Merge pull request #48 from MetaboHUB-MetaToul-FluxoMet/47-conda-vali…

…dation

implement conda envs validation

README.md

@@ -81,7 +81,7 @@ The raws files you provide should be numbered that way:
 
 The workflow will automatically detect the files and process them. Results (metabolites, concentrations, etc) are in `share/cycle/N/` of the run directory. 
 
-For a more detailed guide, see the [Tutorial](https://rtmet.readthedocs.io/en/latest/getting_started/tutorial.html) in the docs.
+For a more detailed guide, see the [Tutorial](https://rtmet.readthedocs.io/en/latest/getting_started/basic_tutorial.html) in the docs.
 
 ## 🪲 Bugs and feature requests
 

cylc-src/bioreactor-workflow/envs/conda.cylc

@@ -8,16 +8,40 @@
     'CONDA_OPENMS': 'wf-pyopenms',
     } %}
 
+[scheduling]
+    [[graph]]
+{% if cfg__validate_conda_envs %}
+        R1/^ = validate_cfg => validate_conda => is_setup
+{% endif %}
+
 [runtime]
-{% for env, conda_env_name in conda_envs.items() %}
-    [[{{env}}]]
+
+{% for FAMILY, conda_env_name in conda_envs.items() %}
+    [[{{FAMILY}}]]
         pre-script = """
             set +eu
             conda activate {{ conda_env_name }}
             set -eu
         """
         [[[meta]]]
-            title = "{{ env }}"
+            title = "{{ FAMILY }}"
             description = "Conda environment activation: {{ conda_env_name }}"
             categories = family task, conda
-{% endfor %}
+{% endfor %}
+
+    [[validate_conda]]
+        script = """
+            >&2 echo "conda compare outputs errors to stdout. Check 'job.out' for errors."
+{% for conda_env_name in conda_envs.values() %}
+            echo "Checking conda environment: {{ conda_env_name }}"
+            conda compare \
+                ${CYLC_WORKFLOW_RUN_DIR}/envs/{{ conda_env_name }}.yml \
+                -n {{ conda_env_name }}
+{% endfor %}
+        """
+        [[[meta]]]
+            title = "Validate Conda Environments"
+            description = """
+                Check if the conda environments are up-to-date and all packages are present.
+            """
+            categories = family task, conda

cylc-src/bioreactor-workflow/envs/wf-binner.yml

@@ -3,16 +3,12 @@ channels:
   - conda-forge
   - bioconda
 dependencies:
-  - r-base=4.3.1
-  - r-essentials=4.3
   - r-remotes=2.4
-  - r-furrr=0.3.1
-  - r-ggthemes=5.1.0
+  - r-essentials=4.3
   - r-optparse=1.7.4
-  - r-tidyselect=1.2.1
-  - r-rcpp=1.0.12
+  - r-ggthemes=5.1.0
+  - r-furrr=0.3.1
   - bioconductor-mzr=2.36
-  - bioconductor-rhdf5lib=1.24
 
 # Inside R, install the binneR package from GitHub
-# remotes::install_github('aberHRML/binneR')
+# conda activate wf-binner && Rscript -e "remotes::install_github('aberHRML/binneR', upgrade='never')"

cylc-src/bioreactor-workflow/meta/rose-meta.conf

@@ -31,6 +31,10 @@ compulsory=true
 range=0:
 type=real
 
+[template variables=cfg__validate_conda_envs]
+compulsory=true
+type=python_boolean
+
 [template variables=cfg__scans_window]
 compulsory=false
 pattern=^'(|(\d+)-(\d+))'$

cylc-src/bioreactor-workflow/rose-suite.conf

@@ -8,7 +8,37 @@
 # It is rendered directly from the docstrings here.
 
 
-########## Scientific parameters ##########
+
+################################### Global workflow parameters ####################################
+
+# Should the workflow validate the conda environments for packages' presence
+# and versions before running.
+#
+# :default: False
+# :type: bool
+cfg__validate_conda_envs = False
+
+# `internal | local`.
+#
+# * *internal*: will look for .raw files in :file:`${CYLC_WORKFLOW_RUN_DIR}/raws`.
+# * *local*: will look for .raw files in :file:`{cfg__local_runs_dir}/${CYLC_WORKFLOW_RUN_NAME}`.
+#
+# :default: 'internal'
+# :type: str
+cfg__input_strategy='internal'
+
+# Used if :rose:conf:`cfg__input_strategy` is set to *local*. If there is a
+# subdirectory in the one defined here that shares the same name as the
+# :term:`workflow run`, the workflow will look for .raw files inside.
+#
+# :optional: yes
+# :default: None
+# :type: str (path)
+cfg__local_runs_dir=''
+
+
+
+###################################### Scientific parameters ######################################
 
 # Tuple for the number of scans to trim at the start and at the end of the
 # :term:`flowgram`. If set to [0, 0], no trimming will be done.
@@ -42,29 +72,7 @@ cfg__ppm_tol=5
 
 
 
-########## Input Strategy ##########
-
-# `internal | local`.
-#
-# * *internal*: will look for .raw files in :file:`${CYLC_WORKFLOW_RUN_DIR}/raws`.
-# * *local*: will look for .raw files in :file:`{cfg__local_runs_dir}/${CYLC_WORKFLOW_RUN_NAME}`.
-#
-# :default: 'internal'
-# :type: str
-cfg__input_strategy='internal'
-
-# Used if :rose:conf:`cfg__input_strategy` is set to *local*. If there is a
-# subdirectory in the one defined here that shares the same name as the
-# :term:`workflow run`, the workflow will look for .raw files inside.
-#
-# :optional: yes
-# :default: None
-# :type: str (path)
-cfg__local_runs_dir=''
-
-
-
-########## [OPTIONAL] InfluxDB Setup ##########
+#################################### [OPTIONAL] InfluxDB Setup ####################################
 
 # Should the workflow send data to an InfluxDB instance for visualization.
 # If set to *True*, the following configuration options must be set.

docs/source/getting_started/installation.rst

@@ -96,7 +96,7 @@ Bioinformatics tools are installed in separate Conda environments, for isolation
 purposes. For binneR, you'll need to install it from the R console::
 
     $ for file in wf-*.yml; do conda env create -f $file; done
-    $ conda activate wf-binner && Rscript -e "remotes::install_github('aberHRML/binneR', dependencies=FALSE, upgrade_dependencies=FALSE)"
+    $ conda activate wf-binner && Rscript -e "remotes::install_github('aberHRML/binneR', upgrade='never')"
     $ conda deactivate
 
 .. _installation.influxdb: