llamactl 0.3.0a12__py3-none-any.whl → 0.3.0a14__py3-none-any.whl

llama_deploy/cli/options.py

@@ -2,6 +2,9 @@ import logging
 from typing import Callable, ParamSpec, TypeVar
 
 import click
+from llama_deploy.cli.interactive_prompts.session_utils import is_interactive_session
+
+from .debug import setup_file_logging
 
 P = ParamSpec("P")
 R = TypeVar("R")
@@ -9,7 +12,6 @@ R = TypeVar("R")
 
 def global_options(f: Callable[P, R]) -> Callable[P, R]:
     """Common decorator to add global options to command groups"""
-    from .debug import setup_file_logging
 
     def debug_callback(ctx: click.Context, param: click.Parameter, value: str) -> str:
         if value:
@@ -27,3 +29,15 @@ def global_options(f: Callable[P, R]) -> Callable[P, R]:
         is_eager=True,
         hidden=True,
     )(f)
+
+
+def interactive_option(f: Callable[P, R]) -> Callable[P, R]:
+    """Add an interactive option to the command"""
+
+    default = is_interactive_session()
+    return click.option(
+        "--interactive/--no-interactive",
+        help="Run in interactive mode. If not provided, will default to the current session's interactive state.",
+        is_flag=True,
+        default=default,
+    )(f)

llama_deploy/cli/platform_client.py

@@ -0,0 +1,52 @@
+from __future__ import annotations
+
+from datetime import datetime
+
+import httpx
+from pydantic import BaseModel, TypeAdapter
+
+
+class PlatformClient:
+    def __init__(self, auth_token: str, platform_url: str):
+        self.auth_token = auth_token
+        self.platform_url = platform_url
+        self.client = httpx.AsyncClient(base_url=platform_url)
+
+    async def list_projects(self) -> list[Project]:
+        response = await self.client.get("/api/v1/projects")
+        response.raise_for_status()
+        return ProjectList.validate_python(response.json())
+
+    async def list_organizations(self) -> list[Organization]:
+        response = await self.client.get("/api/v1/organizations")
+        response.raise_for_status()
+        return OrganizationList.validate_python(response.json())
+
+    async def validate_auth_token(self) -> bool:
+        response = await self.client.get("/api/v1/organizations/default")
+        try:
+            response.raise_for_status()
+            return True
+        except httpx.HTTPStatusError:
+            if response.status_code == 401:
+                return False
+            raise
+
+
+class Organization(BaseModel):
+    id: str
+    name: str
+    created_at: datetime
+    updated_at: datetime
+
+
+class Project(BaseModel):
+    id: str
+    name: str
+    organization_id: str
+    created_at: datetime
+    updated_at: datetime
+
+
+OrganizationList = TypeAdapter(list[Organization])
+ProjectList = TypeAdapter(list[Project])

llama_deploy/cli/textual/deployment_form.py

@@ -405,7 +405,9 @@ class DeploymentEditApp(App[DeploymentResponse | None]):
         self.save_error = ""  # Clear any previous errors
         self.current_state = "validation"
 
-    def on_validation_result_message(self, message: ValidationResultMessage) -> None:
+    async def on_validation_result_message(
+        self, message: ValidationResultMessage
+    ) -> None:
         """Handle validation success from git validation widget"""
         logging.info("validation result message", message)
         # Update form data with validated PAT if provided
@@ -417,8 +419,8 @@ class DeploymentEditApp(App[DeploymentResponse | None]):
                 updated_form.has_existing_pat = False
             self.form_data = updated_form
 
-        # Proceed with save
-        self._perform_save()
+        # Proceed with save (async)
+        await self._perform_save()
 
     def on_validation_cancel_message(self, message: ValidationCancelMessage) -> None:
         """Handle validation cancellation from git validation widget"""
@@ -435,18 +437,17 @@ class DeploymentEditApp(App[DeploymentResponse | None]):
         """Return from help to form, keeping form state intact."""
         self.current_state = "form"
 
