fix: make omicsintegrator2 work offline

docker-wrappers/OmicsIntegrator2/0001-disable-graph-exports.patch

@@ -0,0 +1,26 @@
+From 5dc0e69fa3d1049ae8e1d8f51859335910245ad7 Mon Sep 17 00:00:00 2001
+From: "Tristan F.-R." <[email protected]>
+Date: Mon, 26 May 2025 09:52:34 -0700
+Subject: [PATCH] fix: disable graph exports
+
+this allows OI2 to work offline; plus, SPRAS already has graph visualizers.
+---
+ src/__main__.py | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/src/__main__.py b/src/__main__.py
+index 49ef402..b7afbca 100644
+--- a/src/__main__.py
++++ b/src/__main__.py
+@@ -80,7 +80,7 @@ def main():
+         forest, augmented_forest = graph.output_forest_as_networkx(vertex_indices, edge_indices)
+
+     #oi.output_networkx_graph_as_graphml_for_cytoscape(augmented_forest, args.output_dir)
+-    oi.output_networkx_graph_as_interactive_html(augmented_forest, args.output_dir, args.filename+'.html')
++    # oi.output_networkx_graph_as_interactive_html(augmented_forest, args.output_dir, args.filename+'.html')
+     augmented_forest_df = oi.get_networkx_graph_as_dataframe_of_edges(augmented_forest)
+     output_dataframe_to_tsv(augmented_forest_df, args.output_dir, args.filename+'.tsv')
+
+-- 
+2.47.0
+

docker-wrappers/OmicsIntegrator2/Dockerfile

@@ -2,5 +2,14 @@
 # https://github.com/fraenkel-lab/OmicsIntegrator2
 FROM continuumio/miniconda3:4.9.2
 
+COPY 0001-disable-graph-exports.patch .
+
+RUN git clone https://github.com/agitter/OmicsIntegrator2 && \
+    cd OmicsIntegrator2 && \
+    git reset --hard 568f170eae388e42e923c478ac9f3308b487760b && \
+    git config user.email "[email protected]" && \
+    git config user.name "Non-existent User" && \
+    git am /0001-disable-graph-exports.patch
+
 COPY environment.yml .
 RUN conda env update --name base --file environment.yml --prune

docker-wrappers/OmicsIntegrator2/README.md

@@ -25,6 +25,7 @@ The Docker wrapper can be tested with `pytest`.
 ## Versions:
 - v1: Created a named conda environment in the container and used `ENTRYPOINT` to execute commands inside that environment. Not compatible with Singularity.
 - v2: Used the environment file to update the base conda environment so the `ENTRYPOINT` command was no longer needed. Compatible with Singularity.
+- v3: Patch to work offline by never running `output_networkx_graph_as_interactive_html` ([#226](https://github.com/Reed-CompBio/spras/pull/226))
 
 ## TODO
 - Attribute https://github.com/fraenkel-lab/OmicsIntegrator2

docker-wrappers/OmicsIntegrator2/environment.yml

@@ -18,4 +18,4 @@ dependencies:
       - pcst_fast==1.0.7
       - goenrich==1.7.0
       - axial==0.1.10
-      - git+https://github.com/agitter/OmicsIntegrator2@568f170eae388e42e923c478ac9f3308b487760b
+      - git+file:///OmicsIntegrator2

docs/contributing/patching.rst

@@ -0,0 +1,23 @@
+Patching Algorithms
+===================
+
+Some wrapped algorithms require extra fixes inside their code. For permissively licensed algorithms,
+we use ``.patch`` files generated from ``git format-patch``.
+
+To create patch files using ``git format-patch`` (assuming your wrapped algorithm is in a git repository):
+
+#. Clone the repository locally.
+#. Commit the changes you want to make (with good commit messages and descriptions).
+
+    * Distinct changes should be made in different commits to make patch files easy to read
+    * For removing code, we prefer to comment out code instead of removing it, to make potential stacktraces less confusing for end users.
+
+#. Run ``git format-patch HEAD~[N]`` where ``N`` is the number of commits you made.
+
+To use ``.patch`` files in ``Dockerfiles``, we create a fake user for ``git`` and apply the patch files using ``git am``:
+
+.. code:: shell
+
+    git config user.email "[email protected]"
+    git config user.name "Non-existent User"
+    git am /0001-my-patch.patch

docs/index.rst

@@ -55,6 +55,7 @@ methods (PRMs) to omics data.
 
