Add ability to run in fully automated mode

.github/dependabot.yml

@@ -0,0 +1,15 @@
+# To get started with Dependabot version updates, you'll need to specify which
+# package ecosystems to update and where the package manifests are located.
+# Please see the documentation for all configuration options:
+# https://docs.github.com/github/administering-a-repository/configuration-options-for-dependency-updates
+
+version: 2
+updates:
+  - package-ecosystem: "github-actions" # See documentation for possible values
+    directory: ".github/workflows" # Location of package manifests
+    schedule:
+      interval: "weekly"
+    groups:
+      actions:
+        patterns:
+          - "*"

.github/workflows/main.yml

@@ -0,0 +1,22 @@
+name: Automated test
+
+on:
+  workflow_dispatch:
+  pull_request:
+  push:
+
+jobs:
+  test:
+    uses: OpenAstronomy/github-actions-workflows/.github/workflows/tox.yml@v1
+    with:
+      envs: |
+        - linux: py310-test
+        - linux: py311-test
+        - linux: py312-test
+        - linux: py313-test
+        - macos: py310-test
+        - macos: py311-test
+        - macos: py312-test
+        - macos: py313-test
+        - windows: py312-test
+        - windows: py313-test

.gitignore

@@ -4,3 +4,4 @@ __pycache__
 dist
 build
 .ipynb_checkpoints
+__pycache__

.pre-commit-config.yaml

@@ -0,0 +1,29 @@
+ci:
+  autofix_prs: false
+  autoupdate_schedule: 'monthly'
+
+repos:
+  - repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v4.5.0
+    hooks:
+      - id: check-added-large-files
+        args: ["--enforce-all", "--maxkb=300"]
+      - id: check-case-conflict
+      - id: check-json
+      - id: check-merge-conflict
+      - id: check-symlinks
+      - id: check-toml
+      - id: check-xml
+      - id: check-yaml
+        exclude: ".*(.github.*)$"
+      - id: detect-private-key
+      - id: end-of-file-fixer
+        exclude: ".*(data.*|extern.*|licenses.*|_static.*|_parsetab.py)$"
+      - id: trailing-whitespace
+
+  - repo: https://github.com/astral-sh/ruff-pre-commit
+    rev: "v0.3.4"
+    hooks:
+      - id: ruff
+        args: ["--fix", "--show-fixes"]
+      - id: ruff-format

README.md

@@ -1,28 +1,53 @@
 This repository contains an experimental utility to monitor the visual output of
 cells from Jupyter notebooks.
 
-## Requirements
-
-On the machine being used to run the ``monitor_cells.py``:
+## Installing
 
-* [numpy](https://numpy.org)
-* [click](https://click.palletsprojects.com/en/stable/)
-* [pillow](https://python-pillow.org/)
-* [playwright](https://pypi.org/project/playwright/)
+To install, check out this repository and:
 
-On the Jupyter Lab server, optionally (but recommended):
+    pip install -e .
 
-* [jupyter-collaboration](https://github.com/jupyterlab/jupyter-collaboration)
+Python 3.10 or later is supported (Python 3.12 or later on Windows).
 
-If this is the first time using playwright, you will need to run::
+If this is the first time using playwright, you will also need to run:
 
     playwright install firefox
 
-## Installing
+## Quick start
 
-To install, check out this repository and:
+First, write one or more blocks of code you want to benchmark each in a cell. In
+addition, as early as possible in the notebook, make sure you set the border
+color on any ipywidget layout you want to record:
 
-    pip install -e .
+    widget.layout.border = '1px solid rgb(143, 56, 3)'
+
+The R and G values should be kept as (143, 56), and the B color should be unique for each widget and be a value between 0 and 255 (inclusive).
+
+Then, to run the notebook and monitor the changes in widget output, run:
+
+    jupyter-output-monitor --notebook mynotebook.ipynb
+
+Where ``mynotebook.ipynb`` is the name of your notebook. By default, this will
+open a window showing you what is happening, but you can also pass ``--headless``
+to run in headless mode.
+
+## Using this on a remote Jupyter Lab instance
+
+If you want to test this on an existing Jupyter Lab instance, including
+remote ones, you can use ``--url`` instead of ``--notebook``:
+
+    jupyter-output-monitor http://localhost:8987/lab/tree/notebook.ipynb?token=7bb9a...
+
+Note that the URL should include the path to the notebook, and will likely
+require the token too.
+
+You should make sure that all output cells in the notebook have been cleared
+before running the above command, and that the widget border color has been
+set as mention in the **Quick start** guide above.
+
+If you make use of the [jupyter-collaboration](https://github.com/jupyterlab/jupyter-collaboration) plugin on the Jupyter Lab server, you will be able to
+more easily e.g. clear the output between runs and edit the notebook in
+between runs of ``jupyter-output-monitor``.
 
 ## How this works
 
@@ -83,46 +108,8 @@ and if using jdaviz:
 To stop recording output for a given cell, you can set the border attribute to
 ``''``.
 
