Add endpoint to get project workflow

backend/app/api/routes/projects.py

@@ -13,6 +13,7 @@
 import yaml
 from app import users
 from app.api.deps import CurrentUser, SessionDep
+from app.dvc import make_mermaid_diagram
 from app.models import (
     Dataset,
     Figure,
@@ -21,6 +22,7 @@
     ProjectCreate,
     ProjectsPublic,
     Question,
+    Workflow,
 )
 from fastapi import APIRouter, HTTPException, Request
 from fastapi.responses import StreamingResponse
@@ -357,3 +359,27 @@ def post_project_sync(
     # Publications
     # TODO: Update files in Git repo with IDs?
     return Message(message="success")
+
+
+@router.get("/projects/{owner_name}/{project_name}/workflow")
+def get_project_workflow(
+    owner_name: str,
+    project_name: str,
+    current_user: CurrentUser,
+    session: SessionDep,
+) -> Workflow:
+    content = get_project_git_contents(
+        owner_name=owner_name,
+        project_name=project_name,
+        session=session,
+        current_user=current_user,
+        path="/dvc.yaml",
+    )
+    content = base64.b64decode(content["content"]).decode()
+    dvc_pipeline = yaml.safe_load(content)
+    # Generate Mermaid diagram
+    mermaid = make_mermaid_diagram(dvc_pipeline)
+    return Workflow(
+        stages=dvc_pipeline["stages"],
+        mermaid=mermaid,
+    )

backend/app/dvc.py

@@ -0,0 +1,32 @@
+"""Functionality for working with DVC."""
+
+import os
+import tempfile
+from dvc.commands import dag
+from dvc.repo import Repo
+import yaml
+
+
+def make_mermaid_diagram(pipeline: dict) -> str:
+    """Create a Mermaid diagram from a pipeline file (typically ``dvc.yaml``).
+
+    This is a little hacky since we need to create a Git and DVC repo in order
+    to run the commands in DVC.
+    """
+    wd_orig = os.getcwd()
+    try:
+        with tempfile.TemporaryDirectory() as tmpdirname:
+            os.chdir(tmpdirname)
+            with open("dvc.yaml", "w") as f:
+                yaml.safe_dump(pipeline, f)
+            with Repo.init(
+                ".",
+                no_scm=True,
+                force=False,
+                subdir=False,
+            ) as repo:
+                d = dag._build(repo)
+            mm = dag._show_mermaid(d, markdown=False)
+    finally:
+        os.chdir(wd_orig)
+    return mm

backend/app/models.py

@@ -176,6 +176,20 @@ class ProjectCreate(ProjectBase):
     pass
 
 
+class WorkflowStage(SQLModel):
+    cmd: str
+    deps: list[str] | None = None
+    outs: list[str]
+    desc: str | None = None
+    meta: dict | None = None
+    wdir: str | None = None
+
+
+class Workflow(SQLModel):
+    mermaid: str
+    stages: dict[str, WorkflowStage]
+
+
 class Question(SQLModel, table=True):
     id: uuid.UUID = Field(default_factory=uuid.uuid4, primary_key=True)
     project_id: uuid.UUID = Field(foreign_key="project.id")
@@ -219,5 +233,6 @@ class Dataset(SQLModel, table=True):
 
 class ImportedDataset(SQLModel, table=True):
     """A dataset imported into a project in a read-only fashion."""
+
     project_id: uuid.UUID = Field(foreign_key="project.id", primary_key=True)
     dataset_id: uuid.UUID = Field(foreign_key="dataset.id", primary_key=True)

backend/app/tests/test_dvc.py

@@ -0,0 +1,22 @@
+"""Tests for the ``dvc`` module."""
+
+from app.dvc import make_mermaid_diagram
+
+
+def test_make_mermaid_diagram():
+    pipeline = {
+        "stages": {
+            "do-something": {
+                "cmd": "echo sup",
+                "deps": ["somefile.py"],
+                "outs": ["something.png"],
+            },
+            "do-something-else": {
+                "cmd": "echo sup2",
+                "deps": ["something.png"],
+                "outs": ["else.pdf"],
+            },
+        }
+    }
+    mm = make_mermaid_diagram(pipeline)
+    return mm