    contributing/index
    contributing/maintain
+   contributing/patching
 
 Indices and tables
 ==================

spras/containers.py

@@ -132,31 +132,32 @@ def env_to_items(environment: dict[str, str]) -> Iterator[str]:
 # TODO consider a better default environment variable
 # Follow docker-py's naming conventions (https://docker-py.readthedocs.io/en/stable/containers.html)
 # Technically the argument is an image, not a container, but we use container here.
-def run_container(framework: str, container_suffix: str, command: List[str], volumes: List[Tuple[PurePath, PurePath]], working_dir: str, out_dir: str | os.PathLike, environment: Optional[dict[str, str]] = None):
+def run_container(framework: str, container_suffix: str, command: List[str], volumes: List[Tuple[PurePath, PurePath]], working_dir: str, out_dir: str | os.PathLike, environment: Optional[dict[str, str]] = None, network_disabled = False):
     """
     Runs a command in the container using Singularity or Docker
     @param framework: singularity or docker
     @param container_suffix: name of the DockerHub container without the 'docker://' prefix
     @param command: command to run in the container
     @param volumes: a list of volumes to mount where each item is a (source, destination) tuple
     @param working_dir: the working directory in the container
-    @param environment: environment variables to set in the container
     @param out_dir: output directory for the rule's artifacts. Only passed into run_container_singularity for the purpose of profiling.
+    @param environment: environment variables to set in the container
+    @param network_disabled: Disables the network on the container. Only works for docker for now. This acts as a 'runtime assertion' that a container works w/o networking.
     @return: output from Singularity execute or Docker run
     """
     normalized_framework = framework.casefold()
 
     container = config.config.container_prefix + "/" + container_suffix
     if normalized_framework == 'docker':
-        return run_container_docker(container, command, volumes, working_dir, environment)
+        return run_container_docker(container, command, volumes, working_dir, environment, network_disabled)
     elif normalized_framework == 'singularity' or normalized_framework == "apptainer":
         return run_container_singularity(container, command, volumes, working_dir, out_dir, environment)
     elif normalized_framework == 'dsub':
         return run_container_dsub(container, command, volumes, working_dir, environment)
     else:
         raise ValueError(f'{framework} is not a recognized container framework. Choose "docker", "dsub", or "singularity".')
 
-def run_container_and_log(name: str, framework: str, container_suffix: str, command: List[str], volumes: List[Tuple[PurePath, PurePath]], working_dir: str, out_dir: str | os.PathLike, environment: Optional[dict[str, str]] = None):
+def run_container_and_log(name: str, framework: str, container_suffix: str, command: List[str], volumes: List[Tuple[PurePath, PurePath]], working_dir: str, out_dir: str | os.PathLike, environment: Optional[dict[str, str]] = None, network_disabled=False):
     """
     Runs a command in the container using Singularity or Docker with associated pretty printed messages.
     @param name: the display name of the running container for logging purposes
@@ -166,14 +167,15 @@ def run_container_and_log(name: str, framework: str, container_suffix: str, comm
     @param volumes: a list of volumes to mount where each item is a (source, destination) tuple
     @param working_dir: the working directory in the container
     @param environment: environment variables to set in the container
+    @param network_disabled: Disables the network on the container. Only works for docker for now. This acts as a 'runtime assertion' that a container works w/o networking.
     @return: output from Singularity execute or Docker run
     """
     if not environment:
         environment = {'SPRAS': 'True'}
 
     print('Running {} on container framework "{}" on env {} with command: {}'.format(name, framework, list(env_to_items(environment)), ' '.join(command)), flush=True)
     try:
-        out = run_container(framework=framework, container_suffix=container_suffix, command=command, volumes=volumes, working_dir=working_dir, out_dir=out_dir, environment=environment)
+        out = run_container(framework=framework, container_suffix=container_suffix, command=command, volumes=volumes, working_dir=working_dir, out_dir=out_dir, environment=environment, network_disabled=network_disabled)
         if out is not None:
             if isinstance(out, list):
                 out = ''.join(out)
@@ -199,7 +201,7 @@ def run_container_and_log(name: str, framework: str, container_suffix: str, comm
         raise err
 
 # TODO any issue with creating a new client each time inside this function?
-def run_container_docker(container: str, command: List[str], volumes: List[Tuple[PurePath, PurePath]], working_dir: str, environment: Optional[dict[str, str]] = None):
+def run_container_docker(container: str, command: List[str], volumes: List[Tuple[PurePath, PurePath]], working_dir: str, environment: Optional[dict[str, str]] = None, network_disabled=False):
     """
     Runs a command in the container using Docker.
     Attempts to automatically correct file owner and group for new files created by the container, setting them to the
@@ -244,6 +246,7 @@ def run_container_docker(container: str, command: List[str], volumes: List[Tuple
                                 stderr=True,
                                 volumes=bind_paths,
                                 working_dir=working_dir,
+                                network_disabled=network_disabled,
                                 environment=environment).decode('utf-8')
 
     # TODO does this cleanup need to still run even if there was an error in the above run command?
@@ -278,6 +281,7 @@ def run_container_docker(container: str, command: List[str], volumes: List[Tuple
                                 stderr=True,
                                 volumes=bind_paths,
                                 working_dir=working_dir,
+                                network_disabled=network_disabled,
                                 environment=environment).decode('utf-8')
 
     # Raised on non-Unix systems

spras/omicsintegrator2.py

@@ -128,14 +128,15 @@ def run(edges=None, prizes=None, output_file=None, w=None, b=None, g=None, noise
         if seed is not None:
             command.extend(['--seed', str(seed)])
 
-        container_suffix = "omics-integrator-2:v2"
+        container_suffix = "omics-integrator-2:v3"
         run_container_and_log('Omics Integrator 2',
                              container_framework,
                              container_suffix,
                              command,
                              volumes,
                              work_dir,
-                             out_dir)
+                             out_dir,
+                             network_disabled=True)
 
         # TODO do we want to retain other output files?
         # TODO if deleting other output files, write them all to a tmp directory and copy