steve-cli 0.3.8__tar.gz → 0.3.11__tar.gz

{steve_cli-0.3.8 → steve_cli-0.3.11}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: steve-cli
-Version: 0.3.8
+Version: 0.3.11
 Summary: A simple CLI tool to run jobs from jobs.yaml with proper environment setup
 Author: Frank
 License: MIT
@@ -44,6 +44,8 @@ Requires-Dist: great-expectations>=0.18.0; extra == "great-expectations"
 Requires-Dist: pandas>=2.0.3; extra == "great-expectations"
 Provides-Extra: excel
 Requires-Dist: openpyxl>=3.1.0; extra == "excel"
+Provides-Extra: visidata
+Requires-Dist: visidata>=3.0; extra == "visidata"
 Provides-Extra: all
 Requires-Dist: openlineage-python>=1.0.0; extra == "all"
 Requires-Dist: pyarrow>=17.0.0; extra == "all"
@@ -52,6 +54,7 @@ Requires-Dist: validoopsie>=0.1.0; extra == "all"
 Requires-Dist: great-expectations>=0.18.0; extra == "all"
 Requires-Dist: pandas>=2.0.3; extra == "all"
 Requires-Dist: openpyxl>=3.1.0; extra == "all"
+Requires-Dist: visidata>=3.0; extra == "all"
 
 # Steve CLI
 

{steve_cli-0.3.8 → steve_cli-0.3.11}/pyproject.toml

@@ -5,7 +5,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "steve-cli"
-version = "0.3.8"
+version = "0.3.11"
 description = "A simple CLI tool to run jobs from jobs.yaml with proper environment setup"
 readme = "README.md"
 license = {text = "MIT"}
@@ -60,6 +60,9 @@ great-expectations = [
 excel = [
     "openpyxl>=3.1.0",
 ]
+visidata = [
+    "visidata>=3.0",
+]
 all = [
     "openlineage-python>=1.0.0",
     "pyarrow>=17.0.0",
@@ -68,6 +71,7 @@ all = [
     "great-expectations>=0.18.0",
     "pandas>=2.0.3",
     "openpyxl>=3.1.0",
+    "visidata>=3.0",
 ]
 
 [project.scripts]

{steve_cli-0.3.8 → steve_cli-0.3.11}/steve_cli/cli.py

@@ -9,6 +9,7 @@ Usage:
 """
 
 import os
+import shutil
 import signal
 import socket
 import subprocess
@@ -18,7 +19,7 @@ from pathlib import Path
 from typing import Dict, List, Any, Optional
 
 import click
-from dotenv import load_dotenv
+from dotenv import load_dotenv, dotenv_values
 import questionary
 import yaml
 
@@ -69,6 +70,10 @@ def run_job_command(job: Dict[str, Any]) -> int:
         return 1
 
     env = os.environ.copy()
+    for env_filename in [".env", ".workspaces.env"]:
+        env_path = Path.cwd() / env_filename
+        if env_path.exists():
+            env.update({k: v for k, v in dotenv_values(env_path).items() if v is not None})
     job_env = job.get('env', {})
     for key, value in job_env.items():
         env[key] = str(value)
@@ -369,6 +374,10 @@ def apps_start(app_name: str, apps_file: Optional[Path]):
         sys.exit(1)
 
     env = os.environ.copy()
+    for env_filename in [".env", ".workspaces.env"]:
+        env_path = config.apps_file.parent / env_filename
+        if env_path.exists():
+            env.update({k: v for k, v in dotenv_values(env_path).items() if v is not None})
     for key, value in env_vars.items():
         env[key] = str(value)
 
@@ -537,7 +546,7 @@ def _print_tree(node: Dict[str, Any], prefix: str = "", is_last: bool = True) ->
             _print_tree(child, prefix + extension, last)
 
 
-def _list_bucket(storage_kwargs: dict, label: str, bucket_name: str) -> None:
+def _list_bucket(storage_kwargs: dict, label: str, bucket_name: str) -> tuple:
     click.echo(f"  {click.style(label, fg='cyan')} ({bucket_name})")
     try:
         from steve_cli.storage import S3Storage
@@ -548,10 +557,12 @@ def _list_bucket(storage_kwargs: dict, label: str, bucket_name: str) -> None:
         else:
             tree = _build_tree(keys)
             _print_tree(tree, prefix="    ")
+        return storage, keys
     except EnvironmentError as e:
         click.echo(f"    ⚠️  {e}", err=True)
     except Exception as e:
         click.echo(f"    ❌ {e}", err=True)
+    return None, []
 
