Fix pip install if torch isn't already installed

[anth2o]

Hello !

Currently if you are working on a fresh new environment without torch installed, you have issues when installing this package:

$ pip install git+https://github.com/parlance/ctcdecode.git
Collecting git+https://github.com/parlance/ctcdecode.git
  Cloning https://github.com/parlance/ctcdecode.git to /tmp/pip-req-build-tu1xf42c
  Running command git clone --filter=blob:none -q https://github.com/parlance/ctcdecode.git /tmp/pip-req-build-tu1xf42c
  Resolved https://github.com/parlance/ctcdecode.git to commit c90ad94a0b19554f80804fb7812f2a1447a34a70
  Running command git submodule update --init --recursive -q
  Preparing metadata (setup.py) ... error
  ERROR: Command errored out with exit status 1:
   command: /afs/melodis.com/home/ahoorelbeke/venv/bin/python -c 'import io, os, sys, setuptools, tokenize; sys.argv[0] = '"'"'/tmp/pip-req-build-tu1xf42c/setup.py'"'"'; __file__='"'"'/tmp/pip-req-build-tu1xf42c/setup.py'"'"';f = getattr(tokenize, '"'"'open'"'"', open)(__file__) if os.path.exists(__file__) else io.StringIO('"'"'from setuptools import setup; setup()'"'"');code = f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, __file__, '"'"'exec'"'"'))' egg_info --egg-base /tmp/pip-pip-egg-info-wn60pvg7
       cwd: /tmp/pip-req-build-tu1xf42c/
  Complete output (5 lines):
  Traceback (most recent call last):
    File "<string>", line 1, in <module>
    File "/tmp/pip-req-build-tu1xf42c/setup.py", line 10, in <module>
      from torch.utils.cpp_extension import BuildExtension, CppExtension, include_paths
  ModuleNotFoundError: No module named 'torch'
  ----------------------------------------
WARNING: Discarding git+https://github.com/parlance/ctcdecode.git. Command errored out with exit status 1: python setup.py egg_info Check the logs for full command output.
ERROR: Command errored out with exit status 1: python setup.py egg_info Check the logs for full command output.

This PR suggests different improvements:

• pip install in a single line, useful to add this package in a requirements.txt for instance
• add torch to install_requires in setup.py so that torch is installed after the pip install
• add torch to requires in pyproject.toml so that torch is present before calling the setup.py and we avoid the error above

You can test this by doing:

pip install git+https://github.com/anth2o/ctcdecode.git

[LuluW8071]

Can you list the dependencies like torch version, c++ and others
Also I am facing same error in the windows. Will it be different for linux based ?