juba / pyobsplot Goto Github PK

View Code? Open in Web Editor NEW

184.0 6.0 8.0 92.74 MB

Observable Plot in Jupyter notebooks and Quarto documents

Home Page: https://juba.github.io/pyobsplot/

License: MIT License

JavaScript 39.92% CSS 0.06% Python 2.00% Jupyter Notebook 0.80% HTML 57.12% Typst 0.09%

anywidget jupyter jupyter-widget observable-plot python quarto

pyobsplot's People

Contributors

Stargazers

Watchers

Forkers

tomkallo genostack jaedukseo vladschkolnik harrylojames wirhabenzeit mhuebert gauravcodepro

pyobsplot's Issues

Interactions and static tips not possible with `jsdom` renderer

For the moment "static" plots generated as HTML or SVG by the jsdom renderer cannot have pointer interactions.

Make shorthand definition possible

Hi,

In line with having a syntax as close as possible to javascript it would nice to be able to use the shorthand syntax as seen here.

import pandas as pd
from pyobsplot import Plot

df = pd.DataFrame({'Date': {0: '1880-01-01', 1: '1880-02-01'}, 'Anomaly': {0: -0.3, 1: -0.21}})
Plot.lineY(df, {"x": "Date", "y": "Anomaly"}).plot()

AttributeError                            Traceback (most recent call last)
Cell In[15], line 5
      2 from pyobsplot import Plot
      4 df = pd.DataFrame({'Date': {0: '1880-01-01', 1: '1880-02-01'}, 'Anomaly': {0: -0.3, 1: -0.21}})
----> 5 Plot.lineY(df, {"x": "Date", "y": "Anomaly"}).plot()

AttributeError: 'dict' object has no attribute 'plot'

Not sure how feasible this is and not a dealbreaker for me but thought it was worth noting in case it wasn't on the radar.

Many thanks,

Harry

Mixing renderers in Jupyter lab moves all outputs to widgets

Unexpected scale for pandas datetime column

datetime64[ns] objects aren't converted to javascript date objects. Presume this is intended as per the section on datetimes. Would be great if it were possible!

import pandas as pd
from pyobsplot import Plot

df = pd.DataFrame({'full_date_time': {0: pd.Timestamp('2022-12-01 00:00:00'),
  1: pd.Timestamp('2022-12-01 01:00:00')},
 'value': {0: 20.0, 1: 18.0}})

Plot.plot({"marks":[Plot.dot(df, {"x":"full_date_time", "y": "value"})]})

Plots with legends don't work in Quarto in formats other than HTML

As plots with legends generate HTML output instead of SVG, they are not included in Quarto outputs for formats other than HTML.

Syntax error in documentation

A comma is missing in below lines of code after color object.

Plot.plot(
    {
        color: {legend: true} #comma missing here
        grid: false,
        marks: [Plot.dot(data, {x: "x", y: "y", fill: "type", r: 5})]
    }
)

op({
    "color": {"legend": True} #missig comma here as well
    "grid": False,
    "marks": [Plot.dot(data, {"x": "x", "y": "y", "fill": "type", "r": 5})],
})

https://juba.github.io/pyobsplot/usage.html

"Exception not rethrown" in pytest

Look what is the cause of the FATAL: Exception not rethrown messages in pytest.

question about ObservableHQ

Hi, I have a few questions related to ObservableHQ and how you've built pyobsplot:

Is it possible to import an ObservableHQ notebook into anywidget?
Are you importing the observable runtime in your anywidget front-end code?

I'm curious if I could reproduce something similar to this widget using notebooks I've made on ObservableHQ.

Renderer, format and path API

As for the general question, it seems a bit problematic that there is no one-to-one correspondence between format and path-extension. What I mean is that:

there is no format="pdf" (I assume because displaying PDFs in something like Jupyter notebooks seems problematic)

there are two formats for extension ".html", namely "widget", and "html".

Without knowing the internals it seems weird that
Plot.plot({}, format="png", path="out.svg") => ok
Plot.plot({}, format="widget", path="out.svg") => error.
I first thought that it would be best to completely decouple format and path: format dictates the display format, path dictates the export format. But this does not account for the html case...

Originally posted by @wirhabenzeit in #32 (comment)

Widget renderer output lost when notebook is closed in VSCode

VSCode doesn't save widget output state when the file is closed, so all cells have to be recomputed to see the output.

See: microsoft/vscode-jupyter#4404

Use type declarations to improve IDE support

See if it would be possible to use the typescript type declarations introduced in Plot 0.6.5 to improve autocompletion and IDE support.

