Installation of SLM-LAB

[diarmuid-corcoran]

Hi,
Great book! Well done!

I have been unable to install the SLM-LAB environment by following the instructions and running ./bin/setup.
I have attempted many times on both ubuntu and macOS. Am I missing an important setup step somewhere?

On mac I get the following error on running "python run_lab.py slm_lab/spec/demo.json dqn_cartpole train"

Traceback (most recent call last):
File "run_lab.py", line 80, in
main()
File "run_lab.py", line 72, in main
read_spec_and_run(*args)
File "run_lab.py", line 50, in read_spec_and_run
spec = spec_util.get(spec_file, spec_name)
File "/Users/XXXXX/git/SLM-Lab/slm_lab/spec/spec_util.py", line 154, in get
spec_dict = util.read(spec_file)
File "/Users/XXXXX/git/SLM-Lab/slm_lab/lib/util.py", line 412, in read
data = read_as_plain(data_path, **kwargs)
File "/Users/XXXXX/git/SLM-Lab/slm_lab/lib/util.py", line 435, in read_as_plain
data = ujson.load(open_file, **kwargs)
ValueError: Unexpected character in found when decoding object value

On Ubuntu I reach the end of the road with the following error:
File "/home/XXXX/git/SLM-Lab/slm_lab/experiment/analysis.py", line 2, in
from slm_lab.spec import random_baseline
File "/home/XXXXX/git/SLM-Lab/slm_lab/spec/random_baseline.py", line 7, in
import roboschool
File "/home/XXXXX/anaconda3/envs/lab/lib/python3.7/site-packages/roboschool/init.py", line 112, in
from roboschool.gym_pendulums import RoboschoolInvertedPendulum
File "/home/XXXXX/anaconda3/envs/lab/lib/python3.7/site-packages/roboschool/gym_pendulums.py", line 1, in
from roboschool.scene_abstract import SingleRobotEmptyScene
File "/home/XXXXX/anaconda3/envs/lab/lib/python3.7/site-packages/roboschool/scene_abstract.py", line 12, in
from roboschool import cpp_household as cpp_household
ImportError: cannot import name 'cpp_household' from 'roboschool' (/home/XXXXX/anaconda3/envs/lab/lib/python3.7/site-packages/roboschool/init.py)

[kengz]

Hi, seems like the first error the json is invalid. Could you make sure it's not edited? One way to do this is to just do git stash and run it again.

For the second error, seems like it's a common issue with roboschool as seen in their issues. Could you try this solution in the lab conda environment? Worst case - you may comment out the roboschool import at SLM-Lab/slm_lab/spec/random_baseline.py line 7 to get you unblocked for the mean time.