[Feature/interactive cli] - adding interactive CLI for first time users and easy setup #10

README.md

@@ -63,6 +63,7 @@ Configure your preferred LLM provider. Supports both cloud providers (OpenAI, An
 ```bash
 how setup
 ```
+*(Features a keyboard-navigable selection menu for an easier first-time setup experience! Use your arrow keys to navigate and `Enter` to select.)*
 
 **Non-Interactive Mode**:
 ```bash

how/actions.py

@@ -5,6 +5,7 @@
 from rich.console import Console
 from rich.panel import Panel
 from rich.prompt import Prompt
+import questionary
 
 from how.core.safety import assess_risk
 
@@ -82,7 +83,7 @@ def execute_commands(commands: list[str]) -> int:
     if not check_safety_gate(commands):
         return 1
 
-    if not typer.confirm("Are you sure you want to execute these commands?"):
+    if not questionary.confirm("Are you sure you want to execute these commands?").ask():
         console.print("[yellow]Execution cancelled.[/yellow]")
         return 0
 
@@ -95,10 +96,10 @@ def execute_commands(commands: list[str]) -> int:
                     f"[bold red]Command failed with exit code {proc.returncode}[/bold red]"
                 )
                 if idx < len(commands):
-                    continue_run = typer.confirm(
+                    continue_run = questionary.confirm(
                         "Do you want to continue running the remaining commands?",
                         default=False,
-                    )
+                    ).ask()
                     if not continue_run:
                         console.print("[yellow]Execution stopped.[/yellow]")
                         return proc.returncode
@@ -130,7 +131,9 @@ def hook() -> None:
         readline.set_startup_hook(hook)
         edited = input("> ")
     except Exception:  # noqa: BLE001
-        edited = Prompt.ask("Edit command", default=joined)
+        edited = questionary.text("Edit command:", default=joined).ask()
+        if edited is None:
+            edited = joined
     finally:
         with contextlib.suppress(Exception):
             import readline
@@ -151,18 +154,20 @@ def interactive_action_menu(commands: list[str]) -> None:
 
     while True:
         try:
-            console.print(
-                "\n[bold]Options:[/bold] "
-                "[bold green][E]xecute[/bold green] | "
-                "[bold cyan][C]opy[/bold cyan] | "
-                "[bold yellow][M]odify[/bold yellow] | "
-                "[bold red][A]bort[/bold red]"
-            )
-            choice = Prompt.ask(
-                "Select an action",
-                choices=["e", "c", "m", "a", "E", "C", "M", "A"],
-                default="a",
-            ).lower()
+            choice = questionary.select(
+                "Select an action:",
+                choices=[
+                    questionary.Choice("Execute commands", "e"),
+                    questionary.Choice("Copy to clipboard", "c"),
+                    questionary.Choice("Modify command(s)", "m"),
+                    questionary.Choice("Abort", "a"),
+                ],
+                default="a"
+            ).ask()
+
+            if choice is None:
+                console.print("\n[yellow]Aborted.[/yellow]")
+                break
 
             if choice == "e":
                 execute_commands(current_commands)

how/how.py

@@ -247,28 +247,46 @@ def setup(
         selected_model = model or provider_info.get("model", "")
         selected_endpoint = endpoint or provider_info.get("endpoint", "")
     else:
-        provider = Prompt.ask(
-            "Select the LLM Provider", choices=list(LLM_PROVIDERS.keys())
-        )
+        import questionary
+
+        provider = questionary.select(
+            "Select the LLM Provider:",
+            choices=list(LLM_PROVIDERS.keys())
+        ).ask()
+
+        if provider is None:
+            raise typer.Abort()
+
         provider_info = LLM_PROVIDERS[provider]
 
         if provider_info.get("requires_key", True):
-            api_key = Prompt.ask(f"Enter {provider} API Key", password=True)
+            api_key = questionary.password(f"Enter {provider} API Key:").ask()
+            if api_key is None:
+                raise typer.Abort()
         else:
             api_key = ""
 
         default_model = provider_info.get("model", "")
-        selected_model = Prompt.ask("Enter model name", default=model or default_model)
+        selected_model = questionary.text(
+            "Enter model name:", default=model or default_model
+        ).ask()
+
+        if selected_model is None:
+            raise typer.Abort()
 
         if provider == "Ollama" or provider_info.get("endpoint"):
             default_endpoint = provider_info.get("endpoint", "http://localhost:11434")
-            selected_endpoint = Prompt.ask(
-                "Enter endpoint URL", default=endpoint or default_endpoint
-            )
+            selected_endpoint = questionary.text(
+                "Enter endpoint URL:", default=endpoint or default_endpoint
+            ).ask()
+            if selected_endpoint is None:
+                raise typer.Abort()
         else:
             selected_endpoint = endpoint or ""
 
-        typer.confirm("Do you want to save the configuration?", abort=True)