Combining tree and text (as table)?

Hello,

I am trying to replicate the visualization here, where a table is displayed right next to a tree.

https://observablehq.com/@d3/indented-tree

Is this out of scope for your library? The key code from that script is:

  for (const {label, value, format, x} of columns) {
    svg.append("text")
        .attr("dy", "0.32em")
        .attr("y", -nodeSize)
        .attr("x", x)
        .attr("text-anchor", "end")
        .attr("font-weight", "bold")
        .text(label);

    node.append("text")
        .attr("dy", "0.32em")
        .attr("x", x)
        .attr("text-anchor", "end")
        .attr("fill", d => d.children ? null : "#555")
      .data(root.copy().sum(value).descendants())
        .text(d => format(d.value, d));
  }

I've tried replicating this with Plot.text but it appears combining these marks disrupts how inset handles the Plot.tree. Here is the code for the tree alone.

indent = js("""
() => {
  return (root) => {
    root.eachBefore((node, i) => {
      node.x = i;
      node.y = node.depth;
    });
  };
}
""")

Plot.plot({
    'axis': None,
    'inset': 10,
    'insetRight': 800,
    'round': True,
    'width': 1000,
    'height': 6000,
    'marks': [
        Plot.tree(path_utility, {
            'path': 'Path',
            'strokeWidth': 1,
            'curve': "step-before",
            'treeLayout': indent,
            'treeSort': "node:height",
            'delimiter': '^',
        }),
    ]
})

I don't actually need all the dynamic logic for calculating sizes. I can precompute all the numbers. Just want to align the table with the tree.

Thanks again!

Update documentation with new PNG and PDF outputs

Update Getting started and Usage pages.

Investigate topojson "frame" issue

When converting topojson data to geojson in Python there seems to be a global "frame" object that hides everything when filled.

error cb.apply is not a function

I am trying to use the jsdom renderer but I get the following error (latest pyobsplot==0.4.2):

ValueError: Server not started: npm ERR! cb.apply is not a function

npm ERR! A complete log of this run can be found in:
npm ERR!     [/home/jarvis/.npm/_logs/2024-05-03T08_53_13_050Z-debug.log](https://jupyter.imaging.ms.uhbs.ch/home/jarvis/.npm/_logs/2024-05-03T08_53_13_050Z-debug.log)
Install for [ '[email protected]' ] failed with code 1

0 info it worked if it ends with ok
1 verbose cli [
1 verbose cli   '/home/jarvis/.nvm/versions/node/v18.17.1/bin/node',
1 verbose cli   '/home/jarvis/.nvm/versions/node/v18.17.1/lib/node_modules/npx/node_modules/npm/bin/npm-cli.js',
1 verbose cli   'install',
1 verbose cli   '[email protected]',
1 verbose cli   '--global',
1 verbose cli   '--prefix',
1 verbose cli   '/home/jarvis/.npm/_npx/3055885',
1 verbose cli   '--loglevel',
1 verbose cli   'error',
1 verbose cli   '--json'
1 verbose cli ]
2 info using [email protected]
3 info using [email protected]
4 verbose npm-session 70f0ac5533eebc63
5 silly install loadCurrentTree
6 silly install readGlobalPackageData
7 http fetch GET 304 https://registry.npmjs.org/pyobsplot 884ms (from cache)
8 http fetch GET 200 https://registry.npmjs.org/pyobsplot/-/pyobsplot-0.4.2.tgz 7ms (from cache)
9 silly pacote version manifest for [email protected] fetched in 930ms
10 verbose stack TypeError: cb.apply is not a function
10 verbose stack     at /home/jarvis/.nvm/versions/node/v18.17.1/lib/node_modules/npx/node_modules/npm/node_modules/graceful-fs/polyfills.js:287:18
10 verbose stack     at FSReqCallback.oncomplete (node:fs:211:5)
11 verbose cwd /var/www/jarvis/Auswertung/ge-migration
12 verbose Linux 5.10.0-25-amd64
13 verbose argv "/home/jarvis/.nvm/versions/node/v18.17.1/bin/node" "/home/jarvis/.nvm/versions/node/v18.17.1/lib/node_modules/npx/node_modules/npm/bin/npm-cli.js" "install" "[email protected]" "--
global" "--prefix" "/home/jarvis/.npm/_npx/3055885" "--loglevel" "error" "--json"
14 verbose node v18.17.1
15 verbose npm  v5.1.0
16 error cb.apply is not a function
17 verbose exit [ 1, true ]

Thanks for any help!

