calkit-python 0.3.0__tar.gz → 0.3.1__tar.gz

{calkit_python-0.3.0 → calkit_python-0.3.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: calkit-python
-Version: 0.3.0
+Version: 0.3.1
 Summary: Reproducibility simplified.
 Project-URL: Homepage, https://github.com/calkit/calkit
 Project-URL: Issues, https://github.com/calkit/calkit/issues

{calkit_python-0.3.0 → calkit_python-0.3.1}/calkit/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.3.0"
+__version__ = "0.3.1"
 
 from .core import *
 from . import git

{calkit_python-0.3.0 → calkit_python-0.3.1}/calkit/cli/office.py

@@ -44,7 +44,19 @@ def excel_chart_to_png(
 
 @office_app.command(name="word-to-pdf", help="Convert a Word document to PDF.")
 def word_to_pdf(
-    input_fpath: Annotated[str, typer.Argument(help="Input Excel file path.")],
-    output_fpath: Annotated[str, typer.Argument(help="Output PNG file path.")],
+    input_fpath: Annotated[
+        str, typer.Argument(help="Input Word document file path.")
+    ],
+    output_fpath: Annotated[
+        str,
+        typer.Option(
+            "-o",
+            "--output",
+            help=(
+                "Output file path. If not specified, "
+                "will be the same as input with a .pdf extension."
+            ),
+        ),
+    ] = None,
 ):
     docx2pdf.convert(input_path=input_fpath, output_path=output_fpath)

{calkit_python-0.3.0 → calkit_python-0.3.1}/calkit/git.py

@@ -8,7 +8,10 @@ import git
 def detect_project_name(path=None) -> str:
     """Read the project owner and name from the remote.
 
-    TODO: Currently only works with GitHub remotes.
+    TODO: Currently only works with GitHub remotes where the GitHub repo
+    name is identical to the Calkit project name, which is not guaranteed.
+    We should probably look inside ``calkit.yaml`` at ``name``
+    first, and fallback to the GitHub remote URL if we can't find that.
     """
     url = git.Repo(path=path).remote().url
     return url.split("github.com")[-1][1:].removesuffix(".git")

{calkit_python-0.3.0 → calkit_python-0.3.1}/calkit/models.py

@@ -107,6 +107,11 @@ class ProjectInfo(BaseModel):
         distinguish what has been newly created here.
     """
 
+    title: str | None = None
+    owner: str | None = None
+    description: str | None = None
+    name: str | None = None
+    git_repo_url: str | None = None
     parent: str | None = None
     questions: list[str] = []
     datasets: list[Dataset] = []

{calkit_python-0.3.0 → calkit_python-0.3.1}/calkit/office.py

@@ -23,6 +23,7 @@ def excel_chart_to_png(
     wb = excel.Workbooks.Open(os.path.abspath(input_fpath))
     factor = 1.0
     # Extract sheet
+    # TODO: Close workbook if something fails
     sheet = excel.Sheets(sheet)
     shape = sheet.Shapes[chart_index]
     shape.Copy()
@@ -31,6 +32,9 @@ def excel_chart_to_png(
     size = int(factor * length_x), int(factor * width_y)
     image_resize = image.resize(size)
     # Save the image into the existing png file, overwriting if exists
+    dirname = os.path.dirname(output_fpath)
+    if dirname and not os.path.isdir(dirname):
+        os.makedirs(dirname)
     image_resize.save(
         os.path.abspath(output_fpath), "png", quality=95, dpi=(300, 300)
     )

{calkit_python-0.3.0 → calkit_python-0.3.1}/calkit/server.py

@@ -59,7 +59,7 @@ class LocalProject(BaseModel):
 
 
 @app.get("/")
-def get_root() -> list[LocalProject]:
+def get_root(get_jupyter_servers: bool = True) -> list[LocalProject]:
     """Return information about the current running server.
 
     - The project owner
@@ -68,9 +68,15 @@ def get_root() -> list[LocalProject]:
     - A Jupyter server running here, if applicable
     """
     resp = []
+    logger.info("Finding project directories")
     project_dirs = calkit.find_project_dirs()
-    servers = calkit.jupyter.get_servers()
+    if get_jupyter_servers:
+        logger.info("Getting Jupyter servers")
+        servers = calkit.jupyter.get_servers()
+    else:
+        servers = []
     for pdir in project_dirs:
+        logger.info(f"Inspecting {pdir}")
         try:
             project = calkit.git.detect_project_name(path=pdir)
         except ValueError:
@@ -96,8 +102,10 @@ def get_root() -> list[LocalProject]:
 
 
 @app.get("/projects/{owner_name}/{project_name}")
-def get_local_project(owner_name: str, project_name: str) -> LocalProject:
-    all_projects = get_root()
+def get_local_project(
+    owner_name: str, project_name: str, get_jupyter_server: bool = True
+) -> LocalProject:
+    all_projects = get_root(get_jupyter_servers=get_jupyter_server)
     for project in all_projects:
         if (
             project.owner_name == owner_name
@@ -235,14 +243,22 @@ class Status(BaseModel):
 
 
 @app.get("/projects/{owner_name}/{project_name}/status")
-def get_status(owner_name: str, project_name: str):
+def get_status(
+    owner_name: str,
+    project_name: str,
+    fetch_git: bool = True,
+    fetch_dvc: bool = True,
+):
     """Get status in working directory, from both Git and DVC."""
     errors = []
     logger.info(f"Looking for project {owner_name}/{project_name}")
-    project = get_local_project(owner_name, project_name)
+    project = get_local_project(
+        owner_name, project_name, get_jupyter_server=False
+    )
     logger.info(f"Found project at {project.wdir}")
     git_repo = git.Repo(project.wdir)
-    git_repo.git.fetch()
+    if fetch_git:
+        git_repo.git.fetch()
     untracked_git_files = git_repo.untracked_files
     # Get a list of diffs of the working tree to the index
     git_diff = git_repo.index.diff(None)
@@ -281,13 +297,17 @@ def get_status(owner_name: str, project_name: str):
         dvc_pipeline_status = dvc.repo.status.status(dvc_repo)
         # Remove any always changed entries so the pipeline doesn't look
         # out of date
+        logger.info(f"Raw DVC pipeline status: {dvc_pipeline_status}")
         dvc_pipeline_status = {
-            k: v
+            k.split("dvc.yaml:")[-1]: v
             for k, v in dvc_pipeline_status.items()
-            if v != ["always changed"] and "dvc.yaml:" in k
+            if v != ["always changed"] and not k.endswith(".dvc")
         }
+        logger.info(
+            f"DVC pipeline status after filtering: {dvc_pipeline_status}"
+        )
         dvc_data_status = dvc.repo.data.status(
-            dvc_repo, not_in_remote=True, remote_refresh=True
+            dvc_repo, not_in_remote=fetch_dvc, remote_refresh=fetch_dvc
         )
         # Reformat this a bit, since it can be a little hard to understand
         # DVC calls a path committed when its DVC file is staged