Use CLI install script instead of custom instructions

.gitlab-ci.yml

 variables:
-  CLI_CONDA_ENV: "$CI_PROJECT_DIR/.conda_envs/cli.env"
+  CONDA_ENVS: "$CI_PROJECT_DIR/.conda_envs"
+  CLI_CONDA_ENV: "$CONDA_ENVS/cli.env"
+  SCORER_CONDA_ENV: "$CONDA_ENVS/scorer.env"
   PIP_CACHE_DIR: "$CI_PROJECT_DIR/.pip"
 
 default:
@@ -19,8 +21,9 @@ install:
   stage: setup
   script:
     - python3 -m pip install pycodestyle
-    - if [[ -f requirements.txt ]]; then python3 -m pip install -r requirements.txt; python3 -m pip install -e .; fi
-    - if [[ -f environment.yml ]]; then mkdir $CONDA_ENVS; conda create -f environment.yml -p $CLI_CONDA_ENV; fi
+    #- if [[ -f requirements.txt ]]; then python3 -m pip install -r requirements.txt; python3 -m pip install -e .; fi
+    #- if [[ -f environment.yml ]]; then mkdir $CONDA_ENVS; conda create -f environment.yml -p $CLI_CONDA_ENV; fi
+    - bash diva_evaluation_cli/bin/install.sh --all -c $CLI_CONDA_ENV -s $SCORER_CONDA_ENV
 
 test-code:
   stage: test
@@ -28,5 +31,5 @@ test-code:
   script:
     - python3 -m pycodestyle --filename=*.py diva_evaluation_cli/bin/ test/
     - python3 -m unittest discover -s test/
-    - if [[ -f environment.yml ]]; then conda activate $CONDA_ENVS/cli.env; fi
+    - if [[ -f environment.yml ]]; then conda activate $CLI_CONDA_ENV; fi
     - actev -h