Closed malcook closed 2 years ago
I had the same problem before, I uninstalled cellbrowser in conda and then "pip install cellbrowser". Then the problem was solved
Thanks, @ckfromCN , I'm hoping to find a resolution whilst maintaining the conda install, though I may undo/redo as you suggest as a workaround.
We don't manage the bioconda install, but probably we should. Is there a reason Malcom why you prefer conda over pip?
On Sun, Jan 2, 2022 at 4:35 AM Malcolm Cook @.***> wrote:
Thanks, @ckfromCN https://github.com/ckfromCN , I'm hoping to find a resolution whilst maintaining the conda install, though I may undo/redo as you suggest as a workaround.
— Reply to this email directly, view it on GitHub https://github.com/maximilianh/cellBrowser/issues/233#issuecomment-1003658392, or unsubscribe https://github.com/notifications/unsubscribe-auth/AACL4TNYFN6GANC24XRLRKLUT7BX3ANCNFSM5KVPS5OA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.
You are receiving this because you are subscribed to this thread.Message ID: @.***>
I am flexing my conda muscles over pip as it appears to me to be the more generic of the two (e.g. not just addressing python).
I could I suppose use pip within a conda env.
FWIW: If I had my druthers, I would rather than conda or pip be managing scientific computing software installations with guix, but it is not sufficiently broadly used (yet?) and requires buy-in from my IT group which I don't (yet?) have. Anyway, this opinion is mainly based on reading, and not (yet?) experience.
I guess we could update the conda recipe through a PR. I have a note in the README that the conda recipe is usually out of date, as it's managed by the EBI and the EBI doesn't track all our releases. Let me know if this is a problem, I can look into it. Probably we should trigger a conda update whenever we release through pip (they should be a way to achieve this automatically)
On Mon, Jan 3, 2022 at 4:00 PM Malcolm Cook @.***> wrote:
I am flexing my conda muscles over pip as it appears to me to be the more generic of the two (e.g. not just addressing python).
I could I suppose use pip within a conda env.
FWIW: If I had my druthers, I would rather than conda or pip be managing scientific computing software installations with guix, but it is not sufficiently broadly used (yet?) and requires buy-in from my IT group which I don't (yet?) have. Anyway, this opinion is mainly based on reading, and not (yet?) experience.
— Reply to this email directly, view it on GitHub https://github.com/maximilianh/cellBrowser/issues/233#issuecomment-1004149574, or unsubscribe https://github.com/notifications/unsubscribe-auth/AACL4TPBLFMXOGPLT4VZXLDUUG2YVANCNFSM5KVPS5OA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.
You are receiving this because you commented.Message ID: @.***>
I think the conda version should be the same as pip. One of the last times I asked @pcm32 he said conda should auto-update to reflect the latest in pip.
conda installs v1.1.1 of the cell browser, which is the same as pip version (minus the pre-release versions), so it's not clear why this error pops up only for the conda version.
Oops, so then ignore what I said Malcolm! Sounds like Matt is much more on top of the conda questions than I am.
On Mon, Jan 3, 2022 at 5:23 PM Matt Speir @.***> wrote:
I think the conda version should be the same as pip. One of the last times I asked @pcm32 https://github.com/pcm32 he said conda should auto-update to reflect the latest in pip.
conda installs v1.1.1 of the cell browser, which is the same as pip version (minus the pre-release versions), so it's not clear why this error pops up only for the conda version.
— Reply to this email directly, view it on GitHub https://github.com/maximilianh/cellBrowser/issues/233#issuecomment-1004204776, or unsubscribe https://github.com/notifications/unsubscribe-auth/AACL4TN7HQB7XJSRIZOPM33UUHEQJANCNFSM5KVPS5OA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.
You are receiving this because you commented.Message ID: @.***>
@matthewspeir if there is any improved diagnostic (or screenshare?) I can provide to help you help me further sleuth this issue, I'm at glad to oblige...
If I decide to try reinstalling (using pip or otherwise) I will report result here.
For now my successful workaround for creating custom cbs is a hybrid of R interface and CLI.
I can replicate the error with the conda version. It's just not clear why the conda version would be experiencing this error when the pip version seems to be fine.
Looking at the _version.py files from the zip files for both conda and pip seem to state that they're both based on Braney's commit 7cef644bb9e843c3e0fb0d8fedce8ffa18be4212. But clearly there's something going on with how the conda file is generated that's leading to this cbImportSeurat2Cli reference leaking into the conda version as it seems the reference to to cbImportSeurat2Cli was removed all the way back in 8e8df7bba6f5b5d36438a9613da2f54440cbc733 from 2019?
@maximilianh any ideas as to why? Or do we need @pcm32 to chime in with more details about how the conda version is built and where it grabs tools like cbImportSeurat from (as they don't seem to be in the zips from neither conda or pip)?
