calkit-python 0.17.2__tar.gz → 0.17.3__tar.gz

{calkit_python-0.17.2 → calkit_python-0.17.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: calkit-python
-Version: 0.17.2
+Version: 0.17.3
 Summary: Reproducibility simplified.
 Project-URL: Homepage, https://calkit.org
 Project-URL: Issues, https://github.com/calkit/calkit/issues

{calkit_python-0.17.2 → calkit_python-0.17.3}/calkit/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.17.2"
+__version__ = "0.17.3"
 
 from .core import *
 from . import git

{calkit_python-0.17.2 → calkit_python-0.17.3}/calkit/cli/main.py

@@ -51,9 +51,31 @@ app.add_typer(update_app, name="update", help="Update objects.")
 app.add_typer(check_app, name="check", help="Check things.")
 
 # Constants for version control auto-ignore
-AUTO_IGNORE_SUFFIXES = [".DS_Store", ".env"]
+AUTO_IGNORE_SUFFIXES = [".DS_Store", ".env", ".pyc"]
 AUTO_IGNORE_PATHS = [os.path.join(".dvc", "config.local")]
-AUTO_IGNORE_PREFIXES = [".venv"]
+AUTO_IGNORE_PREFIXES = [".venv", "__pycache__"]
+# Constants for version control auto-add to DVC
+DVC_EXTENSIONS = [
+    ".png",
+    ".jpeg",
+    ".jpg",
+    ".gif",
+    ".h5",
+    ".parquet",
+    ".pickle",
+    ".mp4",
+    ".avi",
+    ".webm",
+    ".pdf",
+    ".xlsx",
+    ".docx",
+    ".xls",
+    ".doc",
+    ".nc",
+    ".nc4",
+    ".zarr",
+]
+DVC_SIZE_THRESH_BYTES = 1_000_000
 
 
 def _to_shell_cmd(cmd: list[str]) -> str:
@@ -195,6 +217,14 @@ def add(
             help="Automatically commit and use this as a message.",
         ),
     ] = None,
+    auto_commit_message: Annotated[
+        bool,
+        typer.Option(
+            "--auto-message",
+            "-M",
+            help="Commit with an automatically-generated message.",
+        ),
+    ] = False,
     disable_auto_ignore: Annotated[
         bool, typer.Option("--no-auto-ignore", help="Disable auto-ignore.")
     ] = False,
@@ -215,6 +245,18 @@ def add(
     Note: This will enable the 'autostage' feature of DVC, automatically
     adding any .dvc files to Git when adding to DVC.
     """
+    if auto_commit_message:
+        if commit_message is not None:
+            raise_error(
+                "Commit message should not be provided if using "
+                "automatic message"
+            )
+        if "." in paths:
+            raise_error("Cannot auto-generate commit message for '.'")
+        if len(paths) > 1:
+            raise_error(
+                "Cannot auto-generate commit message for more than one path"
+            )
     if to is not None and to not in ["git", "dvc"]:
