pip3 install pyknowhere.xx.whl, but "import pyknowhere" Error: ModuleNotFoundError: No module named 'pyknowhere'

[zjx1230]

Hello, Everyone! I want to use the package pyknowhere. I followed github's guide and successfully compiled the package, and after the installation was successful, it still prompted that the python package could not be found, what is going on? The installed package only has the directory pyknowhere-0.0.0.dist-info and no directory pyknowhere Here are the details:

(z00xxxxxxx) [root@node57 python]# pip3 install ./dist/pyknowhere-0.0.0-cp310-cp310-linux_x86_64.whl Looking in indexes: http://mirrors.tools.huawei.com/pypi/simple Processing ./dist/pyknowhere-0.0.0-cp310-cp310-linux_x86_64.whl Installing collected packages: pyknowhere Successfully installed pyknowhere-0.0.0 WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv

(z00xxxxxxx) [root@node57 python]# python Python 3.10.12 (main, Jul 5 2023, 18:54:27) [GCC 11.2.0] on linux Type "help", "copyright", "credits" or "license" for more information. import pyknowhere Traceback (most recent call last): File "", line 1, in ModuleNotFoundError: No module named 'pyknowhere' exit()

(z00xxxxxxx) [root@node57 python]# ls /root/miniconda3/envs/z00878108/lib /python3.10/site-packages/ | grep knowhere* knowhere knowhere-1.0.0.dist-info pyknowhere-0.0.0.dist-info

OS: CentOS 7

[SimFG]

It seems that you should use import knowhere

[zjx1230]

It seems that you should use import knowhere

But, the project 'ann-benchmarks' has source code pyknowhere.Index(), import knowhere report error 'No Attribute'.

[SimFG]

This seems to be due to version inconsistency. I see that 'ann-benchmarks' uses version 1.x, but now knowhere should be version 2.x

[zjx1230]

This seems to be due to version inconsistency. I see that 'ann-benchmarks' uses version 1.x, but now knowhere should be version 2.x

However, when I checkout 1.x branch to generate the pyknowhere in the knowhere project, it still report error: Not Found module. Here is the details:

(z00xxxxx) [root@node57 knowhere]# git branch -v * 1.x c0aa5fb Fix DiskANN's Async cache making (#250)

(z00xxxxx) [root@node57 knowhere]# cd python/

(z00xxxxx) [root@node57 python]# pip3 install ./dist/knowhere-1.0.0-cp310-cp310-linux_x86_64.whl Looking in indexes: http://mirrors.tools.huawei.com/pypi/simple Processing ./dist/knowhere-1.0.0-cp310-cp310-linux_x86_64.whl Installing collected packages: knowhere Successfully installed knowhere-1.0.0 WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv

(z00xxxxx) [root@node57 python]# python Python 3.10.12 (main, Jul 5 2023, 18:54:27) [GCC 11.2.0] on linux Type "help", "copyright", "credits" or "license" for more information. import pyknowhere Traceback (most recent call last): File "", line 1, in ModuleNotFoundError: No module named 'pyknowhere'

[zjx1230]

How to import pyknowhere? I can't use it.

[SimFG]

I think this project 'ann-benchmarks' should use a historical version, but I am not sure. If you want to run it, I suggest you modify the relevant code in 'ann-benchmarks' according to the current knowhere code

[zjx1230]

I think this project 'ann-benchmarks' should use a historical version, but I am not sure. If you want to run it, I suggest you modify the relevant code in 'ann-benchmarks' according to the current knowhere code

Sorry, I can't do it. The work so hard. I think there must be a version of the source code that can generate this corresponding package, otherwise the ann-benchmark project cannot run the algorithm. If anyone knows please let me know, thank you very much!

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. Rotten issues close after 30d of inactivity. Reopen the issue with /reopen.