-    def _perform_save(self) -> None:
+    async def _perform_save(self) -> None:
         """Actually save the deployment after validation"""
         logging.info("saving form data", self.form_data)
         result = self.form_data
         client = get_client()
         try:
-            if result.is_editing:
-                update_deployment = client.update_deployment(
-                    result.id, result.to_update()
-                )
-            else:
-                update_deployment = client.create_deployment(result.to_create())
+            update_deployment = (
+                await client.update_deployment(result.id, result.to_update())
+                if result.is_editing
+                else await client.create_deployment(result.to_create())
+            )
             # Save and navigate to embedded monitor screen
             self.saved_deployment = update_deployment
             # Ensure form_data carries the new ID for any subsequent operations

llama_deploy/cli/textual/deployment_monitor.py

@@ -7,11 +7,11 @@ import hashlib
 import threading
 import time
 from pathlib import Path
-from typing import Iterator
 
-from llama_deploy.cli.client import get_project_client as get_client
-from llama_deploy.core.client.manage_client import Closer
-from llama_deploy.core.schema.base import LogEvent
+from llama_deploy.cli.client import (
+    project_client_context,
+)
+from llama_deploy.core.schema import LogEvent
 from llama_deploy.core.schema.deployments import DeploymentResponse
 from rich.text import Text
 from textual import events
@@ -78,7 +78,6 @@ class DeploymentMonitorWidget(Widget):
     error_message = reactive("", recompose=False)
     wrap_enabled = reactive(False, recompose=False)
     autoscroll_enabled = reactive(True, recompose=False)
-    stream_closer: Closer | None = None
 
     def __init__(self, deployment_id: str) -> None:
         super().__init__()
@@ -87,12 +86,12 @@ class DeploymentMonitorWidget(Widget):
         # Persist content written to the RichLog across recomposes
         self._log_buffer: list[Text] = []
 
-    def on_mount(self) -> None:
+    async def on_mount(self) -> None:
         # Kick off initial fetch and start logs stream in background
-        self.run_worker(self._fetch_deployment(), exclusive=True)
-        self.run_worker(self._stream_logs, exclusive=False, thread=True)
+        self.run_worker(self._fetch_deployment())
+        self.run_worker(self._stream_logs())
         # Start periodic polling of deployment status
-        self.run_worker(self._poll_deployment_status(), exclusive=False)
+        self.run_worker(self._poll_deployment_status())
 
     def compose(self) -> ComposeResult:
         yield Static("Deployment Status", classes="primary-message")
@@ -141,81 +140,117 @@ class DeploymentMonitorWidget(Widget):
 
     async def _fetch_deployment(self) -> None:
         try:
-            client = get_client()
-            self.deployment = client.get_deployment(
-                self.deployment_id, include_events=True
-            )
+            async with project_client_context() as client:
+                self.deployment = await client.get_deployment(
+                    self.deployment_id, include_events=True
+                )
             # Clear any previous error on success
             self.error_message = ""
         except Exception as e:  # pragma: no cover - network errors
             self.error_message = f"Failed to fetch deployment: {e}"
 
-    def _stream_logs(self) -> None:
-        """Consume the blocking log iterator in a single worker thread.
+    async def _stream_logs(self) -> None:
+        """Consume the async log iterator, batch updates, and reconnect with backoff."""
 
-        Cooperative cancellation uses `self._stop_stream` to exit cleanly.
-        """
-        client = get_client()
-
-        def _sleep_with_cancel(total_seconds: float) -> None:
+        async def _sleep_with_cancel(total_seconds: float) -> None:
             step = 0.2
             remaining = total_seconds
             while remaining > 0 and not self._stop_stream.is_set():
-                time.sleep(min(step, remaining))
+                await asyncio.sleep(min(step, remaining))
                 remaining -= step
 