         raise_error(f"Invalid option for 'to': {to}")
     try:
@@ -239,40 +281,25 @@ def add(
     # Ensure autostage is enabled for DVC
     subprocess.call(["dvc", "config", "core.autostage", "true"])
     subprocess.call(["git", "add", ".dvc/config"])
+    dvc_paths = [obj.get("path") for obj in dvc_repo.ls(".", dvc_only=True)]
+    untracked_git_files = repo.untracked_files
+    if auto_commit_message:
+        # See if this path is in the repo already
+        if paths[0] in dvc_paths or repo.git.ls_files(paths[0]):
+            commit_message = f"Update {paths[0]}"
+        else:
+            commit_message = f"Add {paths[0]}"
     if to is not None:
         subprocess.call([to, "add"] + paths)
     else:
-        dvc_extensions = [
-            ".png",
-            ".jpeg",
-            ".jpg",
-            ".gif",
-            ".h5",
-            ".parquet",
-            ".pickle",
-            ".mp4",
-            ".avi",
-            ".webm",
-            ".pdf",
-            ".xlsx",
-            ".docx",
-            ".xls",
-            ".doc",
-        ]
-        dvc_size_thresh_bytes = 1_000_000
-        dvc_paths = [
-            obj.get("path") for obj in dvc_repo.ls(".", dvc_only=True)
-        ]
         if "." in paths:
             paths.remove(".")
-            # TODO: There is some copy/paste from the `save` function here
             dvc_status = dvc_repo.data_status()
             for dvc_uncommitted in dvc_status["uncommitted"].get(
                 "modified", []
             ):
                 typer.echo(f"Adding {dvc_uncommitted} to DVC")
                 dvc_repo.commit(dvc_uncommitted, force=True)
-            untracked_git_files = repo.untracked_files
             if not disable_auto_ignore:
                 for untracked_file in untracked_git_files:
                     if (
@@ -315,10 +342,10 @@ def add(
                     f"Adding {path} to DVC since it's already tracked with DVC"
                 )
                 subprocess.call(["dvc", "add", path])
-            elif os.path.splitext(path)[-1] in dvc_extensions:
+            elif os.path.splitext(path)[-1] in DVC_EXTENSIONS:
                 typer.echo(f"Adding {path} to DVC per its extension")
                 subprocess.call(["dvc", "add", path])
-            elif calkit.get_size(path) > dvc_size_thresh_bytes:
+            elif calkit.get_size(path) > DVC_SIZE_THRESH_BYTES:
                 typer.echo(
                     f"Adding {path} to DVC since it's greater than 1 MB"
                 )

