Reputation: 1
I have a windows installation and plan to run GPU-accelerated code with RAPIDS 24.12, enabling me to handle dataframes with dask-cudf. RAPIDS Installation with Conda successful, test code works as it should.
Now I would like to install a specific python library: ta-lib, to the venv.
Procedure: (and no, I can’t use pip or conda because, loads of reasons and my grampa has a toothache… I have to build it from source. Moving on...:)
Prep:
create venv with : python3 -m venv <myenvpath>
as usual
login to venv:
Correct syntax to activate venv in ubuntu shell:
<<<< source ~/miniforge3/envs/rapids-24.12/bin/activate
>>>>>>>>>
Download the library at: https://github.com/TA-Lib/ta-lib/archive/refs/tags/v0.6.4.tar.gz
Gently place the library in /home/<username>/miniforge3/envs/rapids-24.12
. If it squeaks, slap it and do it again.
extract the tarball
tar -xzf ta-lib-0.6.4-src.tar.gz
cd to the resulting dir
cd ta-lib-0.6.4
run install script:
Now, first:
`./configure
make
sudo make install`
.....will install the library to a default directory, and not the dir where my venv is housed. I tried that: ta-lib
doesn't show up when i run pip list
Also tried installing ta-lib into other dir than usr/local:
cd to: cd ~/miniforge3/envs/rapids-24.12/ta-lib-0.6.4
then run:
./configure --prefix=/home/<username>/miniforge3/envs/rapids-24.12
make
sudo make install
This makes the installation proceed normally and everything works, as far as I can see from shell output, but when I run pip list, ta-lib is nowhere to be seen.
All of this SUDO wrestling has made me tired and SUDO is winning? I’m all out of ideas. How do I proceeed with this?
Upvotes: 0
Views: 86
Reputation: 2732
I would like to install a specific python library: ta-lib, to the venv
ta-lib
doesn't show up when I runpip list
The approach you took downloading the source of https://github.com/TA-Lib/ta-lib and running make install
builds the Ta-Lib C library.
pip list
only knows about packages installed with pip
and similar tools.
Since it appears that you're using conda
, you can use the published releases on conda-forge
to get the Python package.
conda install \
-c conda-forge \
--name rapids-24.12 \
ta-lib
That package contains built versions of https://github.com/TA-Lib/ta-lib-python, created from https://github.com/conda-forge/ta-lib-feedstock.
Alternatively, you could:
pip install ta-lib
(https://pypi.org/project/ta-lib/)Upvotes: 1