Misc improvements to display=plain (#1166)

- Handle multiple models and multiple tasks (display task/model to disambiguate if required)
- Pad display so columns always line up
- Throttle updates to once every second
- Condense task panel when there is no body or footer
- Call rich_initialise in PlainDisplay constructor so ansi colors, etc. are disabled globally for rich printing.
- Remove spurious plain print from rich display (since plain now has its own display handler)

CHANGELOG.md

@@ -13,6 +13,7 @@
 - Log warning when a non-fatal sample error occurs (i.e. errors permitted by the `fail_on_error` option) 
 - Inspect View: allow filtering samples by compound expressions including multiple scorers. (thanks @andrei-apollo)
 - Inspect View: improve rendering performance and stability for the viewer when viewing very large eval logs or samples with a large number of steps.
+- Task display: Improved `plain` mode with periodic updates on progress, metrics, etc.
 
 ## v0.3.58 (16 January 2025)
 

src/inspect_ai/_display/core/panel.py

@@ -50,9 +50,13 @@ def task_panel(
         table.add_row(subtitle_table)
 
     # main progress and task info
-    table.add_row()
-    table.add_row(body)
-    table.add_row()
+    if body:
+        table.add_row()
+        table.add_row(body)
+
+    # spacing if there is more ocontent
+    if footer or log_location:
+        table.add_row()
 
     # footer if specified
     if footer:

src/inspect_ai/_display/plain/display.py

@@ -4,6 +4,10 @@
 
 import rich
 
+from inspect_ai._display.core.rich import rich_initialise
+from inspect_ai._util.text import truncate
+from inspect_ai._util.throttle import throttle
+
 from ...util._concurrency import concurrency_status
 from ..core.config import task_config
 from ..core.display import (
@@ -28,6 +32,7 @@ def __init__(self) -> None:
         self.total_tasks: int = 0
         self.tasks: list[TaskWithResult] = []
         self.parallel = False
+        rich_initialise()
 
     def print(self, message: str) -> None:
         print(message)
@@ -48,6 +53,8 @@ async def task_screen(
         self, tasks: list[TaskSpec], parallel: bool
     ) -> AsyncIterator[TaskScreen]:
         self.total_tasks = len(tasks)
+        self.multiple_task_names = len({task.name for task in tasks}) > 1
+        self.multiple_model_names = len({str(task.model) for task in tasks}) > 1
         self.tasks = []
         self.parallel = parallel
         try:
@@ -58,7 +65,6 @@ async def task_screen(
         finally:
             # Print final results
             if self.tasks:
-                print("\nResults:")
                 self._print_results()
 
     @contextlib.contextmanager
@@ -72,13 +78,16 @@ def task(self, profile: TaskProfile) -> Iterator[TaskDisplay]:
             footer=None,
             log_location=None,
         )
-        print("Running task:")
         rich.print(panel)
 
         # Create and yield task display
         task = TaskWithResult(profile, None)
         self.tasks.append(task)
-        yield PlainTaskDisplay(task)
+        yield PlainTaskDisplay(
+            task,
+            show_task_names=self.multiple_task_names,
+            show_model_names=self.multiple_model_names,
+        )
 
     def _print_results(self) -> None:
         """Print final results using rich panels"""
@@ -100,8 +109,12 @@ def complete(self) -> None:
 
 
 class PlainTaskDisplay(TaskDisplay):
-    def __init__(self, task: TaskWithResult):
+    def __init__(
+        self, task: TaskWithResult, *, show_task_names: bool, show_model_names: bool
+    ):
         self.task = task
+        self.show_task_names = show_task_names
+        self.show_model_names = show_model_names
         self.progress_display: PlainProgress | None = None
         self.samples_complete = 0
         self.samples_total = 0
@@ -113,6 +126,10 @@ def progress(self) -> Iterator[Progress]:
         self.progress_display = PlainProgress(self.task.profile.steps)
         yield self.progress_display
 
+    @throttle(1)
+    def _print_status_throttled(self) -> None:
+        self._print_status()
+
     def _print_status(self) -> None:
         """Print status updates on new lines when there's meaningful progress"""
         if not self.progress_display:
@@ -125,16 +142,25 @@ def _print_status(self) -> None:
 