{calkit_python-0.17.2 → calkit_python-0.17.3}/calkit/cli/new.py

@@ -3,7 +3,6 @@
 from __future__ import annotations
 
 import os
-import re
 import shutil
 import subprocess
 
@@ -109,7 +108,7 @@ def new_project(
     if os.path.isdir(abs_path) and os.listdir(abs_path):
         warn(f"{abs_path} is not empty")
     if name is None:
-        name = re.sub(r"[-_,\.\ ]", "-", os.path.basename(abs_path).lower())
+        name = calkit.to_kebab_case(os.path.basename(abs_path))
     if " " in name:
         warn("Invalid name; replacing spaces with hyphens")
         name = name.replace(" ", "-")
@@ -242,8 +241,10 @@ def new_project(
     ck_info = dict(name=name, title=title, description=description) | ck_info
     with open(os.path.join(abs_path, "calkit.yaml"), "w") as f:
         ryaml.dump(ck_info, f)
+    repo.git.add("calkit.yaml")
     # Create dev container spec
     update_devcontainer(wdir=abs_path)
+    repo.git.add(".devcontainer")
     # Create README
     readme_fpath = os.path.join(abs_path, "README.md")
     if os.path.isfile(readme_fpath) and not overwrite:
@@ -257,6 +258,7 @@ def new_project(
         )
         with open(readme_fpath, "w") as f:
             f.write(readme_txt)
+        repo.git.add("README.md")
     if git_repo_url and not repo.remotes:
         typer.echo(f"Adding Git remote {git_repo_url}")
         repo.git.remote(["add", "origin", git_repo_url])
@@ -267,7 +269,6 @@ def new_project(
         typer.echo("Setting up Calkit Cloud DVC remote")
         calkit.dvc.configure_remote(wdir=abs_path)
         calkit.dvc.set_remote_auth(wdir=abs_path)
-    repo.git.add(".")
     if repo.git.diff("--staged") and not no_commit:
         repo.git.commit(["-m", "Initialize Calkit project"])
     typer.echo(success_message)
@@ -916,16 +917,24 @@ def new_conda_env(
     name: Annotated[
         str, typer.Option("--name", "-n", help="Environment name.")
     ],
+    conda_name: Annotated[
+        str,
+        typer.Option(
+            "--conda-name",
+            help=(
+                "Name to use in the Conda environment file, if desired. "
+                "Will be automatically generated if not provided. "
+                "Note that these should be unique since Conda environments are "
+                "a system-wide collection."
+            ),
+        ),
+    ] = None,
     path: Annotated[
         str, typer.Option("--path", help="Environment YAML file path.")
     ] = "environment.yml",
     pip_packages: Annotated[
         list[str], typer.Option("--pip", help="Packages to install with pip.")
     ] = [],
-    stage: Annotated[
-        str,
-        typer.Option("--stage", help="DVC pipeline stage name for checking."),
-    ] = None,
     description: Annotated[
         str, typer.Option("--description", help="Description.")
     ] = None,
@@ -957,9 +966,14 @@ def new_conda_env(
             f"Environment with name {name} already exists "
             "(use -f to overwrite)"
         )
+    if conda_name is None:
+        project_name = ck_info.get("name")
+        if project_name is None:
+            project_name = os.path.basename(os.getcwd())
+        conda_name = calkit.to_kebab_case(project_name) + "-" + name
     # Write environment to path
     conda_env = dict(
-        name=name, channels=["conda-forge"], dependencies=packages
+        name=conda_name, channels=["conda-forge"], dependencies=packages
     )
     if pip_packages:
         conda_env["dependencies"].append(dict(pip=pip_packages))
@@ -968,42 +982,13 @@ def new_conda_env(
     repo.git.add(path)
     typer.echo("Adding environment to calkit.yaml")
     env = dict(path=path, kind="conda")
-    if stage is not None:
-        env["stage"] = stage
     if description is not None:
         env["description"] = description
     envs[name] = env
     ck_info["environments"] = envs
     with open("calkit.yaml", "w") as f:
         ryaml.dump(ck_info, f)
-    # If we're creating a stage, do so with DVC
-    if stage:
-        typer.echo(f"Creating DVC stage {stage}")
-        if not os.path.isfile(".dvc/config"):
-            typer.echo("Running dvc init")
-            subprocess.check_call(["dvc", "init"])
-        ck_cmd = f"calkit check-conda-env -f {path}"
-        fname, ext = os.path.splitext(path)
-        lock_path = f"{fname}-lock{ext}"
-        subprocess.check_call(
-            [
-                "dvc",
-                "stage",
-                "add",
-                "-f",
-                "-n",
-                stage,
-                "--always-changed",
-                "-d",
-                path,
-                "--outs-persist-no-cache",
-                lock_path,
-                ck_cmd,
-            ]
-        )
     repo.git.add("calkit.yaml")
-    if stage:
-        repo.git.add("dvc.yaml")
     if not no_commit and repo.git.diff("--staged"):
         repo.git.commit(["-m", f"Add Conda environment {name}"])
 

{calkit_python-0.17.2 → calkit_python-0.17.3}/calkit/core.py

@@ -380,3 +380,8 @@ def get_size(path: str):
             if not os.path.islink(fp):
                 total_size += os.path.getsize(fp)
     return total_size
+
+
+def to_kebab_case(str) -> str:
+    """Convert a string to kebab-case."""
+    return re.sub(r"[-_,\.\ ]", "-", str.lower())