+        # Batching configuration: small latency to reduce UI churn while staying responsive
+        batch_max_latency_seconds = 0.1
+        batch_max_items = 200
+
         base_backoff_seconds = 0.2
         backoff_seconds = base_backoff_seconds
         max_backoff_seconds = 30.0
 
         while not self._stop_stream.is_set():
-            try:
-                connect_started_at = time.monotonic()
-                closer, stream = client.stream_deployment_logs(
-                    self.deployment_id,
-                    include_init_containers=True,
-                )
-                # On any (re)connect, clear existing content
-                self.app.call_from_thread(self._reset_log_view_for_reconnect)
-
-                buffered_stream = _buffer_log_lines(stream)
-
-                def close_stream():
+            connect_started_at = time.monotonic()
+            # On any (re)connect, clear existing content
+            self._reset_log_view_for_reconnect()
+
+            queue: asyncio.Queue[LogEvent] = asyncio.Queue(maxsize=10000)
+            producer_done = asyncio.Event()
+
+            async def _producer() -> None:
+                try:
+                    async with project_client_context() as client:
+                        async for event in client.stream_deployment_logs(
+                            self.deployment_id,
+                            include_init_containers=True,
+                            tail_lines=10000,
+                        ):
+                            if self._stop_stream.is_set():
+                                break
+                            try:
+                                await queue.put(event)
+                            except Exception:
+                                # If queue put fails due to cancellation/shutdown, stop
+                                break
+                except Exception as e:
+                    # Surface error via error message and rely on reconnect loop
+                    if not self._stop_stream.is_set():
+                        self._set_error_message(
+                            f"Log stream failed: {e}. Reconnecting..."
+                        )
+                finally:
+                    producer_done.set()
+
+            async def _consumer() -> None:
+                batch: list[LogEvent] = []
+                next_deadline = time.monotonic() + batch_max_latency_seconds
+                while not self._stop_stream.is_set():
+                    # Stop once producer finished and queue drained
+                    if producer_done.is_set() and queue.empty():
+                        if batch:
+                            self._handle_log_events(batch)
+                            batch = []
+                        break
+                    timeout = max(0.0, next_deadline - time.monotonic())
                     try:
-                        closer()
+                        item = await asyncio.wait_for(queue.get(), timeout=timeout)
+                        batch.append(item)
+                        if len(batch) >= batch_max_items:
+                            self._handle_log_events(batch)
+                            batch = []
+                            next_deadline = time.monotonic() + batch_max_latency_seconds
+                    except asyncio.TimeoutError:
+                        if batch:
+                            self._handle_log_events(batch)
+                            batch = []
+                        next_deadline = time.monotonic() + batch_max_latency_seconds
                     except Exception:
-                        pass
-
-                self.stream_closer = close_stream
-                # Stream connected; consume until end
-                for events in buffered_stream:
-                    if self._stop_stream.is_set():
+                        # On any unexpected error, flush and exit, reconnect will handle
+                        if batch:
+                            self._handle_log_events(batch)
                         break
-                    # Marshal UI updates back to the main thread via the App
-                    self.app.call_from_thread(self._handle_log_events, events)
-                if self._stop_stream.is_set():
-                    break
-                # Stream ended without explicit error; attempt reconnect
-                self.app.call_from_thread(
-                    self._set_error_message, "Log stream disconnected. Reconnecting..."
-                )
-            except Exception as e:
-                if self._stop_stream.is_set():
-                    break
-                # Surface the error to the UI and attempt reconnect with backoff
-                self.app.call_from_thread(
-                    self._set_error_message, f"Log stream failed: {e}. Reconnecting..."
-                )
 
-            # Duration-aware backoff: subtract how long the last connection lived
+            producer_task = asyncio.create_task(_producer())
+            try:
+                await _consumer()
+            finally:
+                # Ensure producer is not left running
+                try:
+                    producer_task.cancel()
+                except Exception:
+                    pass
+
+            if self._stop_stream.is_set():
+                break
+
+            # If we reached here, the stream ended or failed; attempt reconnect with backoff
+            self._set_error_message("Log stream disconnected. Reconnecting...")
+
+            # Duration-aware backoff (smaller when the previous connection lived longer)
             connection_lifetime = 0.0
             try:
                 connection_lifetime = max(0.0, time.monotonic() - connect_started_at)
             except Exception:
                 connection_lifetime = 0.0
 