The executable scripts are created by pip and conda. Weirdly, conda doesn't use the pip definitions, but has its own. Since I changed one of the function names, this broke the conda recipe. I've created a PR now on bioconda and hope that they'll add it soon which should fix this issue. Thanks Malcolm for reporting this! you can watch progress on the PR here: https://github.com/bioconda/bioconda-recipes/pull/32270
On Mon, Jan 3, 2022 at 8:43 PM Matt Speir @.***> wrote:
I can replicate the error with the conda version. It's just not clear why the conda version would be experiencing this error when the pip version seems to be fine.
Looking at the _version.py files from the zip files for both conda https://anaconda.org/bioconda/ucsc-cell-browser/1.1.1/download/noarch/ucsc-cell-browser-1.1.1-pyhdfd78af_0.tar.bz2 and pip https://files.pythonhosted.org/packages/77/54/8a19205e3cffbfb62469695619026b0a511f720c96d402a9617784753635/cellbrowser-1.1.1.tar.gz seem to state that they're both based on Braney's commit 7cef644 https://github.com/maximilianh/cellBrowser/commit/7cef644bb9e843c3e0fb0d8fedce8ffa18be4212. But clearly there's something going on with how the conda file is generated that's leading to this cbImportSeurat2Cli reference leaking into the conda version as it seems the reference to to cbImportSeurat2Cli was removed all the way back in 8e8df7b https://github.com/maximilianh/cellBrowser/commit/8e8df7bba6f5b5d36438a9613da2f54440cbc733 from 2019?
@maximilianh https://github.com/maximilianh any ideas as to why? Or do we need @pcm32 https://github.com/pcm32 to chime in with more details about how the conda version is built and where it grabs tools like cbImportSeurat from (as they don't seem to be in the zips from neither conda or pip)?
— Reply to this email directly, view it on GitHub https://github.com/maximilianh/cellBrowser/issues/233#issuecomment-1004314466, or unsubscribe https://github.com/notifications/unsubscribe-auth/AACL4TNAE32PI4XQM3MBXXDUUH34LANCNFSM5KVPS5OA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.
You are receiving this because you were mentioned.Message ID: @.***>
I'm sorry, I had a long holiday break. I have added some minor changes to Max's PR on bioconda, hopefully that should clear CI lint/tests and enable a merge.
Could someone try to see if conda installing using the artifact generated by the bioconda CI the issue gets resolved? The artifacts are here, you will need to unzip the linux download and point conda to the tar.gz within images: https://dev.azure.com/bioconda/bioconda-recipes/_build/results?buildId=9290&view=artifacts&pathAsName=false&type=publishedArtifacts (this comes from https://github.com/bioconda/bioconda-recipes/pull/32270/checks?check_run_id=4704074100 ).
Sorry what are artifacts? "point conda to the tar.gz within images" ? Or was this comment for someone else than me?
On Tue, Jan 4, 2022 at 5:21 PM Pablo Moreno @.***> wrote:
Could someone try to see if conda installing using the artifact generated by the bioconda CI the issue gets resolved? The artifacts are here, you will need to unzip the linux download and point conda to the tar.gz within images: https://dev.azure.com/bioconda/bioconda-recipes/_build/results?buildId=9290&view=artifacts&pathAsName=false&type=publishedArtifacts (this comes from https://github.com/bioconda/bioconda-recipes/pull/32270/checks?check_run_id=4704074100 ).
— Reply to this email directly, view it on GitHub https://github.com/maximilianh/cellBrowser/issues/233#issuecomment-1004949459, or unsubscribe https://github.com/notifications/unsubscribe-auth/AACL4TLZG5HAEXSJBS7R2YDUUMNBJANCNFSM5KVPS5OA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.
You are receiving this because you were mentioned.Message ID: @.***>
I think everything seems to be fixed here? I installed the cell browser via conda and was able to run cbImportSeurat w/o arguments to see the usage message.
@pcm32 We just updated Cell Browser on pypi to 1.2.0, but it doesn't look like the conda version is updated: https://anaconda.org/bioconda/ucsc-cell-browser? I was hoping to test the conda install after a new Cell Browser release just to be sure everything was fixed.
It might be that someone needs to approve the automatic PR with the auto bump on bioconda, did you check that?
I don't see the auto bump PR, when was it updated?
It was updated on Friday, May 13th.
Thanks for taking a look!
Well, it looks like the Cell Browser in conda still hasn't updated to 1.2.0, but it looks like the original error this issue was made for is fixed with the 1.1.1 version available, so closing this for now.
is occurring whether calling cbImportSeurat with or without command-line options
ucsc-cell-browser was installed on CentOS7 into
using
resulting in
in case it helps:
FWIW: I was successful running
cbBuild
and browse the results using the examplemini
data set.