Problem with conda environment between build steps

Hey everyone, 

I created a build configuration and can't figure out how to connect the dots here. 

My configuration is supposed to run a python project using conda. It consists of these steps:

1. Command-Line script to activate and update the conda environment for this project (there are three projects and the environment has to be activated accordingly)

2. Python Runner to run pytest

3. Qodana analysis

The Commandline script

The command line script runs first and looks like this (some of the things are a bit experimental):

sudo ln -sf bash /bin/sh                #makes sure that the agent uses bash, since the default is dash
ls -l `which sh`                        # print to confirm
. /opt/miniconda/etc/profile.d/conda.sh # init conda for bash and reload the environment
conda activate <my-environments-name>   # activate the environment
conda env update                        # update the dependencies                            
conda list                              # print the packages in this environment

The pytest-step

The pytest step runs with the following configuration:

• Command: pytest
• Test-reporting: true
• Python version: <Custom>
• custom path: /opt/miniconda/bin/python
• Environment tool: none

Running this step throws an "Module not found"-error for one of the dependencies. But I see in the list in step 1, that the dependency has been installed. 

I also tried to set the python path in the second step to conda instead of python, but this will prevent the step from loading the stuff for test-reporting.

What am I missing? I am fairly new to TeamCity (and conda) and could need some help here. What am I doing wrong or what is a proper way to achieve what I want to do? 

Thanks in advance!