feat: Update env command to prefer current environment values in .env… (#193)

Author: sam-hey

docs/developer-guide/cli.md

@@ -80,7 +80,7 @@ Use the `wurzel env` helper to inspect or validate the variables your pipeline n
 # Show required env vars (toggle optional ones via --only-required)
 wurzel env examples.pipeline.pipelinedemo:pipeline --only-required
 
-# Generate a .env snippet with defaults
+# Generate a .env snippet (prefers values already present in your shell)
 wurzel env examples.pipeline.pipelinedemo:pipeline --gen-env > .env.sample
 
 # Fail fast when something is missing

tests/cli/cmd_env_test.py

@@ -45,6 +45,21 @@ def test_format_env_snippet_matches_expected_layout():
     )
 
 
+def test_format_env_snippet_prefers_current_env_values():
+    reqs = cmd_env.collect_env_requirements(pipelinedemo.pipeline)
+
+    snippet = cmd_env.format_env_snippet(
+        reqs,
+        current_env={
+            "MANUALMARKDOWNSTEP__FOLDER_PATH": "/tmp/data",
+            "SIMPLESPLITTERSTEP__BATCH_SIZE": "256",
+        },
+    )
+
+    assert "MANUALMARKDOWNSTEP__FOLDER_PATH=/tmp/data" in snippet
+    assert "SIMPLESPLITTERSTEP__BATCH_SIZE=256" in snippet
+
+
 def test_validate_env_vars_reports_missing(env):
     env.clear()
     issues = cmd_env.validate_env_vars(pipelinedemo.pipeline, allow_extra_fields=False)

tests/cli/main_test.py

@@ -199,7 +199,30 @@ def test_env_only_required_filters_optional(capsys, monkeypatch):
     assert "SIMPLESPLITTERSTEP__BATCH_SIZE" not in captured.out
 
 
-def test_env_gen_env_outputs_env_file(capsys, monkeypatch):
+def test_env_gen_env_outputs_env_file(capsys, monkeypatch, env):
+    monkeypatch.setattr(main, "console", main.console.__class__(force_terminal=False, width=200))
+    env.set("MANUALMARKDOWNSTEP__FOLDER_PATH", "/tmp/custom")
+    env.set("SIMPLESPLITTERSTEP__BATCH_SIZE", "256")
+
+    main.env_cmd("examples.pipeline.pipelinedemo:pipeline", gen_env=True)
+    captured = capsys.readouterr()
+    expected = (
+        "# Generated env vars\n\n"
+        "# ManualMarkdownStep\n"
+        "MANUALMARKDOWNSTEP__FOLDER_PATH=/tmp/custom\n\n"  # pragma: allowlist secret
+        "# SimpleSplitterStep\n"
+        "SIMPLESPLITTERSTEP__BATCH_SIZE=256\n"
+        "SIMPLESPLITTERSTEP__NUM_THREADS=4\n"
+        "SIMPLESPLITTERSTEP__TOKEN_COUNT_MIN=64\n"
+        "SIMPLESPLITTERSTEP__TOKEN_COUNT_MAX=1024\n"
+        "SIMPLESPLITTERSTEP__TOKEN_COUNT_BUFFER=32\n"
+        "SIMPLESPLITTERSTEP__TOKENIZER_MODEL=gpt-3.5-turbo\n"
+        "SIMPLESPLITTERSTEP__SENTENCE_SPLITTER_MODEL=de_core_news_sm\n\n\n"
+    )
+    assert captured.out == expected
+
+
+def test_env_gen_env_outputs_env_file_empty(capsys, monkeypatch):
     monkeypatch.setattr(main, "console", main.console.__class__(force_terminal=False, width=200))
     main.env_cmd("examples.pipeline.pipelinedemo:pipeline", gen_env=True)
     captured = capsys.readouterr()

wurzel/cli/_main.py

@@ -500,7 +500,7 @@ def env_cmd(
         return
 
     if gen_env:
-        typer.echo(format_env_snippet(to_display))
+        typer.echo(format_env_snippet(to_display, current_env=os.environ))
         return
 
     _print_requirements(to_display)

wurzel/cli/cmd_env.py

@@ -5,6 +5,7 @@
 
 from __future__ import annotations
 
+from collections.abc import Mapping
 from json import dumps
 from types import NoneType
 from typing import TYPE_CHECKING
@@ -74,19 +75,23 @@ def collect_env_requirements(pipeline: TypedStep) -> list[EnvVarRequirement]:
     return sorted(requirements, key=lambda req: (req.step_index, req.field_index))
 
 
-def format_env_snippet(requirements: list[EnvVarRequirement]) -> str:
+def format_env_snippet(
+    requirements: list[EnvVarRequirement],
+    current_env: Mapping[str, str] | None = None,
+) -> str:
     """Return .env-style representation of requirements grouped by step."""
     lines: list[str] = ["# Generated env vars"]
     current_step = None
+    env_values = current_env or {}
     for req in requirements:
         if req.step_name != current_step:
             lines.append("")
             lines.append(f"# {req.step_name}")
             current_step = req.step_name
-        default = req.default
-        if not default:
-            default = ""
-        lines.append(f"{req.env_var}={default}")
+        value = req.default or ""
+        if current_env is not None and req.env_var in env_values:
+            value = env_values[req.env_var]
+        lines.append(f"{req.env_var}={value}")
     lines.append("")
     return "\n".join(lines) + "\n"