Completion attempt gets wrong-type-argument (epc:manager nil)

[BigBaaadBob]

GNU Emacs 26.3 (build 1, x86_64-w64-mingw32)

Using miniconda3 virtual environments

Company and company-jedi installed from Melpa:

  company            20201014.2251 installed 
  company-jedi       20200324.25   installed
  pyvenv             20191202.1039 installed 

Prereqs installed:

$ conda list | grep 'epc\|jedi'
epc                       0.0.5                    pypi_0    pypi
jedi                      0.17.2                   pypi_0    pypi

Config:

;; company
(eval-after-load "company"
  '(add-to-list 'company-backends 'company-jedi))
(add-hook 'after-init-hook 'global-company-mode)
(setq jedi:server-command '("~/.emacs.d/elpa/jedi-core-20191011.1750/jediepcserver.py"))

jediepcserver.py works at that path:

$ ~/.emacs.d/elpa/jedi-core-20191011.1750/jediepcserver.py --help
usage: jediepcserver.py [-h] [--address ADDRESS] [--port PORT]
                        [--port-file PORT_FILE] [--sys-path SYS_PATH]
                        [--sys-path-append SYS_PATH_APPEND]
                        [--virtual-env VIRTUAL_ENV] [--log LOG]
                        [--log-level {CRITICAL,ERROR,WARN,INFO,DEBUG}]
                        [--log-rotate-max-size LOG_ROTATE_MAX_SIZE]
                        [--log-rotate-max-count LOG_ROTATE_MAX_COUNT]
                        [--log-traceback] [--pdb] [--ipdb]

Jedi EPC server.

Completion fails with:

*** jedi-mode is disabled in #<buffer hello.py> ***
Fix the problem and re-enable it.

*** You may need to run "M-x jedi:install-server". ***
This could solve the problem especially if you haven't run the command yet
since Jedi.el installation or update and if the server complains about
Python module imports.

I enabled debug on error and get this traceback:

Debugger entered--Lisp error: (wrong-type-argument epc:manager nil)
  signal(wrong-type-argument (epc:manager nil))
  jedi:epc--start-epc("~/.emacs.d/elpa/jedi-core-20191011.1750/jediepcserver.py" nil)
  jedi:server-pool--start(("~/.emacs.d/elpa/jedi-core-20191011.1750/jediepcserver.py"))
  jedi:start-server()
  jedi:get-epc()
  jedi:call-deferred(complete)
  company-jedi-candidates(#f(compiled-function (candidates) #<bytecode 0x2a33485>))
  company--fetch-candidates(#("impo" 0 4 (fontified t)))
  company-calculate-candidates(#("impo" 0 4 (fontified t)) nil)
  company--begin-new()
  company--perform()
  company-auto-begin()
  company-idle-begin(#<buffer hello.py> #<window 3 on hello.py> 75 5)
  apply(company-idle-begin (#<buffer hello.py> #<window 3 on hello.py> 75 5))
  timer-event-handler([t 24471 6093 761307 nil company-idle-begin (#<buffer hello.py> #<window 3 on hello.py> 75 5) nil 0])

This happens nomatter what I put in jedi:server-args.

Company completion seems to be working fine for elisp. I haven't tried any other backends. I appreciate this isn't a standard config, but it is difficult to have multiple venv setups in a Conda environment.