Not able to install Point Cloud Library onto python using conda - python

Hi all I am having a hard time installing the PCL or point cloud library onto python using the function conda install -c conda-forge pcl. I am getting all types of errors and it is not installing pcl properly and there is something about a brew? Please help
I have tried many things and even looked at the Point Cloud Library website, but I can't seem to find anything

You can try other channels instead too, sometimes it works, check here for other channels:
pcl anaconda three channels
If that still doesnot work then try in anaconda powershell pip install pcl.
I hope this will help.

As per documentation of the python package:
conda install -c conda-forge -c davidcaron pclpy
The package you installed (pcl) - it is just the binary dependencies. You need to install pclpy, which is the python wrapper.

Related

Cannot install scikit-surprise on my jupyter notebook

I am building a recommendation engine and am not able to install surprise, i thought the problem was because i didn't have a c compiler(since i found some people saying it would solve the problem) so i installed it too but couldn't solve anything. i even tried runing the pip install on my command prompt and it gave an error.
Some users have suggested that i use '!' before pip but pip install works fine for me all the time as u see in the picture for numpy.
the solution i found for this was
Step 1 : Go to anaconda command prompt
Step 2 : Type conda install -c conda-forge scikit-surprise
i hope this is helpful to anyone who encounters this error in the future.
I found the solution, change your enviroment interpreter to python <= 3.7 version
scikit-surprise it´s not support for greater version of python.
For more information see also https://pypi.org/project/scikit-surprise/
I found the solution for me as to first update all my conda packages and then install scikit-surprise. You can follow the steps as well if it works out for you:
Go to anaconda command prompt
Type conda update --all
Type conda install -c conda-forge scikit-surprise
Also I'm running python 3.8.8.
The key error is in Microsoft Visual C++ 14.0 or greater is required. You have to download supporting build tools for C++(sometimes just downloading it from MS website wont work), if you are using visual studio 2022 for python development, you have to modify your installer and download missing build tools, I think its titled "Desktop Development with C++ (not just python developer/IDE) and check almost all tools", however it consumes bit of system space, however it sorted this problem of installing many packages in python which was not working otherwise.
I was able to install it on python 3.7 after installing python3.7-dev
I found the solution in this link https://github.com/NicolasHug/Surprise/issues/283#issuecomment-515878185
Try this in shell: pip install scikit-surprise==1.0.4

No module named "fastai"

I'm trying to use fastai to figure out an optimal learning rate for my neural network. Everything else is working fine I'm just not quite getting the accuracy I want. So I'm trying to use the following lines of code to optimize my learning rate:
learn.lr_find()
learn.sched.plot_lr()
So I pip installed fastai and everything seemed like it installed correctly and into the correct directory, but every time I try to import fastai, I can't. I included pictures of my command prompt and the error message. Thank you all for the help in advance I really appreciate it. If I didn't provide enough info just let me know. I'm new to asking questions on here.
Error Message
Installation of Package
From the screenshots provided, it seems like you didn't install FastAi properly and you were installing PyTorch instead. FastAi comes with a PyTorch build. So you don't need to install PyTorch separately.
You can install FastAi with Pip by the following command
pip install fastai
If you are using Python 3.x sometimes you might need to use pip3 command instead
pip3 install fastai
Another problem might be Python mismatch. Maybe you've installed the both version of Python and you distro couldn't verify from where to grab the package. So make sure to use the correct Python and Pip version
I had a similar issue where I installed fastai with pip. I got the 'No module named fastai' error whenever I tried to import fastai, but if I did "pip freeze | grep fastai" in the terminal, it showed that fastai was clearly installed.
The solution for me was to download anaconda3, enter an anaconda environment, and reinstall fastai with conda using step 6 of fastai's setup instructions for using AWS EC2 instances.
Useful links:
https://docs.anaconda.com/anaconda/install/linux/
https://course.fast.ai/start_aws.html
conda install -c fastai fastai
I would recommend waiting for the new release of fastai, they are currently working on version 2 in July.
Learner.lr_find(start_lr=1e-07, end_lr=10, num_it=100, stop_div=True, show_plot=True, suggestions=True)

What's the difference between conda skeleton and conda install?

Following this tutorial conda skeleton should be used for installing non r-essential R packages. however this hasn't worked for me. However after searching online this post suggests to use conda install -c r r-packagename which worked for me. So what is the difference between conda skeketon and conda install?
The skeleton command is used to jump-start the building of your own package. The install command is for installing someone else's package (or one that you previously finished.)

Path confusion using Conda, PIP, and Jupyter notebook at the same time

I've been using conda and jupyter notebook for a while now and recently I need some packages that are only available for install via pip. My understanding was that they are completely compatible and transparent with each other but apparently not.
I did a pip install pydicom and when I tried to do an import pydicom in my Jupyter Notebook, it says No module named pydicom.
In command prompt (using a Windows machine here) I ran conda list and I see pydicom is listed and marked as < pip >.
In my Jupyter notebook, this is my sys.executable output:
C:\Users\USERNAME\AppData\Local\Continuum\Anaconda2\python.exe
Is that normal/correct???
Disclaimer: I'm thoroughly confused by how the paths, environment, and dependencies work in conda and pip. I can't find a good tutorial or explanation how conda/pip work behind the scenes. Any pointers will be greatly appreciated!
While the package is named pydicom, you need to import dicom:
import dicom
You can also in install pydicom via conda. You need to provide the channel 'conda-forge`:
conda install -c conda-forge pydicom
You can find more packages a https://anaconda.org/. Just enter the package name in the search window. If the package is available you will see the channel you can with the -c option.

How do I go about installing packages into Anaconda?

I've been looking everywhere and cannot find a robust explanation.
I'm brand new to Python, coming from R. I had no issues installing packages there but I'm finding it to be rather confusing in Python.
So, I'm using Anaconda and I want to install this package into Python. It mentions using the conda install -c https://conda.anaconda.org/amueller wordcloud command but I have no idea where I'm supposed to run it.
Any advice would be much appreciated.
If you can't find a package with a simple conda search from the command line, run a search on the Anaconda website.
In your case, you'll find that the contributor amueller has his own channel and the package wordcloud is available.
Just run conda install -c amueller wordcloud=1.2.1 to install it.
You might want to create a separate environment using conda create first.
Run it from command line. You can directly install using pip as mentioned in the link. Don't forget to install the pre-requisite packages.
pip install wordcloud
If you are using windows, make sure your environmental path is set so that you can use the pip command directly from windows command prompt. Usually the environmental variable is updated when you install Anaconda distribution.
Conda is just another command which you can use to install packages. Procedure is exactly the same.
You may install packages from interface
Interface

Categories