Can't open bgen using symbolic links #42
Assignees
Comments
|
Jordan,
Thanks for using Bgen-Reader!
I'll be able to look into this tomorrow. I wonder if you have write access to the *.bgen file's directory?
In any case, adding an option to put the *.mmm file in any directory seems like the right answer and that is likely what we'll do.
* Carl
From: Jordan Rossen ***@***.***>
Sent: Tuesday, June 07, 2022 7:58 AM
To: limix/bgen-reader-py ***@***.***>
Cc: Subscribed ***@***.***>
Subject: [limix/bgen-reader-py] Can't open bgen using symbolic links (Issue #42)
Hello,
I want to use the numpy inspired API to open an bgen using a symbolic link. I want to do this because I don't want to create the .metadata2.mmm file in the shared directory where the bgen lives. I thought I recalled being able to do this in a previous version of bgen-reader, but am running into the following problem now.
from bgen_reader import open_bgen
wha = open_bgen('/n/holylfs05/LABS/liang_lab_l3/Lab/jordan/ssctpr/data/ukbb_symbolic_link/ukb_imp_chr1_v3.bgen')
Traceback (most recent call last):
File "", line 1, in
File "/n/home_fasse/jrossen/.conda/envs/ssctpr2/lib/python3.9/site-packages/bgen_reader/_bgen2.py", line 133, in init
assert_file_exist(filepath)
File "/n/home_fasse/jrossen/.conda/envs/ssctpr2/lib/python3.9/site-packages/bgen_reader/_file.py", line 14, in assert_file_exist
raise FileNotFoundError(errno.ENOENT, os.strerror(errno.ENOENT), filepath)
I don't have a very strong understanding of how PosixPath works, but after a little playing around, it seems like PosixPath precludes the use of symbolic links.
Either allowing for the use of symbolic links, or allowing the user to specify a directory where .metadata2.mmm should be created would solve my problem completely.
Thanks for making this software!
-
Reply to this email directly, view it on GitHub<https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Flimix%2Fbgen-reader-py%2Fissues%2F42&data=05%7C01%7C%7Cf122f3122a244133143308da48962d84%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C637902107101785825%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Csmdm8wELv3aKeKlHLljUI4eIs5fH8UKEJFh2SCk9Kc%3D&reserved=0>, or unsubscribe<https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FABR65P4NZAC4U3P7ZSZZC3TVN5PRFANCNFSM5YDIGANA&data=05%7C01%7C%7Cf122f3122a244133143308da48962d84%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C637902107101785825%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=ZEqkxE39TrV6eWCALOmeeG4cTCjb5EY4ZbchudfhbGg%3D&reserved=0>.
You are receiving this because you are subscribed to this thread.Message ID: ***@***.******@***.***>>
|
|
Jordan,
Here is a possible temporary fix.
Uninstall bgen-reader (for example, pip uninstall bgen-reader)
Then install my private version from GitHub:
pip install ***@***.***
Then you should be able to use a new 'metadata_filepath' in open_bgen. For example:
filepath = example_filepath("example.32bits.bgen")
with open_bgen(
filepath, metadata_filepath=r"m:\deldir\cmk.mmm", verbose=False
) as bgen2:
for num_threads in [1, 2]:
for slice in [np.s_[:, :], np.s_[:, []]]:
val = bgen2.read(index=slice, num_threads=num_threads)
row_count = len(bgen2.samples[slice[0]])
col_count = len(bgen2.ids[slice[1]])
assert val.shape == (row_count, col_count, 3)
Please, let me know if this works for you. In any case, I'll have time tomorrow to work on it.
* Carl
From: Carl K
Sent: Tuesday, June 07, 2022 8:14 AM
To: limix/bgen-reader-py ***@***.***>; limix/bgen-reader-py ***@***.***>
Cc: Subscribed ***@***.***>
Subject: RE: [limix/bgen-reader-py] Can't open bgen using symbolic links (Issue #42)
Jordan,
Thanks for using Bgen-Reader!
I'll be able to look into this tomorrow. I wonder if you have write access to the *.bgen file's directory?