'charmap' codec can't decode byte 0x90 in position 780112: character maps to <undefined>

Hi,

Running into the error 'UnicodeDecodeError: 'charmap' codec can't decode byte 0x90 in position 780112: character maps to ' when using the jupyter widget renderer. It doesn't happen when using the jsdom renderer.

from pyobsplot import Plot
import polars as pl

data = pl.DataFrame(
    {
        "x": [1, 5, 2, 4, 6, 2, 4],
        "y": [2, 1, 3, 4, 5, 1, 2],
        "type": ["T1", "T2", "T1", "T2", "T1", "T1", "T2"],
    }
)

Plot.plot(
    {
        "grid": True,
        "marks": [Plot.dot(data, {"x": "x", "y": "y", "fill": "type", "r": 5})],
    }
)

Here is the error in full.



UnicodeDecodeError                        Traceback (most recent call last)
Cell In[9], line 12
      2 import polars as pl
      4 data = pl.DataFrame(
      5     {
      6         "x": [1, 5, 2, 4, 6, 2, 4],
   (...)
      9     }
     10 )
---> 12 Plot.plot(
     13     {
     14         "grid": True,
     15         "marks": [Plot.dot(data, {"x": "x", "y": "y", "fill": "type", "r": 5})],
     16     }
     17 )

File ~\AppData\Local\pypoetry\Cache\virtualenvs\pyobsplot-test-JUPV2V8N-py3.11\Lib\site-packages\pyobsplot\jsmodules.py:25, in Plot.plot(*args, **kwargs)
     23 if _plot_renderer == "jsdom":
     24     op = ObsplotJsdomCreator()
---> 25 return op(*args, **kwargs)

File ~\AppData\Local\pypoetry\Cache\virtualenvs\pyobsplot-test-JUPV2V8N-py3.11\Lib\site-packages\pyobsplot\obsplot.py:141, in ObsplotWidgetCreator.__call__(self, *args, **kwargs)
    137 """
    138 Method called when an instance is called.
    139 """
    140 spec = self.get_spec(*args, **kwargs)
--> 141 return ObsplotWidget(
    142     spec, theme=self._theme, default=self._default, debug=self._debug
    143 )

File ~\AppData\Local\pypoetry\Cache\virtualenvs\pyobsplot-test-JUPV2V8N-py3.11\Lib\site-packages\pyobsplot\widget.py:44, in ObsplotWidget.__init__(self, spec, theme, default, debug)
     42 self._default = default
     43 # Init widget
---> 44 super().__init__(spec=spec, theme=theme)

File ~\AppData\Local\pypoetry\Cache\virtualenvs\pyobsplot-test-JUPV2V8N-py3.11\Lib\site-packages\anywidget\widget.py:42, in AnyWidget.__init__(self, *args, **kwargs)
     40 if hasattr(self, key) and not self.has_trait(key):
     41     value = getattr(self, key)
---> 42     anywidget_traits[key] = t.Unicode(str(value)).tag(sync=True)
     44     if isinstance(value, FileContents):
     45         value.changed.connect(
     46             lambda new_contents, key=key: setattr(self, key, new_contents)
     47         )

File ~\AppData\Local\pypoetry\Cache\virtualenvs\pyobsplot-test-JUPV2V8N-py3.11\Lib\site-packages\anywidget\_file_contents.py:92, in FileContents.__str__(self)
     90 def __str__(self) -> str:
     91     if self._contents is None:
---> 92         self._contents = self._path.read_text()
     93     return self._contents

File ~\.pyenv\pyenv-win\versions\3.11.1\Lib\pathlib.py:1059, in Path.read_text(self, encoding, errors)
   1057 encoding = io.text_encoding(encoding)
   1058 with self.open(mode='r', encoding=encoding, errors=errors) as f:
-> 1059     return f.read()

File ~\.pyenv\pyenv-win\versions\3.11.1\Lib\encodings\cp1252.py:23, in IncrementalDecoder.decode(self, input, final)
     22 def decode(self, input, final=False):
---> 23     return codecs.charmap_decode(input,self.errors,decoding_table)[0]

UnicodeDecodeError: 'charmap' codec can't decode byte 0x90 in position 780112: character maps to <undefined>

Let me know if there are any other details I can provide!

Many thanks,

Harry

Problems with faceting

Same data sources are seen as distinct, generating warning and disabling faceting by default.