{calkit_python-0.17.2 → calkit_python-0.17.3}/calkit/tests/cli/test_main.py

@@ -276,3 +276,23 @@ def test_add(tmp_dir):
     assert "data" in calkit.dvc.list_paths()
     # Check that we can run `calkit add .`
     subprocess.check_call(["calkit", "add", "."])
+    # Test the auto commit message feature
+    subprocess.check_call(["git", "reset"])
+    subprocess.check_call(["calkit", "add", "large.bin", "-M"])
+    repo = git.Repo()
+    assert repo.head.commit.message.strip() == "Update large.bin"
+    subprocess.check_call(["calkit", "add", "src", "-M"])
+    assert repo.head.commit.message.strip() == "Add src"
+    with open("src/code.py", "w") as f:
+        f.write("# This is the new code")
+    subprocess.check_call(["calkit", "add", "src/code.py", "-M"])
+    assert repo.head.commit.message.strip() == "Update src/code.py"
+    with open("data/raw/file2.bin", "wb") as f:
+        f.write(os.urandom(2_000_000))
+    subprocess.check_call(["calkit", "add", "data", "-M"])
+    assert repo.head.commit.message.strip() == "Update data"
+    os.makedirs("data2")
+    with open("data2/large2.bin", "wb") as f:
+        f.write(os.urandom(2_000_000))
+    subprocess.check_call(["calkit", "add", "data2", "-M"])
+    assert repo.head.commit.message.strip() == "Add data2"

{calkit_python-0.17.2 → calkit_python-0.17.3}/calkit/tests/cli/test_new.py

@@ -3,6 +3,7 @@
 import os
 import subprocess
 
+import git
 import pytest
 
 import calkit
@@ -212,3 +213,18 @@ def test_new_uv_venv(tmp_dir):
     assert env.path == "requirements-2.txt"
     assert env.prefix == ".venv2"
     assert env.kind == "uv-venv"
+
+
+def test_new_project(tmp_dir):
+    subprocess.check_call(["touch", "some-existing-file.txt"])
+    subprocess.check_call(
+        ["calkit", "new", "project", ".", "--title", "My new project"]
+    )
+    repo = git.Repo()
+    assert "some-existing-file.txt" in repo.untracked_files
+    assert repo.git.ls_files("calkit.yaml")
+    assert repo.git.ls_files("README.md")
+    assert not repo.git.ls_files("some-other-file.txt")
+    assert repo.git.ls_files(".devcontainer")
+    ck_info = calkit.load_calkit_info()
+    assert ck_info["title"] == "My new project"

{calkit_python-0.17.2 → calkit_python-0.17.3}/calkit/tests/test_conda.py

@@ -1,12 +1,15 @@
 """Tests for the ``conda`` module."""
 
+import os
 import subprocess
-import uuid
 
 import pytest
 
+import calkit
 from calkit.conda import _check_list, _check_single, check_env
 
+ENV_NAME = "main"
+
 
 def test_check_single():
     assert _check_single("python=3.12", "python=3.12.18", conda=True)
@@ -31,15 +34,14 @@ def delete_env(name: str):
 
 
 @pytest.fixture
-def env_name():
-    # Setup code
-    name = "tmp_" + str(uuid.uuid4())[:12]
+def conda_env_name():
+    name = calkit.to_kebab_case(os.path.basename(os.getcwd())) + "-" + ENV_NAME
     yield name
     # Teardown code
     delete_env(name)
 
 
-def test_check_env(tmp_dir, env_name):
+def test_check_env(tmp_dir, conda_env_name):
     subprocess.check_call(["git", "init"])
     subprocess.check_call(["dvc", "init"])
     subprocess.check_call(
@@ -48,7 +50,7 @@ def test_check_env(tmp_dir, env_name):
             "new",
             "conda-env",
             "-n",
-            env_name,
+            ENV_NAME,
             "python",
             "pip",
             "--pip",
@@ -69,7 +71,7 @@ def test_check_env(tmp_dir, env_name):
             "conda-env",
             "--overwrite",
             "-n",
-            env_name,
+            ENV_NAME,
             "python=3.11.0",
             "pip",
             "--pip",
@@ -91,7 +93,7 @@ def test_check_env(tmp_dir, env_name):
             "conda-env",
             "--overwrite",
             "-n",
-            env_name,
+            ENV_NAME,
             "python=3.11.0",
             "pip",
             "sqlalchemy",
@@ -102,7 +104,7 @@ def test_check_env(tmp_dir, env_name):
             "conda",
             "run",
             "-n",
-            env_name,
+            ENV_NAME,
             "pip",
             "install",
             "--upgrade",
@@ -118,7 +120,7 @@ def test_check_env(tmp_dir, env_name):
             "conda-env",
             "--overwrite",
             "-n",
-            env_name,
+            ENV_NAME,
             "python=3.11.0",
             "pip",
             "sqlalchemy",
@@ -133,7 +135,7 @@ def test_check_env(tmp_dir, env_name):
             "conda-env",
             "--overwrite",
             "-n",
-            env_name,
+            ENV_NAME,
             "python=3.11.0",
             "pip",
             "--pip",
@@ -150,7 +152,7 @@ def test_check_env(tmp_dir, env_name):
             "conda-env",
             "--overwrite",
             "-n",
-            env_name,
+            ENV_NAME,
             "python=3.12",
             "--pip",
             "numpy>=1",