-            # If the connection lived longer than the current backoff window,
-            # reset to base so the next reconnect is immediate.
             if connection_lifetime >= backoff_seconds:
                 backoff_seconds = base_backoff_seconds
             else:
@@ -223,7 +258,7 @@ class DeploymentMonitorWidget(Widget):
 
             delay = max(0.0, backoff_seconds - connection_lifetime)
             if delay > 0:
-                _sleep_with_cancel(delay)
+                await _sleep_with_cancel(delay)
 
     def _reset_log_view_for_reconnect(self) -> None:
         """Clear UI and buffers so new stream replaces previous content."""
@@ -326,9 +361,6 @@ class DeploymentMonitorWidget(Widget):
     def on_unmount(self) -> None:
         # Attempt to stop the streaming loop
         self._stop_stream.set()
-        if self.stream_closer is not None:
-            self.stream_closer()
-            self.stream_closer = None
 
     # Reactive watchers to update widgets in place instead of recomposing
     def watch_error_message(self, message: str) -> None:
@@ -383,12 +415,12 @@ class DeploymentMonitorWidget(Widget):
 
     async def _poll_deployment_status(self) -> None:
         """Periodically refresh deployment status to reflect updates in the UI."""
-        client = get_client()
         while not self._stop_stream.is_set():
             try:
-                self.deployment = client.get_deployment(
-                    self.deployment_id, include_events=True
-                )
+                async with project_client_context() as client:
+                    self.deployment = await client.get_deployment(
+                        self.deployment_id, include_events=True
+                    )
                 # Clear any previous error on success
                 if self.error_message:
                     self.error_message = ""
@@ -434,42 +466,3 @@ def monitor_deployment_screen(deployment_id: str) -> None:
     """Launch the standalone deployment monitor screen."""
     app = DeploymentMonitorApp(deployment_id)
     app.run()
-
-
-def _buffer_log_lines(iter: Iterator[LogEvent]) -> Iterator[list[LogEvent]]:
-    """Batch log events into small lists using a background reader.
-
-    This reduces UI churn while still reacting quickly. On shutdown we
-    absorb stream read errors that are expected when the connection is
-    closed from another thread.
-    """
-    buffer: list[LogEvent] = []
-    bg_error: Exception | None = None
-    done = threading.Event()
-
-    def pump() -> None:
-        nonlocal bg_error
-        try:
-            for event in iter:
-                buffer.append(event)
-        except Exception as e:
-            bg_error = e
-        finally:
-            done.set()
-
-    t = threading.Thread(target=pump, daemon=True)
-    t.start()
-    try:
-        while not done.is_set():
-            if buffer:
-                # Yield a snapshot and clear in-place to avoid reallocating list
-                yield list(buffer)
-                buffer.clear()
-            time.sleep(0.5)
-        if bg_error is not None:
-            raise bg_error
-    finally:
-        try:
-            t.join(timeout=0.1)
-        except Exception:
-            pass

llama_deploy/cli/textual/git_validation.py

@@ -282,14 +282,14 @@ class GitValidationWidget(Widget):
         self.error_message = ""
         try:
             client = get_client()
-            response = client.validate_repository(
+            self.validation_response = await client.validate_repository(
                 repo_url=self.repo_url, deployment_id=self.deployment_id, pat=pat
             )
-            self.validation_response = response
 
-            if response.accessible:
+            resp = self.validation_response
+            if resp and resp.accessible:
                 # Success - post result message with appropriate messaging
-                if response.pat_is_obsolete:
+                if resp.pat_is_obsolete:
                     # Show success message about PAT obsolescence before proceeding
                     self.current_state = "success"
                     self.error_message = "Repository accessible via GitHub App. Your Personal Access Token is now obsolete and will be removed."
@@ -308,22 +308,24 @@ class GitValidationWidget(Widget):
         self.error_message = ""
         try:
             client = get_client()
-            response = client.validate_repository(
+            self.validation_response = await client.validate_repository(
                 repo_url=self.repo_url, deployment_id=self.deployment_id
             )
-            self.validation_response = response
 