Example:

import polars as pl
from pyobsplot import Obsplot, Plot

penguins = pl.read_csv("data/penguins.csv")

Obsplot({
    "facet": {"data": penguins, "x": "island"},
    "marks": [
        Plot.dot(penguins, {"x": "flipper_length_mm", "y": "body_mass_g", "fill": "species"}),
    ]
})

Implement Reference pages

Add reference pages to documentation with quartodoc.

Using custom TreeLayout?

Hello!

This is an amazing library.

I am interested in making a custom tree layout as shown here for an Indented Tree Plot. As described here: https://observablehq.com/@observablehq/plot-custom-tree-layout

indent = js("""
function indent() {
  return (root) => {
    root.eachBefore((node, i) => {
      node.y = node.depth;
      node.x = i;
    });
  };
}
""")

Plot.plot({
    'axis': None,
    'inset': 10,
    'insetRight': 120,
    'round': True,
    'width': 2000,
    'height': 3600,
    'margin': 200,
    'marks': Plot.tree([all_paths, {
        'strokeWidth': 1,
        'curve': "step-before",
        'treeLayout': indent,
    })
})

However as far as I can tell, this is just ignoring the treeLayout provided. Am I missing something? Thank you!

Integrating Observable Inputs

Hi,

Really excited to see progress being made on making observable plot available in python!

What is the feasibility of integrating Observable Inputs with this to make the following example possible?

https://observablehq.com/@observablehq/plot-density-weighted?intent=fork

Many thanks,

Harry

Add "Open in colab" button

With an example notebook.

Add !pip install pyobsplot as first cell. See if it is possible to include jsdom renderer by installing nodejs.

Plot.rectY not plotting bars

The following produces an empty chart.

import polars as pl
from pyobsplot import Plot

penguins = pl.read_csv("https://github.com/juba/pyobsplot/raw/main/doc/data/penguins.csv")

Plot.plot({
  "marks": [
    Plot.rectY(penguins, Plot.groupX({"y": "count"}, {"x": "species"})),
  ]
})

The following produces the chart as expected.

import polars as pl
from pyobsplot import Plot

penguins = pl.read_csv("https://github.com/juba/pyobsplot/raw/main/doc/data/penguins.csv")

Plot.plot({
  "marks": [
    Plot.barY(penguins, Plot.groupX({"y": "count"}, {"x": "species"})),
  ]
})

Using rectY works via the observable website.

https://observablehq.com/d/1536c13b55654b3e

Apologies if I'm missing something obvious!

Add padding for `jsdom` renderer output

When rendering plots with jsdom, the generated SVG or HTML as no padding at all. It could be nice to add some in VSCode or Jupyter lab, but maybe not in Quarto.

Exporting an image

How would I export a plot to html/png/svg?

Ideally I would like to be able to do something similar to Plotly i.e. fig.write_html("path/to/file.html") / fig.write_image("images/fig1.png") / fig.write_image("images/fig1.svg")

Implement "dark mode"

Add a setting to allow to generate plots in dark mode.

Replace IPC/arrow with parquet/hyparquet ?

See if IPC serialization could be replaced with parquet, using hyparquet.

Deprecate `kwargs` syntax ?

The kwargs alternative syntax allows to define a plot with keyword arguments instead of a dict:

op(
    grid = True,
    color = {"legend": True},
    marks = [
        Plot.dot(data, {"x": "x", "y": "y", "fill": "type"})
    ],
)

I'm not sure this syntax is very useful, and it tends to make some of the code less readable. Maybe it could be removed ?

Add NEWS file

Add NEWS.md and integrate with doc/.

Simplify dataframe serializations

upgrade to apache-arrow 15.0.0
remove polars schema conversions to 32bits
remove base64 encoding

Widget JS error troubleshooting

The "model not found" error could be related to:

jupyter-widgets/ipywidgets#3577

Widget renderer output lost when notebook is closed in JupyterLab

JupyterLab doesn't save widget output state when the file is closed, so all cells have to be recomputed to see the output.

When using the jsdom renderer, output is saved between sessions.

As with #2 this is probably an upstream issue!

This appears to be the relevant issue jupyter-widgets/ipywidgets#3823 and it looks like people are making progress to fix it.

I tried downgrading jupyterlab and ipywidgets as suggested here without any luck. jupyterlab/jupyterlab#15361 (comment)

juba / pyobsplot Goto Github PK

pyobsplot's People

Contributors

Stargazers

Watchers

Forkers

pyobsplot's Issues

Recommend Projects

Recommend Topics

Recommend Org