search

mindsdb / mindsdb

The platform for building AI from enterprise data
https://mindsdb.com
Other
26.26k stars 4.79k forks source link

Building from source? #247

Closed koyo-jakanees closed 5 years ago

koyo-jakanees commented 5 years ago

Just a thought. I noticed that in there is no documentation neither in the doc or the repo readMe on building from source or cloning, this might seem synonymous to python developers, but arguably not everybody. Really not an immediate issue, but would rather be nice to have out there. include dependencies.

More so to those just shifting to data science and python programming and dev @George3d6 thoughts?

koyo-jakanees commented 5 years ago

like the SKlearn-contrib docs

George3d6 commented 5 years ago

Hi,

There are some docs for contributing here: https://mindsdb.github.io/mindsdb/docs/faq

But they are not obvious right away and they are incomplete.

In terms of installing from source... well, it's one command pip3 install -r requirements.txt + possible custom installation for the given the distro of pytroch and tensorflow + adding to PYTHONPATH (or equivalent env variable on windows). So that should be pretty easy to add... I think we should probably just add that here: https://mindsdb.github.io/mindsdb/docs/installing-mindsdb

But as far as the contributing docs go, well, I think everything might be moving a bit too fast at the moment for us to be able to get some solid docs on that.

For example I just update this: https://mindsdb.github.io/mindsdb/docs/inside-mindsdb 5 days ago, and just now I made a PR which will invalidate some of that...

I shall try to look at the sklearn contributing docs and see if I can get something similar going this month, because I do agree that the good first issue tag on github is not enough to encourage quality code contributions from the community.

If you have any additional feedback on this let me know.

koyo-jakanees commented 5 years ago

@George3d6 i think the sklearn docs might be too comprehensive since it is stable, thus something similar to the ludwig's installation can suffice at the moment but specific to mindsdb. A snippet from ludwig:

Installation¶
Ludwig's requirements are the following:

numpy
pandas
scipy......
sudo apt install python3  # on ubuntu
brew install python3      # on mac

You may want to use a virtual environment to maintain an isolated Python environment.

In order to install Ludwig just run:

pip install ludwig

or install it by building the source code from the repository:

git clone git@github.com:uber/ludwig.git
cd ludwig
virtualenv -p python3 venv
source venv/bin/activate
pip install -r requirements.txt
python setup.py install
koyo-jakanees commented 5 years ago

i have decided to reinstall in a venv then i can write up some bit

George3d6 commented 5 years ago

Hmh, I think ludwig should be auto-installed if you install the requirements.txt file, is it not for you ?

As far as installing ludwig from source, I'm not sure it's worth covering, since we'll be soon moving the standard backend to our own framework (lightwood), which we can develop more closely.

koyo-jakanees commented 5 years ago

sorry @George3d6 for not being specific, i was simply giving the ludwig snippet as an example of simple documentation that could use not including the installing it as part of mindsdb docs

George3d6 commented 5 years ago

Ah, sorry, I miss-read your comment.

George3d6 commented 5 years ago

But yes, I agree, I will add a way to run from source to the docs, the rest should already be in there.

George3d6 commented 5 years ago

Contributing instructions to be added very soon in the next PR, closing this issue.