{calkit_python-0.17.2 → calkit_python-0.17.3}/calkit/tests/test_core.py

@@ -6,3 +6,8 @@ import calkit
 def test_find_project_dirs():
     calkit.find_project_dirs()
     assert calkit.find_project_dirs(relative=False)
+
+
+def test_to_kebab_case():
+    assert calkit.to_kebab_case("THIS IS") == "this-is"
+    assert calkit.to_kebab_case("this_is_my-Project") == "this-is-my-project"

{calkit_python-0.17.2 → calkit_python-0.17.3}/docs/calkit-yaml.md

@@ -19,3 +19,20 @@ for the project's important metadata, which includes its:
 Objects can be imported from other projects,
 which produces a chain of reference to allow tracking reuse
 and reduce redundant storage.
+
+## Showcase
+
+The project showcase is a list of elements that best represent the project,
+shown on the project homepage on the Calkit Cloud web app.
+For example:
+
+```yaml
+showcase:
+  - text: Here is some text.
+  - figure: figures/my-figure.png
+  - text: There is a figure above.
+```
+
+[This project](https://calkit.io/petebachant/strava-analysis)
+has a showcase that includes Plotly figures saved as JSON,
+which render interactively.

calkit_python-0.17.3/docs/local-server.md

@@ -0,0 +1,23 @@
+# Local server for GUI interaction
+
+Calkit includes a local server for interacting with projects locally via the
+Calkit Cloud UI ([calkit.io](https://calkit.io)).
+It can be launched with:
+
+```sh
+calkit local-server
+```
+
+If you then navigate to the project homepage on
+[calkit.io](https://calkit.io) and visit the "local machine"
+tab,
+you'll be able to visually perform tasks like committing and ignoring files,
+running the pipeline, adding stages to the pipeline,
+and pushing to the cloud.
+
+This feature is used in the [Microsoft Office tutorial](tutorials/office.md).
+
+![Local server page](tutorials/img/office/pipeline-out-of-date.png){ width="450px" }
+/// caption
+The local server status widget on the Calkit Cloud UI.
+///

{calkit_python-0.17.2 → calkit_python-0.17.3}/docs/version-control.md

@@ -170,4 +170,6 @@ Options:
 - `--to`, `-t`: Manually specify `git` or `dvc` as the tracking mechanism.
 - `--commit-message`, `-m`: Create a commit after adding
   and use the provided message.
+- `--auto-commit-message`, `-M`: Commit with an automatically-generated message.
+  Only compatible when adding one path.
 - `--push`: Push to the Git or DVC remote after committing.

{calkit_python-0.17.2 → calkit_python-0.17.3}/mkdocs.yml

@@ -47,6 +47,7 @@ nav:
   - References: references.md
   - Calculations: calculations.md
   - Apps: apps.md
+  - Local server (GUI interaction): local-server.md
   - Examples: examples.md
   - Tutorials:
       - tutorials/first-project.md