 
 @main.command("buckets")
@@ -602,7 +613,33 @@ def buckets(env_file: tuple):
         sys.exit(0)
 
     selected = next(o for o in options if o["label"] == choice)
-    _list_bucket(selected["kwargs"], selected["tier"], selected["bucket_name"])
+    storage, keys = _list_bucket(selected["kwargs"], selected["tier"], selected["bucket_name"])
+
+    if not storage or not keys:
+        return
+
+    file_choice = questionary.select(
+        "View a file (or press Esc to exit):",
+        choices=["(done)"] + keys,
+    ).ask()
+
+    if not file_choice or file_choice == "(done)":
+        return
+
+    click.echo(f"\n📄 {click.style(file_choice, fg='cyan')}\n")
+    try:
+        import tempfile
+        data = storage.get_bytes(file_choice)
+        suffix = Path(file_choice).suffix
+        with tempfile.NamedTemporaryFile(suffix=suffix, delete=False) as tmp:
+            tmp.write(data)
+            tmp_path = tmp.name
+        if not shutil.which("vd"):
+            click.secho("visidata not found. Install it with: uv pip install 'steve-cli[visidata]'", fg="yellow")
+            return
+        subprocess.call(["vd", tmp_path])
+    except Exception as e:
+        click.secho(f"❌ Could not open file: {e}", fg="red", err=True)
 
 
 if __name__ == '__main__':

{steve_cli-0.3.8 → steve_cli-0.3.11}/steve_cli/decorators/lineage_job.py

@@ -26,6 +26,10 @@ def lineage_job(
 
         @functools.wraps(fn)
         def wrapper(*args: Any, **kwargs: Any) -> Any:
+            from dotenv import load_dotenv
+            load_dotenv(".env", override=False)
+            load_dotenv(".workspaces.env", override=False)
+
             session = make_session(
                 namespace=namespace,
                 job_name=job_name,
@@ -45,10 +49,11 @@ def lineage_job(
                 result = fn(*args, get_storage=get_storage, **kwargs)
             except Exception as exc:
                 session.fail(exc)
+                import sys
                 from steve_cli.validation.port import DataQualityError
-                if isinstance(exc, DataQualityError):
-                    import sys
-                    print(f"ERROR {exc}", file=sys.stderr)
+                if isinstance(exc, (DataQualityError, EnvironmentError)):
+                    import click
+                    click.secho(f"ERROR {exc}", fg="red", bold=True, err=True)
                     sys.exit(1)
                 raise
 

{steve_cli-0.3.8 → steve_cli-0.3.11}/steve_cli/lineage/collector.py

@@ -18,16 +18,21 @@ class LineageSession:
     _inputs: set[str] = field(default_factory=set)
     _outputs: set[str] = field(default_factory=set)
     _facets: Dict[str, Dict[str, Any]] = field(default_factory=dict)
+    _dataset_namespaces: Dict[str, str] = field(default_factory=dict)
 
-    def record_read(self, dataset: str, facets: Dict[str, Any] | None = None) -> None:
+    def record_read(self, dataset: str, facets: Dict[str, Any] | None = None, namespace: str | None = None) -> None:
         self._inputs.add(dataset)
         if facets:
             self._facets.setdefault(dataset, {}).update(facets)
+        if namespace:
+            self._dataset_namespaces[dataset] = namespace
 
-    def record_write(self, dataset: str, facets: Dict[str, Any] | None = None) -> None:
+    def record_write(self, dataset: str, facets: Dict[str, Any] | None = None, namespace: str | None = None) -> None:
         self._outputs.add(dataset)
         if facets:
             self._facets.setdefault(dataset, {}).update(facets)
+        if namespace:
+            self._dataset_namespaces[dataset] = namespace
 
     def attach_facets(self, dataset: str, facets: Dict[str, Any]) -> None:
         self._facets.setdefault(dataset, {}).update(facets)
@@ -42,8 +47,8 @@ class LineageSession:
             namespace=self.namespace,
             run_id=self.run_id,
             state=state,
-            inputs=[DatasetRef(namespace=self.namespace, name=x, facets=self._facets.get(x, {})) for x in self._inputs],
-            outputs=[DatasetRef(namespace=self.namespace, name=x, facets=self._facets.get(x, {})) for x in self._outputs],
+            inputs=[DatasetRef(namespace=self._dataset_namespaces.get(x, self.namespace), name=x, facets=self._facets.get(x, {})) for x in self._inputs],
+            outputs=[DatasetRef(namespace=self._dataset_namespaces.get(x, self.namespace), name=x, facets=self._facets.get(x, {})) for x in self._outputs],
             run_facets=run_facets or {},
         )
 