+        confirm = questionary.confirm("Do you want to save the configuration?").ask()
+        if not confirm:
+            raise typer.Abort()
 
     # Test the provider connection before saving
     with console.status("[bold green]Testing LLM connection...[/bold green]"):

requirements.txt

@@ -7,3 +7,4 @@ langchain-openai
 langchain-anthropic
 langchain-ollama
 pyperclip
+questionary

tests/test_actions.py

@@ -22,45 +22,49 @@ def test_copy_to_clipboard_failure():
 
 
 def test_execute_commands_declined():
-    with patch("typer.confirm", return_value=False):
+    with patch("how.actions.questionary.confirm") as mock_confirm:
+        mock_confirm.return_value.ask.return_value = False
         ret = execute_commands(["ls -la"])
         assert ret == 0
 
 
 def test_execute_commands_success():
-    with patch("typer.confirm", return_value=True), patch("subprocess.run") as mock_run:
+    with patch("how.actions.questionary.confirm") as mock_confirm, patch("subprocess.run") as mock_run:
+        mock_confirm.return_value.ask.return_value = True
         mock_run.return_value = MagicMock(returncode=0)
         ret = execute_commands(["echo hello", "echo world"])
         assert ret == 0
         assert mock_run.call_count == 2
 
 
 def test_execute_commands_failure_stop():
-    with patch("typer.confirm", side_effect=[True, False]), patch(
-        "subprocess.run"
-    ) as mock_run:
+    with patch("how.actions.questionary.confirm") as mock_confirm, patch("subprocess.run") as mock_run:
         # First confirm is to execute, second confirm is whether to continue after failure
+        mock_confirm.return_value.ask.side_effect = [True, False]
         mock_run.return_value = MagicMock(returncode=42)
         ret = execute_commands(["false", "echo unreachable"])
         assert ret == 42
         assert mock_run.call_count == 1
 
 
 def test_modify_commands():
-    with patch("builtins.input", return_value="echo edited"):
+    with patch("builtins.input", side_effect=Exception("force fallback")), patch("how.actions.questionary.text") as mock_text:
+        mock_text.return_value.ask.return_value = "echo edited"
         res = modify_commands(["echo original"])
         assert res == ["echo edited"]
 
 
 def test_interactive_action_menu_abort():
-    with patch("rich.prompt.Prompt.ask", return_value="a"):
+    with patch("how.actions.questionary.select") as mock_select:
+        mock_select.return_value.ask.return_value = "a"
         # Should exit cleanly without error
         interactive_action_menu(["echo test"])
 
 
 def test_interactive_action_menu_execute():
-    with patch("rich.prompt.Prompt.ask", return_value="e"), patch(
+    with patch("how.actions.questionary.select") as mock_select, patch(
         "how.actions.execute_commands"
     ) as mock_exec:
+        mock_select.return_value.ask.return_value = "e"
         interactive_action_menu(["echo test"])
         mock_exec.assert_called_once_with(["echo test"])

tests/test_cli.py

@@ -0,0 +1,34 @@
+from unittest.mock import patch
+from typer.testing import CliRunner
+from how.how import app
+from how.core.config import Config
+
+runner = CliRunner()
+
+def test_setup_non_interactive(tmp_path):
+    with patch("how.how.config", Config(config_dir=tmp_path)):
+        # Test missing provider flag
+        result = runner.invoke(app, ["setup", "--no-interactive"])
+        assert result.exit_code != 0
+        assert "Please set the --provider" in result.stdout
+
+        # Test missing api key for provider that needs it
+        result = runner.invoke(app, ["setup", "--no-interactive", "--provider", "OpenAI"])
+        assert result.exit_code != 0
+        assert "Please provide --api-key for provider 'OpenAI'" in result.stdout
+
+        # Test valid non-interactive setup (Ollama doesn't require API key)
+        with patch("how.core.llm.get_llm") as mock_get_llm:
+            # Mock get_llm to avoid actually calling the LLM
+            mock_llm = mock_get_llm.return_value
+            mock_llm.invoke.return_value = "Success"
+
+            result = runner.invoke(app, [
+                "setup", 
+                "--no-interactive", 
+                "--provider", "Ollama",
+                "--endpoint", "http://localhost:11434"
+            ])
+
+            assert result.exit_code == 0
+            assert "Configuration saved successfully" in result.stdout

tests/test_context.py

@@ -35,14 +35,15 @@ def test_detect_project_no_markers(tmp_path: Path):
 
 
 def test_detect_shell_unix():
-    with patch.dict("os.environ", {"SHELL": "/bin/zsh"}):
-        assert detect_shell() == "zsh"
+    with patch("platform.system", return_value="Linux"):
+        with patch.dict("os.environ", {"SHELL": "/bin/zsh"}):
+            assert detect_shell() == "zsh"
 
-    with patch.dict("os.environ", {"SHELL": "/usr/bin/bash"}):
-        assert detect_shell() == "bash"
+        with patch.dict("os.environ", {"SHELL": "/usr/bin/bash"}):
+            assert detect_shell() == "bash"
 
-    with patch.dict("os.environ", {"SHELL": "/usr/bin/fish"}):
-        assert detect_shell() == "fish"
+        with patch.dict("os.environ", {"SHELL": "/usr/bin/fish"}):
+            assert detect_shell() == "fish"
 
 
 def test_detect_package_managers():