-            if response.accessible:
+            resp = self.validation_response
+            if resp and resp.accessible:
                 # Success - post result message with appropriate messaging
                 self.current_state = "success"
                 self.post_message(
                     ValidationResultMessage(
-                        self.repo_url, "" if response.pat_is_obsolete else None
+                        self.repo_url, "" if resp.pat_is_obsolete else None
                     )
                 )
             else:
                 # Failed - stay in github_auth and show error
-                self.error_message = f"Still not accessible: {response.message}"
+                self.error_message = (
+                    f"Still not accessible: {resp.message if resp else ''}"
+                )
 
         except Exception as e:
             # Failed - stay in github_auth and show error

llama_deploy/cli/textual/styles.tcss

@@ -30,6 +30,10 @@ Container {
     height: auto;
 }
 
+.two-column-form-grid .full-width {
+    column-span: 2;
+}
+
 /* =============================================== */
 /* FORM ELEMENTS                                  */
 /* =============================================== */
@@ -96,7 +100,7 @@ Input.disabled {
     background: $error-muted;
     border-left: heavy $error;
     margin: 0 0 1 0;
-    padding: 0 0 0 1
+    padding: 0 0 0 1;
 }
 
 .primary-message {
@@ -104,7 +108,7 @@ Input.disabled {
     background: $primary-muted;
     border-left: heavy $primary;
     margin: 0 0 1 0;
-    padding: 0 0 0 1
+    padding: 0 0 0 1;
 }
 
 .secondary-message {
@@ -112,15 +116,23 @@ Input.disabled {
     background: $secondary-muted;
     border-left: heavy $secondary;
     margin: 0 0 1 0;
-    padding: 0 0 0 1
+    padding: 0 0 0 1;
 }
 
 .success-message {
     color: $text-success;
     background: $success-muted;
     border-left: heavy $success;
-    padding: 1;
-    margin: 1 0;
+    padding: 0 0 0 1;
+    margin: 0 0 0 0;
+}
+
+.warning-message {
+    color: $text-warning;
+    background: $warning-muted;
+    border-left: heavy $warning;
+    padding: 0 0 0 1;
+    margin: 0 0 0 0;
 }
 
 .hidden {
@@ -142,6 +154,10 @@ Input.disabled {
     width: 1fr;
 }
 
+.align-right {
+    align: right middle;
+}
+
 /* =============================================== */
 /* BUTTONS & ACTIONS                              */
 /* =============================================== */

{llamactl-0.3.0a12.dist-info → llamactl-0.3.0a14.dist-info}/METADATA

@@ -1,12 +1,12 @@
 Metadata-Version: 2.3
 Name: llamactl
-Version: 0.3.0a12
+Version: 0.3.0a14
 Summary: A command-line interface for managing LlamaDeploy projects and deployments
 Author: Adrian Lyjak
 Author-email: Adrian Lyjak <adrianlyjak@gmail.com>
 License: MIT
-Requires-Dist: llama-deploy-core[client]>=0.3.0a12,<0.4.0
-Requires-Dist: llama-deploy-appserver>=0.3.0a12,<0.4.0
+Requires-Dist: llama-deploy-core[client]>=0.3.0a14,<0.4.0
+Requires-Dist: llama-deploy-appserver>=0.3.0a14,<0.4.0
 Requires-Dist: httpx>=0.24.0
 Requires-Dist: rich>=13.0.0
 Requires-Dist: questionary>=2.0.0

