You are viewing a single comment's thread from:
RE: Install steem-python easily by conda-forge
What do we need to do to get this to work with PIP? Unfortunately for me, the majority of my toolchain is managed by PIP and I kind of need it for keeping things maintained. As such, I can't trivially make the jump to Anaconda.
(I knew making the jump to Python 3.6 was going to be controversial, but I didn't realize how many tools had very subtle inconsistencies that it would take a long time to work out. I also didn't expect that the Python steem module was going to be one of those.)
You could also try beem, a python library for steem which has very few dependencies.
https://steemit.com/utopian-io/@holger80/next-update-for-beem-a-python-library-for-steem
This looks very promising indeed. Thank you very much; I'm pretty sure I can work with this.
Are you talking about PIP on windows? You need visual studio installed.
The main problem is to install pycrypto with pip on windows. If you getting errors recarding import winrandom then you should apply the patches from here: https://github.com/conda-forge/pycrypto-feedstock/tree/master/recipe.
Do you need CLI? This is also a problem.
Yes, PIP on Windows.
The problem with Visual Studio is that it refuses to install the bulk of its code on the system install drive, which in my case (and a lot of people's cases) is an SSD with relatively limited space beyond the operating system core itself. And no, you can't tell it to install elsewhere. It's been brought up to Microsoft that this is perhaps one of the worst install non-options ever seen, to which their response is a dead eyed stare.
pycrypto is the problem, no doubt. The fact that it is an abandoned project only multiplies the issues. Apparently, the migration path is to go off to pycryptodome, but reengineering the steem module to use a different crypto hardness is well beyond the level of "being responsible" that I want to get into.
Really, what I want is a decent Python-side module for dealing with a lot of steem data. The MongoDB interface on SteamData is very nice, and I like it a lot – but it doesn't allow for interaction and that would be nice.