calkit-python 0.34.2__tar.gz → 0.34.4__tar.gz

{calkit_python-0.34.2 → calkit_python-0.34.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: calkit-python
-Version: 0.34.2
+Version: 0.34.4
 Summary: Reproducibility simplified.
 Project-URL: Homepage, https://calkit.org
 Project-URL: Issues, https://github.com/calkit/calkit/issues
@@ -44,6 +44,7 @@ Requires-Dist: python-dotenv>=1
 Requires-Dist: pywin32; platform_system == 'Windows'
 Requires-Dist: requests
 Requires-Dist: sqlitedict>=2.1.0
+Requires-Dist: texsoup>=0.3.1
 Requires-Dist: toml>=0.10.2
 Requires-Dist: tqdm>=4.67.1
 Requires-Dist: typer==0.21.1

{calkit_python-0.34.2 → calkit_python-0.34.4}/calkit/cli/main.py

@@ -1189,6 +1189,7 @@ def run(
         "allow-missing",
         "dry",
         "keep-going",
+        "ignore-errors",
         "force-downstream",
         "glob",
         "no-commit",
@@ -1199,6 +1200,7 @@ def run(
     if pipeline is not None:
         args += ["--pipeline", pipeline]
     if downstream is not None:
+        args.append("--downstream")
         args += downstream
     start_time_no_tz = calkit.utcnow(remove_tz=True)
     start_time = calkit.utcnow(remove_tz=False)

{calkit_python-0.34.2 → calkit_python-0.34.4}/calkit/cli/overleaf.py

@@ -18,6 +18,19 @@ from calkit.cli import raise_error, warn
 overleaf_app = typer.Typer(no_args_is_help=True)
 
 
+def _extract_title_from_tex(tex_file_path: str) -> str | None:
+    """Extract the title from a LaTeX file."""
+    from TexSoup import TexSoup
+
+    try:
+        with open(tex_file_path) as f:
+            overleaf_target_text = f.read()
+        texsoup = TexSoup(overleaf_target_text)
+        return str(texsoup.title.string) if texsoup.title else None
+    except Exception:
+        return None
+
+
 def _get_overleaf_token() -> str:
     """Get the user's Overleaf token from config.
 
@@ -78,13 +91,13 @@ def import_publication(
         ),
     ],
     title: Annotated[
-        str,
+        str | None,
         typer.Option(
             "--title",
             "-t",
             help="Title of the publication.",
         ),
-    ],
+    ] = None,
     target_path: Annotated[
         str | None,
         typer.Option(
@@ -143,6 +156,18 @@ def import_publication(
             help="Force adding the publication even if it already exists.",
         ),
     ] = False,
+    push_only: Annotated[
+        bool,
+        typer.Option(
+            "--push-only",
+            "-P",
+            help=(
+                "Push local files to Overleaf without pulling. "
+                "Useful when initializing a new Overleaf project from local "
+                "files."
+            ),
+        ),
+    ] = False,
 ):
     """Import a publication from an Overleaf project."""
     from calkit.cli.main import ignore as git_ignore
@@ -188,7 +213,9 @@ def import_publication(
     git.Repo.clone_from(git_clone_url, overleaf_project_dir)
     # Detect target path if not specified
     if target_path is None:
-        ol_contents = os.listdir(overleaf_project_dir)
+        ol_contents = os.listdir(
+            dest_dir if push_only else overleaf_project_dir
+        )
         for cand in ["main.tex", "report.tex", "paper.tex"]:
             if cand in ol_contents:
                 target_path = cand
@@ -203,10 +230,26 @@ def import_publication(
             "Target TeX file path cannot be detected; "
             "please specify with --target"
         )
+        return
+    # Try to extract title from the target LaTeX file if not provided
+    if not title:
+        target_tex_path = os.path.join(
+            dest_dir if push_only else overleaf_project_dir, target_path
+        )
+        extracted_title = _extract_title_from_tex(target_tex_path)
+        if extracted_title:
+            typer.echo(f"Detected title: {extracted_title}")
+            title = extracted_title
+    if not title:
+        raise_error(
+            "Title could not be detected from the LaTeX file; "
+            "please specify with --title"
+        )
     # Determine the PDF output path
     pdf_path = target_path.removesuffix(".tex") + ".pdf"  # type: ignore
     typer.echo(f"Using PDF path: {pdf_path}")
     tex_path = pdf_path.removesuffix(".pdf") + ".tex"
+    target_tex_path = Path(dest_dir, tex_path).as_posix()
     pub_path = Path(dest_dir, pdf_path).as_posix()
     pub_paths = [pub.get("path") for pub in pubs]
     if not overwrite and pub_path in pub_paths:
@@ -240,20 +283,18 @@ def import_publication(
         with open("calkit.yaml", "w") as f:
             calkit.ryaml.dump(ck_info, f)
     # Check that we have a build stage
-    # TODO: Use Calkit pipeline for this
     typer.echo("Checking for a build stage in the pipeline")
+    pipeline = ck_info.get("pipeline", {})
+    stages = pipeline.get("stages", {})
     stage_name = None
-    if os.path.isfile("dvc.yaml"):
-        with open("dvc.yaml", "r") as f:
-            dvc_info = calkit.ryaml.load(f)
-        stages = dvc_info.get("stages", {})
-        for stage_name_i, stage in stages.items():
-            if pub_path in stage.get("outs", []):
-                stage_name = stage_name_i
-                typer.echo(f"Found build stage '{stage_name}' in pipeline")
-                break
-    else:
-        stages = {}
+    for stage_name_i, stage in stages.items():
+        if (
+            stage.get("kind") == "latex"
+            and stage.get("target_path") == target_tex_path
+        ):
+            stage_name = stage_name_i
+            typer.echo(f"Found build stage '{stage_name}' in pipeline")
+            break
     if stage_name is None:
         # Create a new stage
         stage_name = calkit.to_kebab_case("build-" + dest_dir)
@@ -265,7 +306,7 @@ def import_publication(
         new_latex_stage(
             name=stage_name,
             environment=tex_env_name,
-            target_path=Path(dest_dir, tex_path).as_posix(),
+            target_path=target_tex_path,
             inputs=[
                 os.path.join(dest_dir, p) for p in sync_paths + push_paths
             ],
@@ -308,7 +349,7 @@ def import_publication(
             ]
         )
     # Sync the project
-    sync(paths=[dest_dir], no_commit=no_commit)
+    sync(paths=[dest_dir], no_commit=no_commit, push_only=push_only)
 
 
 @overleaf_app.command(name="sync")
@@ -347,7 +388,7 @@ def sync(
         typer.Option(
             "--no-push",
             help=(
-                "Do not push the changes to the project remote. "
+                "Do not push the changes to the main project remote. "
                 "Changes will always be pushed to Overleaf."
             ),
         ),
@@ -367,6 +408,19 @@ def sync(
             help="Mark merge conflicts as resolved before committing.",
         ),
     ] = False,
+    push_only: Annotated[
+        bool,
+        typer.Option(
+            "--push-only",
+            "-P",
+            help=(
+                "Only push local files to Overleaf without pulling from "
+                "Overleaf. "
+                "Useful when initializing a new Overleaf project from local "
+                "files."
+            ),
+        ),
+    ] = False,
 ):
     """Sync folders with Overleaf."""
     # TODO: We should probably ensure the pipeline isn't stale
@@ -487,6 +541,7 @@ def sync(
                 verbose=verbose,
                 resolving_conflict=resolve,
                 print_info=typer.echo,
+                push_only=push_only,
             )
         except Exception as e:
             raise_error(str(e))

{calkit_python-0.34.2 → calkit_python-0.34.4}/calkit/overleaf.py

@@ -307,6 +307,7 @@ def sync(
     print_info=print,
     verbose: bool = False,
     resolving_conflict: bool = False,
+    push_only: bool = False,
 ) -> dict:
     """Sync between the main project repo and Overleaf repo.
 
@@ -315,6 +316,10 @@ def sync(
     All files in the Overleaf repo are committed to Git, while some in the
     main project can be ignored, e.g., in cases where they are copied in from
     a map-paths stage.
+
+    When push_only is True, only push local files to Overleaf without pulling
+    or applying changes from Overleaf to local. Useful for initializing a new
+    Overleaf project from local files.
     """
     res = {}
     # Normalize ``path_in_project`` as a posix path
@@ -344,7 +349,13 @@ def sync(
     )
     paths_for_overleaf_patch = paths.paths_to_use_for_git_patch
     res["paths_for_overleaf_patch"] = paths_for_overleaf_patch
-    if last_sync_commit:
+    if push_only:
+        # When push_only is True, skip pulling from Overleaf and applying
+        # patches to local
+        # Simply copy files to Overleaf
+        print_info("Push-only sync; skipping pull from Overleaf")
+        res["patch"] = None
+    elif last_sync_commit:
         # Compute a patch in the Overleaf project between HEAD and the last
         # sync
         patch = overleaf_repo.git.format_patch(

{calkit_python-0.34.2 → calkit_python-0.34.4}/calkit/tests/cli/test_main.py

@@ -13,6 +13,7 @@ import dvc.repo
 import git
 import pytest
 import toml
+import yaml
 from dvc.exceptions import NotDvcRepoError
 from git.exc import InvalidGitRepositoryError
 
@@ -645,6 +646,70 @@ def test_run(tmp_dir):
     assert "stage_run_info" in res
 
 
+def test_run_ignore_errors(tmp_dir):
+    subprocess.check_call(["calkit", "init"])
+    # Create a pipeline with a failing stage and an independent stage
+    dvc_yaml = {
+        "stages": {
+            "failing-stage": {
+                "cmd": 'python -c "import sys; sys.exit(1)"',
+            },
+            "independent-stage": {
+                "cmd": "python -c \"open('out.txt', 'w').write('done')\"",
+                "outs": ["out.txt"],
+            },
+        }
+    }
+    with open("dvc.yaml", "w") as f:
+        yaml.dump(dvc_yaml, f)
+    subprocess.check_call(
+        ["calkit", "save", "-am", "Create pipeline", "--no-push"]
+    )
+    # Without --ignore-errors, the pipeline should fail
+    result = subprocess.run(["calkit", "run"])
+    assert result.returncode != 0
+    # With --ignore-errors, the independent stage should complete
+    if os.path.exists("out.txt"):
+        os.remove("out.txt")
+    subprocess.check_call(["calkit", "run", "--ignore-errors"])
+    assert os.path.exists("out.txt")
+
+
+def test_run_downstream(tmp_dir):
+    subprocess.check_call(["calkit", "init"])
+    # Create a pipeline: stage-a -> stage-b
+    dvc_yaml = {
+        "stages": {
+            "stage-a": {
+                "cmd": "python -c \"open('a.txt', 'w').write('a')\"",
+                "outs": ["a.txt"],
+            },
+            "stage-b": {
+                "cmd": (
+                    "python -c \"open('b.txt', 'w')"
+                    ".write(open('a.txt').read() + 'b')\""
+                ),
+                "deps": ["a.txt"],
+                "outs": ["b.txt"],
+            },
+        }
+    }
+    with open("dvc.yaml", "w") as f:
+        yaml.dump(dvc_yaml, f)
+    # Run all stages to prime the cache
+    subprocess.check_call(["calkit", "run"])
+    assert os.path.exists("a.txt")
+    assert os.path.exists("b.txt")
+    # Delete outputs and run with --downstream stage-a: both stages should run
+    os.remove("a.txt")
+    os.remove("b.txt")
+    subprocess.check_call(
+        ["calkit", "run", "--downstream", "stage-a", "--force"]
+    )
+    assert os.path.exists("a.txt")
+    assert os.path.exists("b.txt")
+
+
 def test_stage_run_info_from_log_content():
     fpath = os.path.join(
         os.path.dirname(__file__), "..", "..", "..", "test", "test-log.log"

{calkit_python-0.34.2 → calkit_python-0.34.4}/calkit/tests/cli/test_overleaf.py

@@ -7,6 +7,7 @@ import uuid
 import git
 
 import calkit
+from calkit.cli.overleaf import _extract_title_from_tex
 from calkit.git import ls_files
 
 
@@ -141,3 +142,18 @@ def test_overleaf(tmp_dir):
     assert "main.pdf" not in ol_repo_git_show
     assert "main.log" not in ol_repo_git_show
     assert "main.aux" not in ol_repo_git_show
+
+
+def test_extract_title_from_tex(tmp_dir):
+    # Test that we can extract a title from a simple LaTeX file
+    tex = r"""
+    \documentclass{article}
+    \title{My Cool Paper}
+    \begin{document}
+    Hello world!
+    \end{document}
+    """
+    with open("test.tex", "w") as f:
+        f.write(tex)
+    title = _extract_title_from_tex("test.tex")
+    assert title == "My Cool Paper"

{calkit_python-0.34.2 → calkit_python-0.34.4}/docs/overleaf.md

@@ -51,10 +51,7 @@ For example:
 ```sh
 calkit overleaf import \
     https://www.overleaf.com/project/68000059d42b134573cb2e35 \
-    paper \
-    --title "My paper title" \
-    --kind journal-article \
-    --push-path figures
+    paper
 ```
 
 This command will link a local project folder, in this case `paper`,

{calkit_python-0.34.2 → calkit_python-0.34.4}/pyproject.toml

@@ -52,6 +52,7 @@ dependencies = [
   "sqlitedict>=2.1.0",
   "json2latex>=0.0.3",
   "toml>=0.10.2",
+  "texsoup>=0.3.1",
 ]
 description = "Reproducibility simplified."
 dynamic = ["version"]

{calkit_python-0.34.2 → calkit_python-0.34.4}/uv.lock

@@ -512,6 +512,7 @@ dependencies = [
     { name = "pywin32", marker = "sys_platform == 'win32'" },
     { name = "requests" },
     { name = "sqlitedict" },
+    { name = "texsoup" },
     { name = "toml" },
     { name = "tqdm" },
     { name = "typer" },
@@ -580,6 +581,7 @@ requires-dist = [
     { name = "pywin32", marker = "sys_platform == 'win32'" },
     { name = "requests" },
     { name = "sqlitedict", specifier = ">=2.1.0" },
+    { name = "texsoup", specifier = ">=0.3.1" },
     { name = "toml", specifier = ">=0.10.2" },
     { name = "tqdm", specifier = ">=4.67.1" },
     { name = "typer", specifier = "==0.21.1" },
@@ -5065,6 +5067,15 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/6a/9e/2064975477fdc887e47ad42157e214526dcad8f317a948dee17e1659a62f/terminado-0.18.1-py3-none-any.whl", hash = "sha256:a4468e1b37bb318f8a86514f65814e1afc977cf29b3992a4500d9dd305dcceb0", size = 14154, upload-time = "2024-03-12T14:34:36.569Z" },
 ]
 
+[[package]]
+name = "texsoup"
+version = "0.3.1"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/84/58/1c503390ed1a81cdcbff811dbf7a54132994acef8dd2194d55cf657a9e97/TexSoup-0.3.1.tar.gz", hash = "sha256:3f6b2ad0abe3688a6656f544c1ba04d0eb25f423f8c377b7369f9ce061ddb70b", size = 26174, upload-time = "2020-08-03T10:14:56.027Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/5c/a7/e9eb0e14633710b51b8472a3beccc7d6a44d55a0a3ef4493ebb4b7977253/TexSoup-0.3.1-py3-none-any.whl", hash = "sha256:ae8f08d17f86a905b7c2ce01c9f2da613fbca0bcea78c71d727719e896045bed", size = 27809, upload-time = "2024-02-28T09:15:28.697Z" },
+]
+
 [[package]]
 name = "tinycss2"
 version = "1.4.0"