         # Only print on percentage changes to avoid too much output
         if current_progress != self.last_progress:
-            status_parts = []
+            status_parts: list[str] = []
+
+            # if this is parallel print task and model to distinguish (limit both to 12 chars)
+            MAX_NAME_WIDTH = 12
+            if self.show_task_names:
+                status_parts.append(truncate(self.task.profile.name, MAX_NAME_WIDTH))
+            if self.show_model_names:
+                status_parts.append(
+                    truncate(str(self.task.profile.model), MAX_NAME_WIDTH)
+                )
 
             # Add step progress
             status_parts.append(
-                f"Steps: {self.progress_display.current}/{self.progress_display.total} ({current_progress}%)"
+                f"Steps: {self.progress_display.current:3d}/{self.progress_display.total} {current_progress:3d}%"
             )
 
             # Add sample progress
             status_parts.append(
-                f"Samples: {self.samples_complete}/{self.samples_total}"
+                f"Samples: {self.samples_complete:3d}/{self.samples_total:3d}"
             )
 
             # Add metrics
@@ -147,7 +173,7 @@ def _print_status(self) -> None:
             # the rich formatting added in the ``task_dict`` call
             resources_dict: dict[str, str] = {}
             for model, resource in concurrency_status().items():
-                resources_dict[model] = f"{resource[0]}/{resource[1]}"
+                resources_dict[model] = f"{resource[0]:2d}/{resource[1]:2d}"
             resources = "".join(
                 [f"{key}: {value}" for key, value in resources_dict.items()]
             )
@@ -166,12 +192,12 @@ def _print_status(self) -> None:
     def sample_complete(self, complete: int, total: int) -> None:
         self.samples_complete = complete
         self.samples_total = total
-        self._print_status()
+        self._print_status_throttled()
 
     def update_metrics(self, metrics: list[TaskDisplayMetric]) -> None:
         self.current_metrics = metrics
-        self._print_status()
+        self._print_status_throttled()
 
     def complete(self, result: TaskResult) -> None:
         self.task.result = result
-        print("Task complete.")
+        self._print_status()

src/inspect_ai/_display/rich/display.py

@@ -129,11 +129,6 @@ async def task_screen(
     @override
     @contextlib.contextmanager
     def task(self, profile: TaskProfile) -> Iterator[TaskDisplay]:
-        # if there is no ansi display than all of the below will
-        # be a no-op, so we print a simple text message for the task
-        if display_type() == "plain":
-            rich.get_console().print(task_no_ansi(profile))
-
         # for typechekcer
         if self.tasks is None:
             self.tasks = []

src/inspect_ai/_util/text.py

@@ -108,3 +108,26 @@ def str_to_float(s: str) -> float:
         exponent = 1  # Default exponent is 1 if no superscript is present
 
     return base**exponent
+
+
+def truncate(text: str, length: int, overflow: str = "...", pad: bool = True) -> str:
+    """
+    Truncate text to specified length with optional padding and overflow indicator.
+
+    Args:
+        text (str): Text to truncate
+        length (int): Maximum length including overflow indicator
+        overflow (str): String to indicate truncation (defaults to '...')
+        pad (bool): Whether to pad the result to full length (defaults to padding)
+
+    Returns:
+        Truncated string, padded if requested
+
+    """
+    if len(text) <= length:
+        return text + (" " * (length - len(text))) if pad else text
+
+    overflow_length = len(overflow)
+    truncated = text[: length - overflow_length] + overflow
+
+    return truncated

tests/util/test_truncate.py

@@ -0,0 +1,21 @@
+from inspect_ai._util.text import truncate
+
+
+def test_basic_truncation():
+    assert truncate("Hello World!", 8) == "Hello..."
+    assert truncate("Short", 8) == "Short   "
+
+
+def test_custom_overflow():
+    assert truncate("Hello World!", 8, overflow=">>") == "Hello >>"
+    assert truncate("Testing", 5, overflow="~") == "Test~"
+
+
+def test_no_padding():
+    assert truncate("Hi", 8, pad=False) == "Hi"
+    assert truncate("Hello World!", 8, pad=False) == "Hello..."
+
+
+def test_exact_length():
+    assert truncate("12345678", 8) == "12345678"
+    assert truncate("1234", 4, pad=False) == "1234"