@@ -86,7 +91,11 @@ def make_session(
         kwargs = {"url": url} if prov == "openlineage" else {}
         try:
             port = LineageRegistry.create(prov, **kwargs)
-        except Exception:
+        except Exception as exc:
+            import logging
+            logging.getLogger(__name__).warning(
+                "Lineage disabled: could not initialize provider %r: %s", prov, exc
+            )
             from .adapters.null import NullLineageAdapter
 
             port = NullLineageAdapter()

{steve_cli-0.3.8 → steve_cli-0.3.11}/steve_cli/lineage/storage.py

@@ -58,23 +58,30 @@ class LineageStorage:
                 self.__resolved = self._storage_or_factory  # type: ignore[assignment]
         return self.__resolved
 
+    @property
+    def _dataset_namespace(self) -> str | None:
+        s = self._storage
+        if hasattr(s, 'endpoint') and hasattr(s, 'bucket'):
+            return f"{s.endpoint.rstrip('/')}/{s.bucket}"
+        return None
+
     def put_file(self, local_path: str, path: str) -> None:
         self._storage.put_file(local_path, path)
-        self._session.record_write(path)
+        self._session.record_write(path, namespace=self._dataset_namespace)
 
     def get_file(self, path: str, local_path: str) -> None:
         self._storage.get_file(path, local_path)
-        self._session.record_read(path)
+        self._session.record_read(path, namespace=self._dataset_namespace)
 
     def put_bytes(self, data: bytes, path: str) -> None:
         self._storage.put_bytes(data, path)
         facets = _extract_facets(data, path)
-        self._session.record_write(path, facets or None)
+        self._session.record_write(path, facets or None, namespace=self._dataset_namespace)
 
     def get_bytes(self, path: str) -> bytes:
         data = self._storage.get_bytes(path)
         facets = _extract_facets(data, path)
-        self._session.record_read(path, facets or None)
+        self._session.record_read(path, facets or None, namespace=self._dataset_namespace)
         return data
 
     def read(self, path: str) -> DatasetHandle:

{steve_cli-0.3.8 → steve_cli-0.3.11}/steve_cli.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: steve-cli
-Version: 0.3.8
+Version: 0.3.11
 Summary: A simple CLI tool to run jobs from jobs.yaml with proper environment setup
 Author: Frank
 License: MIT
@@ -44,6 +44,8 @@ Requires-Dist: great-expectations>=0.18.0; extra == "great-expectations"
 Requires-Dist: pandas>=2.0.3; extra == "great-expectations"
 Provides-Extra: excel
 Requires-Dist: openpyxl>=3.1.0; extra == "excel"
+Provides-Extra: visidata
+Requires-Dist: visidata>=3.0; extra == "visidata"
 Provides-Extra: all
 Requires-Dist: openlineage-python>=1.0.0; extra == "all"
 Requires-Dist: pyarrow>=17.0.0; extra == "all"
@@ -52,6 +54,7 @@ Requires-Dist: validoopsie>=0.1.0; extra == "all"
 Requires-Dist: great-expectations>=0.18.0; extra == "all"
 Requires-Dist: pandas>=2.0.3; extra == "all"
 Requires-Dist: openpyxl>=3.1.0; extra == "all"
+Requires-Dist: visidata>=3.0; extra == "all"
 
 # Steve CLI
 

{steve_cli-0.3.8 → steve_cli-0.3.11}/steve_cli.egg-info/requires.txt

@@ -12,6 +12,7 @@ validoopsie>=0.1.0
 great-expectations>=0.18.0
 pandas>=2.0.3
 openpyxl>=3.1.0
+visidata>=3.0
 
 [dev]
 pytest>=7.0
@@ -43,3 +44,6 @@ pyarrow>=17.0.0
 [validoopsie]
 validoopsie>=0.1.0
 polars>=1.8.2
+
+[visidata]
+visidata>=3.0