In any case, adding an option to put the *.mmm file in any directory seems like the right answer and that is likely what we'll do.
* Carl
From: Jordan Rossen ***@***.******@***.***>>
Sent: Tuesday, June 07, 2022 7:58 AM
To: limix/bgen-reader-py ***@***.******@***.***>>
Cc: Subscribed ***@***.******@***.***>>
Subject: [limix/bgen-reader-py] Can't open bgen using symbolic links (Issue #42)
Hello,
I want to use the numpy inspired API to open an bgen using a symbolic link. I want to do this because I don't want to create the .metadata2.mmm file in the shared directory where the bgen lives. I thought I recalled being able to do this in a previous version of bgen-reader, but am running into the following problem now.
from bgen_reader import open_bgen
wha = open_bgen('/n/holylfs05/LABS/liang_lab_l3/Lab/jordan/ssctpr/data/ukbb_symbolic_link/ukb_imp_chr1_v3.bgen')
Traceback (most recent call last):
File "", line 1, in
File "/n/home_fasse/jrossen/.conda/envs/ssctpr2/lib/python3.9/site-packages/bgen_reader/_bgen2.py", line 133, in init
assert_file_exist(filepath)
File "/n/home_fasse/jrossen/.conda/envs/ssctpr2/lib/python3.9/site-packages/bgen_reader/_file.py", line 14, in assert_file_exist
raise FileNotFoundError(errno.ENOENT, os.strerror(errno.ENOENT), filepath)
I don't have a very strong understanding of how PosixPath works, but after a little playing around, it seems like PosixPath precludes the use of symbolic links.
Either allowing for the use of symbolic links, or allowing the user to specify a directory where .metadata2.mmm should be created would solve my problem completely.
Thanks for making this software!
-
Reply to this email directly, view it on GitHub<https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Flimix%2Fbgen-reader-py%2Fissues%2F42&data=05%7C01%7C%7Cf122f3122a244133143308da48962d84%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C637902107101785825%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Csmdm8wELv3aKeKlHLljUI4eIs5fH8UKEJFh2SCk9Kc%3D&reserved=0>, or unsubscribe<https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FABR65P4NZAC4U3P7ZSZZC3TVN5PRFANCNFSM5YDIGANA&data=05%7C01%7C%7Cf122f3122a244133143308da48962d84%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C637902107101785825%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=ZEqkxE39TrV6eWCALOmeeG4cTCjb5EY4ZbchudfhbGg%3D&reserved=0>.
You are receiving this because you are subscribed to this thread.Message ID: ***@***.******@***.***>>
|
|
Was there some censoring done by github? Your message reads: Then install my private version from GitHub: |
|
Addressed by #44 |
|
Hi Carl, the update is working great! Sorry for taking so long to try it out. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hello,
I want to use the numpy inspired API to open an bgen using a symbolic link. I want to do this because I don't want to create the .metadata2.mmm file in the shared directory where the bgen lives. I thought I recalled being able to do this in a previous version of bgen-reader, but am running into the following problem now.
from bgen_reader import open_bgen
wha = open_bgen('/n/holylfs05/LABS/liang_lab_l3/Lab/jordan/ssctpr/data/ukbb_symbolic_link/ukb_imp_chr1_v3.bgen')
Traceback (most recent call last):
File "", line 1, in
File "/n/home_fasse/jrossen/.conda/envs/ssctpr2/lib/python3.9/site-packages/bgen_reader/_bgen2.py", line 133, in init
assert_file_exist(filepath)
File "/n/home_fasse/jrossen/.conda/envs/ssctpr2/lib/python3.9/site-packages/bgen_reader/_file.py", line 14, in assert_file_exist
raise FileNotFoundError(errno.ENOENT, os.strerror(errno.ENOENT), filepath)
I don't have a very strong understanding of how PosixPath works, but after a little playing around, it seems like PosixPath precludes the use of symbolic links.
Either allowing for the use of symbolic links, or allowing the user to specify a directory where .metadata2.mmm should be created would solve my problem completely.
Thanks for making this software!
The text was updated successfully, but these errors were encountered: