Refresh branch #4116
Merged
Refresh branch #4116
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This change addresses some feedback we've been getting that people who report issues don't know what the Developer Tools console is or how to access it. The "output" can also be unclear but is a more nuanced topic so I've linked it to our troubleshooting page. ### QA Notes N/A, issue template change only
Added an integration test for the connections pane. At this level of abstraction, we are not testing individual clicks (like in the smoke tests), but testing the internal logic in the extension that is used by the TreeView API to render the values. --------- Signed-off-by: Daniel Falbel <[email protected]> Co-authored-by: Jonathan <[email protected]>
… row data frame from a Parquet file and verifies that three rows in the file (the first, the middle, and the last) render correctly in Pandas, Polars, and R. (#3947) This PR adds a Data Explorer smoke test that loads a 100 column x 100 row data frame from a Parquet file and verifies that three rows in the file (the first, the middle, and the last) render correctly in Pandas, Polars, and R. The data for this smoke test is loaded from the `qa-example-content` repo.
Merges upstream VS Code 1.91 into Positron. Co-authored-by: positron-bot[bot] <173392469+positron-bot[bot]@users.noreply.github.com>
…owerShell (#3909) Addresses #3816 by using `vscode.ProcessExecution` instead of `vscode.ShellExecution`, to bypass some tricky quoting issues presented by PowerShell, without doing our own quoting or writing (much) OS- or shell-specific code. The problem is that we want to execute a code snippet like `rmarkdown::render("whatever.Rmd")` (where `"whatever.Rmd"` needs to be interpolated in, which is really neither here nor there). So you need quotes around this file path. But using `ShellExecution` you also need quotes around the whole `rmarkdown::render()` snippet; let's say these are single quotes. And, when executed in PowerShell, the entire command _also_ gets double quoted, which then changes the context for the quotes in the code snippet, leading to much misery: ``` The terminal process "C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe -Command & 'C:\Program Files\R\R-4.3.3\bin\x64\R.exe' -e 'rmarkdown::render("d:\Users\jenny\rmd-render- fun\test.Rmd")'" terminated with exit code: 1. ``` By the time you get to R, the double quotes around the file path have gone missing (since I captured this error, I have changed the path separators here as well; but also that is not the problem): ``` > rmarkdown::render(d:\Users\jenny\rmd-render-fun\test.Rmd) Error: unexpected symbol in "rmarkdown::render(d:\Users" Execution halted ``` I tried all sorts of things, trying to stick with `vscode.ShellExecution` and `vscode.ShellQuoting` but never succeeded in finding a combination of argument strings and quoting that worked on Windows (PowerShell) and macOS. I became increasingly convinced, similar to this issue microsoft/vscode#187661, that it might be impossible. ### QA Notes Exercise these commands on multiple OSes: * *R: Render Document with R Markdown*, find this in the command palette * *R: Source R File*, find this in the command palette or use the "play" button Here's a repo I used for experimentation while working on this PR: https://github.com/jennybc/rmd-render-fun. You could grab it with ```r usethis::create_from_github("https://github.com/jennybc/rmd-render-fun", destdir = "~/tmp") ``` Substitute your favorite destination directory for experiments in `destdir`. This folder sports 2 `.Rmd` files, with and without spaces in the filename, and likewise for `.R`.
This change fixes an annoying issue in Windows dev environments that can cause Visual Studio to open during the incremental build. The problem is that Gulp is invoking `show-version.js`. On Unix-alikes, the shebang comment in this file causes it to run under Node, but on Windows, invoking the file opens it in the default editor for `.js` files. The fix is to use Node explicitly instead of relying on the shebang. ### QA Notes N/A, build change only.
Addresses #3752. I've also given all of our `@keyframes` unique names to avoid unintentionally overwriting keyframes across components, which broke debouncing the activity item to begin with. I also made the activity item indicator animation timing the same as the interrupt button, which I think feels better. Lastly, I added the same fade in to the action bar separator. Before this PR, it would show immediately even though the interrupt button next to it faded in. https://github.com/user-attachments/assets/1c24e135-1910-4f0d-907d-06dab8dbf142 #### QA Notes It's worth playing around with different Positron components that animate to make sure the renames didn't break anything.
Addresses #2170. I also noticed a warning in the console about leaked listeners for the `onDidRenderThumbnail` event so cleaned that up too.
Tests defaults, setting file name and format. Added a couple of helpers for the File Explorer and modals --------- Co-authored-by: Jon Vanausdeln <[email protected]>
There were some missing TestRail Test IDs, so they weren't being reported in TestRail. ### QA Notes They now show up in TestRail
Basic smoke test of the output pane. For R and Python, enters a bad line of code and then asserts that it sees a corresponding log in the output pane. ### QA Notes All smoke tests should pass.
Address the image referencing problem revealed from #3996 Reference SVG in style sheet so optimizer handles it correctly
This is a simple console user input test. In both Python and R, the
"user" is prompted to enter their name and "Hello ${name}" is echoed
back. The test waits until the echo is received in the console output to
determine pass/fail. This PR also adds in a pasteToConsole function that
simulates a user pasting code as opposed to entering code via a command
(to use the console itself more).
### QA Notes
All smoke tests should pass!
…rnel tests (#3951) ## Description - Addresses the Python Project ipykernel tests part of #3879 ### Summary - Reorganizes the automation class for Project Wizard - Each step now has a class which is contained in the Project Wizard class - Selectors are scoped to the class constructors whenever possible - Opting for usage of the `Locator` class when reasonable (sometimes waiting for elements and clicking on them still works better with the BaseElement class) - Adds a couple ipykernel project wizard tests - Reorganize Project Wizard smoke tests so all of the Project Wizard tests can be run with one `only` - Modify the frontend components for the Project Wizard so that wizard steps and substeps have an easily referenceable ID, making constructing selectors easier <img width="1470" alt="proj_wiz_set_up_py_env_ids" src="https://github.com/posit-dev/positron/assets/25834218/c6b9028e-3a5a-456f-8062-cc59843fced8"> - Add option to `PositronPythonFixtures` to install ipykernel as part of selecting the interpreter - Add new method `popupCurrentlyOpen()` to `PositronPopups` to check if a popup is currently open - Support retrieving the active interpreter path from the Interpreter Dropdown - Modify `StartInterpreter` methods to support returning the interpreter path - NOTE: `StartInterpreter` methods don't seem to be used elsewhere, so we should add some tests to confirm the class methods are working. Possibly in test/smoke/src/areas/positron/top-action-bar/top-action-bar.test.ts? ### 🆕 Smoke Tests #### Python Project with existing interpreter: - With ipykernel already installed - With ipykernel not already installed ## QA Notes Please add new test IDs and validate on Windows. Thank you :) --------- Co-authored-by: Jon Vanausdeln <[email protected]>
) ## Description - Addresses the Python Project conda env part of #3879 ### Summary - Add installation of conda to CI script `.github/workflows/positron-full-test.yml` - Add handling for selecting the environment provider in the Project Wizard Python env provider dropdown - Add new smoke test for creating a new Python project with a new Conda environment in the Project Wizard ### 🆕 Smoke Test #### Python Project with new environment: - Create a new Conda environment ### QA Notes If you're testing this locally, make sure you have Conda installed! Either of the following are lighter-weight options than installing Anaconda: - miniforge3: https://github.com/conda-forge/miniforge?tab=readme-ov-file#install - miniconda: https://docs.anaconda.com/miniconda/#quick-command-line-install Please add new test ID and verify this works on Windows too. Thank you!
## Description - Addresses the `git init` part of #3879 ### Summary - Add 🆕 smoke test for creating a new Python Project with `git init` selected ### QA Notes Please add new test ID and verify this works on Windows too. Thank you!
Another step toward #3276. This PR moves the rendering of our notebook output webview service to inside the webview, giving the notebook renderer access to features which previously did not work, including (from the code): > Some features known to be NYI: > - Message passing between the renderer and the host (RenderContext) > - Extending another renderer (RenderContext) > - State management (RenderContext) > - Raw Uint8Array data and blobs We do this by listening for a new `positronRender` message inside the webview. I also changed our `render_complete` message to `positronRenderComplete` and added it to the known message types. I also made some changes to how we determine the required preloads and resource roots to bring it more in line with the backlayer webview: 1. **Preloads:** We get preloads by view type instead of by extension. 2. **Resource roots:** 1. Removed the renderer's extension folder. 2. Added each renderer's parent folder. 3. Added each preload's parent folder. 4. Added each preload's specified `localResourceRoots`. ### QA Notes This should not break any existing widgets/plots/HTML outputs. Here are some tests for different code paths: 1. R html content: `gt::gt(mtcars)` 2. Python notebook renderer: `import plotly.express as px; px.bar(x=["a", "b", "c"], y=[1, 3, 2])` 3. R notebook renderer: `library(plotly); plot_ly(data = iris, x = ~Sepal.Length, y = ~Petal.Length)` 4. Python custom ipywidgets webview: `from ipyleaflet import Map; display(Map(center=(52.204793, 360.121558), zoom=12))`
Very much like the existing variables test except for notebook sessions. Tests Python and R notebooks' variables. ### QA Notes All smoke tests should pass.
Just a minor typo fix and another cleanup. The polars test was using the R fixture. ### QA Notes All smoke tests should pass.
<!-- Describe briefly what problem this pull request resolves, or what new feature it introduces. Include screenshots of any new or altered UI. Link to any GitHub issues but avoid "magic" keywords that will automatically close the issue. If there are any details about your approach that are unintuitive or you want to draw attention to, please describe them here. --> This PR improves the logging for the Data Explorer 100x100 test by outputting row and column information in the `expect` test case. ### QA Notes <!-- Add additional information for QA on how to validate the change, paying special attention to the level of risk, adjacent areas that could be affected by the change, and any important contextual information not present in the linked issues. -->
Just added retries to the console input test so it is less likely to experience an issue with characters being dropped. Also removed a dead function that I thought I had previously deleted. ### QA Notes All smoke tests (besides ones with open product bugs) should pass.
### Intent Add TestRail Test IDs to 100x100 dataframe tests. ### QA Notes Now will show up in TestRail runs
Addresses #3453. There is some refactoring here to improve code reuse with the pandas implementation. ### QA Notes Enable experimental data explorer features and try expanding the column profiles for various data types.
See: #2860 This PR differentiates special values in Data Explorer by rendering them at 65% opacity. For example, execute the following code: ``` import pandas as pd df = pd.DataFrame([['ant', 'bee', None, float('nan')], ['None', None, 'fly', float('nan')]]) ``` And open the resulting data frame in Data Explorer. And you will notice that the cells containing `None` and `float('nan')` are rendered at 65% opacity to differentiate them from normal, non-null values (like 'ant', 'bee', 'None', and 'fly'). <img width="1240" alt="image" src="https://github.com/user-attachments/assets/c4b2a9e5-cadb-40a0-9e0c-0189142f02f5"> This applies to data cell values of the following kinds: NULL NA NaN NotATime None INFINITY NEG_INFINITY UNKNOWN ### QA Notes I'd like to generate some example of these values in the `qa-example-content` repo. I will do this soon.
### Intent Due to #4070 this test needs to be skipped for run stability ### QA Notes Test should be enabled once issue is resolved.
…ates (#4066) ## Description - Addresses remaining pieces of #3879 - Adds Conda installation steps documentation to smoke test README - Fixes and reorders checks in `pythonEnvironmentStep.tsx` to resolve timing issue when selecting Conda as the env provider while interpreter info is still loading - Improves timing handling for clicking project wizard navigation buttons (back, next, create, cancel) - Adds wait for python project wizard dropdown items to load before interacting with them ### QA Notes - The Conda dropdown timing issue should be resolved
Addresses #4018 - Add basic support for HTML output, aka things like pandas dataframe tables. - Only show a single output element per cell output. - A cell output can have multiple data bundles of `mime` + `data`, previously we would render all of them. Now we just choose one to render, like other notebook renderers. - Currently, this is done via a fairly arbitrary ordering of "interesting" output types, but in the future we may be more structured about it. <img width="1199" alt="image" src="https://github.com/user-attachments/assets/3e9b7e89-855c-48c0-82a3-dcf6ebe4b823"> Also includes a large shuffling of the positron instance code to better organize it according to how other code in positron is organized with private and public fields separated. ### QA Notes Notebooks should continue to work as normal, just simple html outputs will now actually render. E.g. the above table. Here's the code to generate one: ```python import pandas as pd import numpy as np # Generate random data pd.DataFrame({ 'A': np.random.randint(0, 100, size=20), 'B': np.random.normal(0, 1, size=20), 'C': np.random.choice(['X', 'Y', 'Z'], size=20), 'D': np.random.uniform(0, 1, size=20) }) ```
Prior to this PR, the parser would be called for almost any interaction with the text (e.g. moving the cursor around). To fix this, cell parsing was moved from the `CellManager` to `DocumentManager`. A map of document managers are maintained for each document.
<!-- Describe briefly what problem this pull request resolves, or what new feature it introduces. Include screenshots of any new or altered UI. Link to any GitHub issues but avoid "magic" keywords that will automatically close the issue. If there are any details about your approach that are unintuitive or you want to draw attention to, please describe them here. --> This PR addresses #705 (prompt flickering) by adjusting the prompt to be spaces immediately after code is executed. This prevents the prompt from being displayed while code is executing, which was the cause of the flickering. Prior to this PR, the Console prompt would noticeably flicker about 30% of the time. The sequence of distinct UI frames that lead to this flicking are: 1: User enters some code and executes it: <img width="538" alt="image" src="https://github.com/user-attachments/assets/33396be9-0f42-41c9-954b-7a985c4c1c60"> 2: The `ConsoleInput` component gets the code from the `CodeEditorWidget`, creates the provisional `ActivityItemInput` component, then clears the `CodeEditorWidget`'s model. This results in the following frame because the `CodeEditorWidget` renders the standard prompt too soon: <img width="528" alt="image" src="https://github.com/user-attachments/assets/505f762b-76df-42b5-bfaf-a4aa2e48ead6"> This is what causes the flicker because once the code executes, the result of the execution is displayed between the `ActivityItemInput` and the `CodeEditorWidget`: <img width="528" alt="image" src="https://github.com/user-attachments/assets/6f1e36c1-0aa7-445f-8d1b-e860a619197e"> ### QA Notes <!-- Add additional information for QA on how to validate the change, paying special attention to the level of risk, adjacent areas that could be affected by the change, and any important contextual information not present in the linked issues. -->
The latest release has been finalized at https://github.com/microsoft/vscode-python/tree/release/2024.10. The new env locator is housed at https://github.com/microsoft/python-environment-tools. The least invasive thing to do for now would be to host builds on GitHub and download then at yarn time like we do for Ark. There are also a handful of other changes having to do with the terminal REPL in vscode-python, so we should investigate those to make sure I didn't break anything when resolving the merge conflicts. --------- Co-authored-by: positron-bot[bot] <173392469+positron-bot[bot]@users.noreply.github.com> Co-authored-by: Jonathan McPherson <[email protected]>
Minor selector change. Looks like the class runtime-items went away in the console. ### QA Notes All smoke tests pass
…0.x to 1.2.x (#4105) polars 1.2.x contains various behavior changes that had to be worked around, causing errors like #4070. I reported some of the issues as bugs upstream in polars (they may or may not be considered bugs in polars -- they look like bugs to me or at minimum unpleasant sharp edges). I was able to get the unit tests working on both 0.20.x and 1.2.x but we have a blind spot in our CI right now in that we do not test > polars 1.0 yet. ### QA Notes These failures arose because the polars version in the smoke test CI setup wasn't pinned -- it would have been sufficient to run the unit test suite for both 0.20.x and 1.2.x to find these issues.
Addresses #2829 This is a brute-force solution to the confusing actions menu in altair plots. We hide the menu by selecting it via css (with the path `.vega-embed details[title="Click to view actions"]`) Also adds a static property to the `PositronNotebookOutputWebviewService` service to store these addon css properties in case we need to do more of them in the future for other plotting libraries etc.. Ultimately the solution should probably be more robust then this and involve stubbing in functions for opening new pages etc that the actions menu tries to use, but that's a lot more work and this gets the job done at the moment. (See #4087) ### Before _Altair plot with action menu that is unusable_  ### After _Altair plot sans actions menu_ <img width="341" alt="image" src="https://github.com/user-attachments/assets/a7d00b80-4f19-4b3a-9851-7c29a6e48689"> <!-- Thank you for submitting a pull request. If this is your first pull request you can find information about contributing here: * https://github.com/posit-dev/positron/blob/main/CONTRIBUTING.md We recommend synchronizing your branch with the latest changes in the main branch by either pulling or rebasing. --> <!-- Describe briefly what problem this pull request resolves, or what new feature it introduces. Include screenshots of any new or altered UI. Link to any GitHub issues but avoid "magic" keywords that will automatically close the issue. If there are any details about your approach that are unintuitive or you want to draw attention to, please describe them here. --> ### QA Notes <!-- Add additional information for QA on how to validate the change, paying special attention to the level of risk, adjacent areas that could be affected by the change, and any important contextual information not present in the linked issues. -->
…for rendering html in positron notebooks (#4078) This PR has three main parts to it as told by the (first) three commits: 1. Update the `IPositronNotebookWebviewService` to expose the `.createRawHtmlOutput()` method and allow that method to create standard `WebviewElement`s in addition to `OverlayWebview`s. (Also adds the ability to create the webview without an active runtime due to runtime requirements preventing notebooks from showing html content until the kernel is connected.) 2. Add knowledge of the current language runtime attached to a notebook to the `PositronNotebookInstance` class. (Also uses that knowledge to make the kernel badge more accurate) 3. Replace the existing rendering of HTML content in notebooks with the `IPositronNotebookWebviewService.createRawHtmlOutput()` and all the goodies that we get (and will get in the future) from that. ### QA Notes This should be a simple under-the-hood swap; nothing should look or work differently.
Updates Ark to 0.1.118; for #3846.
jmcphers
approved these changes
Jul 22, 2024
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
QA Notes