llamactl-0.3.0a14.dist-info/RECORD

@@ -0,0 +1,32 @@
+llama_deploy/cli/__init__.py,sha256=df028686233c4d5a3e244bb50c1c7b84cf2399ae03abe45eb4d01e53caa1be38,476
+llama_deploy/cli/app.py,sha256=9170e4f506c482522bd745eb1cdb700a198cfcfd7204c168c94e5ee2b6b43ffa,2199
+llama_deploy/cli/client.py,sha256=f0f72c90cddfbc9198e154883f3b8f05fb47dbe7ec1f5755106dbb8009d2bb54,1459
+llama_deploy/cli/commands/aliased_group.py,sha256=bc41007c97b7b93981217dbd4d4591df2b6c9412a2d9ed045b0ec5655ed285f2,1066
+llama_deploy/cli/commands/auth.py,sha256=de584d11c1acf5a4e7aee8c8f30184335053ed206d38dbc8509b2d1d0677a092,12640
+llama_deploy/cli/commands/deployment.py,sha256=c99feb73a887063cad86e2cc555f21ebac6d47576749d79c551f26b8567af638,9879
+llama_deploy/cli/commands/env.py,sha256=e0b96b9f4e7921b4370ad5f8bc0a2bfb19b705e73004f72c37c9bed28a208e0d,6702
+llama_deploy/cli/commands/init.py,sha256=51b2de1e35ff34bc15c9dfec72fbad08aaf528c334df168896d36458a4e9401c,6307
+llama_deploy/cli/commands/serve.py,sha256=4d47850397ba172944df56a934a51bedb52403cbd3f9b000b1ced90a31c75049,2721
+llama_deploy/cli/config/_config.py,sha256=31376c3f3ecadaf52545e4aeb4a660dde2c22d7e6c33b232ec93eebeb926e3fb,14953
+llama_deploy/cli/config/auth_service.py,sha256=38a2de9bbcf5780675130d90e4f9116da190fd31a33e1c178a7f3847b943c229,2667
+llama_deploy/cli/config/env_service.py,sha256=6e50ffe2e33dcff4193c87b3df8daac71542d406314a7bcaf48187715a53780e,2445
+llama_deploy/cli/config/schema.py,sha256=086b6161b238c2037068a2b510f5d4bbda917494df764818ff9692e9735a8953,608
+llama_deploy/cli/debug.py,sha256=e85a72d473bbe1645eb31772f7349bde703d45704166f767385895c440afc762,496
+llama_deploy/cli/env.py,sha256=6ebc24579815b3787829c81fd5bb9f31698a06e62c0128a788559f962b33a7af,1016
+llama_deploy/cli/interactive_prompts/session_utils.py,sha256=b996f2eddf70d6c49636c4797d246d212fce0950fe7e9a3f59cf6a1bf7ae26f5,1142
+llama_deploy/cli/interactive_prompts/utils.py,sha256=594cc2a242cc3405d66d0e26a60647496cc5fcb4ce7d0500a4cfec4888c9a0fa,516
+llama_deploy/cli/options.py,sha256=62ee7286c3305ddb4b597783d19e854284d79bf9384800045f15b934dc245c1d,1298
+llama_deploy/cli/platform_client.py,sha256=69de23dc79a8f5922afc9e3bac1b633a531340ebbefeb7838e3a88419faa754c,1451
+llama_deploy/cli/py.typed,sha256=e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855,0
+llama_deploy/cli/textual/deployment_form.py,sha256=1cf186b765d10a1bdd7394f22ddd7598d75dba8c9a8a7a8be74e6151da2f31dc,20941
+llama_deploy/cli/textual/deployment_help.py,sha256=d43e9ff29db71a842cf8b491545763d581ede3132b8af518c73af85a40950046,2464
+llama_deploy/cli/textual/deployment_monitor.py,sha256=7bcf3f0213401c2432fdb5a9d9acf468a4afe83b2d86d7f2852319768e6f2534,17231
+llama_deploy/cli/textual/git_validation.py,sha256=94c95b61d0cbc490566a406b4886c9c12e1d1793dc14038a5be37119223c9568,13419
+llama_deploy/cli/textual/github_callback_server.py,sha256=dc74c510f8a98ef6ffaab0f6d11c7ea86ee77ca5adbc7725a2a29112bae24191,7556
+llama_deploy/cli/textual/llama_loader.py,sha256=33cb32a46dd40bcf889c553e44f2672c410e26bd1d4b17aa6cca6d0a5d59c2c4,1468
+llama_deploy/cli/textual/secrets_form.py,sha256=a43fbd81aad034d0d60906bfd917c107f9ace414648b0f63ac0b29eeba4050db,7061
+llama_deploy/cli/textual/styles.tcss,sha256=b1a54dc5fb0e0aa12cbf48807e9e6a94b9926838b8058dae1336a134f02e92b0,3327
+llamactl-0.3.0a14.dist-info/WHEEL,sha256=66530aef82d5020ef5af27ae0123c71abb9261377c5bc519376c671346b12918,79
+llamactl-0.3.0a14.dist-info/entry_points.txt,sha256=b67e1eb64305058751a651a80f2d2268b5f7046732268421e796f64d4697f83c,52
+llamactl-0.3.0a14.dist-info/METADATA,sha256=9b2916ded8d6a8edfa8ee764b30b115eba00d9bbcb95b5c98e37def2e9bee3f1,3177
+llamactl-0.3.0a14.dist-info/RECORD,,