-## Headless vs non-headless mode
-
-By default, the script will open up a window and show what it is doing. It will
-also wait until it detects any input cells before proceeding. This then gives
-you the opportunity to enter any required passwords, and open the correct
-notebook. However, note that if Jupyter Lab opens up with a different notebook
-to the one you want by default, it will start executing that one! It's also
-better if the notebook starts off with output cells cleared, otherwise the script
-may start taking screenshots straight away.
-
-The easiest way to ensure that the correct notebook gets executed and that it
-has had its output cells cleared is to make use of the
-[jupyter-collaboration](https://github.com/jupyterlab/jupyter-collaboration)
-plugin. With this plugin installed, you can open Jupyter Lab in a regular browser window,
-and set it up so that the correct notebook is open by default and has its cells cleared,
-and you can then launch the monitoring script. In fact, if you do this you can then
-also run the script in headless mode since you know it should be doing the right thing.
-
-One final note is that to avoid any jumping up and down of the notebook during
-execution, the window opened by the script has a very large height so that the
-full notebook fits inside the window without scrolling.
-
-## How to use
-
-* Assuming you have installed
-  [jupyter-collaboration](https://github.com/jupyterlab/jupyter-collaboration),
-  start up Jupyter Lab instance on a regular browser and go to the notebook you
-  want to profile.
-* If not already done, write one or more blocks of code you want to benchmark
-  each in a cell. In addition, as early as possible in the notebook, make sure
-  you set the border color on any ipywidget layout you want to record.
-* Make sure the notebook you want to profile is the main one opened and that
-  you have cleared any output cells.
-* Run the main command in this package, specifying the URL to connect to for Jupyter Lab, e.g.:
-
-      jupyter-output-monitor http://localhost:8987
-
 ## Settings
 
-
 ### Headless
 
 To run in headless mode, include ``--headless``

jupyter_output_monitor/__init__.py

@@ -1,2 +1,4 @@
 from ._monitor import monitor
 from ._version import __version__
+
+__all__ = ["monitor", "__version__"]

jupyter_output_monitor/__main__.py

@@ -0,0 +1,4 @@
+from ._monitor import monitor
+
+if __name__ == "__main__":
+    monitor()

jupyter_output_monitor/_convert.py

@@ -0,0 +1,107 @@
+# Experimental command to convert a notebook to a script that tests widget
+# output with solara but without launching a Jupyter Lab instance. Not yet
+# exposed via the public API.
+
+import os
+from textwrap import indent
+
+import click
+import nbformat
+
+
+def remove_magics(source):
+    lines = [line for line in source.splitlines() if not line.startswith("%")]
+    return os.linesep.join(lines)
+
+
+def remove_excludes(source):
+    lines = [
+        line for line in source.splitlines() if not line.strip().endswith("# EXCLUDE")
+    ]
+    return os.linesep.join(lines)
+
+
+HEADER = """
+import time
+import solara
+import playwright
+import pytest_playwright
+from IPython.display import display
+
+def watch_screenshot(widget):
+    display_start = time.time()
+    last_change_time = display_start
+    last_screenshot = None
+    while time.time() - display_start < 5:
+        screenshot_bytes = widget.screenshot()
+        if screenshot_bytes != last_screenshot:
+            last_screenshot = screenshot_bytes
+            last_change_time = time.time()
+    return last_screenshot, last_change_time - display_start
+
+def test_main(page_session, solara_test):
+
+"""
+
+DISPLAY_CODE = """
+object_to_capture.add_class("test-object")
+display(object_to_capture)
+captured_object = page_session.locator(".test-object")
+captured_object.wait_for()
+"""
+
+PROFILING_CODE = """
+last_screenshot, time_elapsed = watch_screenshot(captured_object)
+print(f"Extra time waiting for display to update: {time_elapsed:.2f}s")
+"""
+
+
+@click.command()
+@click.argument("input_notebook")
+@click.argument("output_script")
+def convert(input_notebook, output_script):
+    nb = nbformat.read(input_notebook, as_version=4)
+
+    with open(output_script, "w") as f:
+        f.write(HEADER)
+
+        captured = False
+
+        for icell, cell in enumerate(nb["cells"]):
+            if cell.cell_type == "markdown":
+                f.write(indent(cell.source, "    # ") + "\n\n")
+            elif cell.cell_type == "code":
+                if cell.source.strip() == "":
+                    continue
+
+                lines = cell.source.splitlines()
+
+                new_lines = []
+
+                new_lines.append("cell_start = time.time()\n\n")
+
+                for line in lines:
+                    if line.startswith("%") or line.strip().endswith("# EXCLUDE"):
+                        continue
+                    elif line.endswith("# SCREENSHOT"):
+                        new_lines.append("object_to_capture = " + line)
+                        new_lines.extend(DISPLAY_CODE.splitlines())
+                        captured = True
+                    else:
+                        new_lines.append(line)
+
+                new_lines.append("cell_end = time.time()\n")
+                new_lines.append(
+                    f'print(f"Cell {icell:2d} Python code executed in           {{cell_end - cell_start:.2f}}s")',
+                )
+
+                if captured:
+                    new_lines.extend(PROFILING_CODE.splitlines())
+
+                source = os.linesep.join(new_lines)
+
+                f.write(indent(source, "    ") + "\n\n")
+
+
+if __name__ == "__main__":
+    convert()