MSSpectrum plots by axelwalter · Pull Request #3 · OpenMS/pyopenms_viz

axelwalter · 2024-05-15T10:04:07Z

No description provided.

- input is pyopenms.MSExperiment().get_df(long=True) - relative intensity support - can bin peaks to reduce complexity and increase performance - binning done with pandas internally, simpler and less dependencies then datashader and holoviews

jcharkow

Great work! Had a few minor comments with the main one being possibly we should make the ion mobility source code in a separate function so it is easier to locate.
I think some of this code is overlapping with the chromatogram plotting interface so we will need to do some work with integrating them upon merging.

Also please include test files so that I can run it on my end.

jcharkow · 2024-06-05T15:07:58Z

pyopenms_viz/BasePlotter.py

+# A colorset suitable for color blindness
+class Colors(str, Enum):
+    BLUE = "#4575B4"
+    RED = "#D73027"
+    LIGHTBLUE = "#91BFDB"
+    ORANGE = "#FC8D59"
+    PURPLE = "#7B2C65"
+    YELLOW = "#FCCF53"
+    DARKGRAY = "#555555"
+    LIGHTGRAY = "#BBBBBB"
+


I think I would rather allow users to specify a palette rather than hardcode colors. We have an implementation for this with the chromatogram plotter can adapt to here

Since each peak is represented by a row in the dataframe, users can specify extra columns for custom peak color, custom annotation color and custom annotation text. This will override this default coloring. What do you think about that solution?

Just annotating mz with default colors:

Setting custom peak (all green) and annotation colors (all blue) defined in df:

Annotating ions default colors:

Annotating ions AND custom color (all blue) includes peak and annotation text:

I like that!

jcharkow · 2024-06-05T15:08:15Z

pyopenms_viz/BasePlotter.py

    def engine_enum(self):
        return Engine[self.engine]

+


remove extra line

Should we use the same formatter? Using black results in these double empty lines after class definition.

Sorry I'm not quite sure what you mean

jcharkow · 2024-06-05T15:10:01Z

pyopenms_viz/MSExperimentPlotter.py

+        ax.set_xlabel(self.config.xlabel, fontsize=10, color=Colors["DARKGRAY"])
+        ax.set_ylabel(
+            self.config.ylabel, fontsize=10, style="italic", color=Colors["DARKGRAY"]
+        )
+        ax.xaxis.label.set_color(Colors["DARKGRAY"])
+        ax.tick_params(axis="x", colors=Colors["DARKGRAY"])
+        ax.yaxis.label.set_color(Colors["DARKGRAY"])
+        ax.tick_params(axis="y", colors=Colors["DARKGRAY"])
+        ax.spines[["right", "top"]].set_visible(False)


Not sure if hardcoding colors is the best idea

This is just for axis and label text colors. Should that be customizable by the user? Where to stop (font sizes, font weight, padding, tickformatters, etc...) ?

I think if we are changing the defaults then we should have a global variable somewhere for different plot settings (e.g. jupyter notebook vs streamlit) does not necessarily have to be accessible to user. It will also be easier to change later if we want to change defaults

pyopenms_viz/MSExperimentPlotter.py

jcharkow · 2024-06-05T15:13:36Z

pyopenms_viz/SpectrumPlotter.py

+        colormap = {
+            "a": Colors["PURPLE"],
+            "b": Colors["BLUE"],
+            "c": Colors["LIGHTBLUE"],
+            "x": Colors["YELLOW"],
+            "y": Colors["RED"],
+            "z": Colors["ORANGE"],
+        }
+        return colormap.get(annotation.lower(), Colors["DARKGRAY"])


TODO integrate this into common color framework

How would that look like? We can discuss in person.

jcharkow · 2024-06-05T15:25:13Z

pyopenms_viz/SpectrumPlotter.py

+                        x=peak["mz"],
+                        y=intensity,
+                        text=text,
+                        text_font_size="8pt",


add this to config instead of harcoding?

Agreed, there should be some universal settings defined in BasePlotter.

pyopenms_viz/SpectrumPlotter.py

axelwalter · 2024-06-05T17:55:14Z

Thanks for the comments, I will work on them tomorrow. Agree on the ion mobility part. We will see how to resolve the overlaps. As for testing I have started working on a demo streamlit app to showcase stuff and parameters easily. Also working with the new plotly events from streamlit 1.35 to update the plot while zooming in with the MSExperiments.

- separate MSExperiment 2D and 3D plotting - separate MSSpectrum ion mobility plotting - Spectrum plot x-axis adjustment to not cut off peaks and annotations - fix legend always visible in Bokeh spectrum plot

jcharkow

Looks good to merge. Next steps is to determine parameters to be added to the base plotter however that will be easier to address when everything is merged and we can see which parameters are used across the different plotting tools.

jcharkow · 2024-06-06T13:40:19Z

pyopenms_viz/BasePlotter.py

+# A colorset suitable for color blindness
+class Colors(str, Enum):
+    BLUE = "#4575B4"
+    RED = "#D73027"
+    LIGHTBLUE = "#91BFDB"
+    ORANGE = "#FC8D59"
+    PURPLE = "#7B2C65"
+    YELLOW = "#FCCF53"
+    DARKGRAY = "#555555"
+    LIGHTGRAY = "#BBBBBB"
+


I like that!

jcharkow · 2024-06-06T13:41:11Z

pyopenms_viz/BasePlotter.py

    def engine_enum(self):
        return Engine[self.engine]

+


Sorry I'm not quite sure what you mean

jcharkow · 2024-06-06T13:42:42Z

pyopenms_viz/MSExperimentPlotter.py

+        ax.set_xlabel(self.config.xlabel, fontsize=10, color=Colors["DARKGRAY"])
+        ax.set_ylabel(
+            self.config.ylabel, fontsize=10, style="italic", color=Colors["DARKGRAY"]
+        )
+        ax.xaxis.label.set_color(Colors["DARKGRAY"])
+        ax.tick_params(axis="x", colors=Colors["DARKGRAY"])
+        ax.yaxis.label.set_color(Colors["DARKGRAY"])
+        ax.tick_params(axis="y", colors=Colors["DARKGRAY"])
+        ax.spines[["right", "top"]].set_visible(False)


I think if we are changing the defaults then we should have a global variable somewhere for different plot settings (e.g. jupyter notebook vs streamlit) does not necessarily have to be accessible to user. It will also be easier to change later if we want to change defaults

axelwalter added 15 commits May 14, 2024 16:07

complete MSSpectrum plot with Plotly

8a9f91b

add matplotlib spectrum plot

8b23c27

grayscale colors for peaks

0909d10

format relative intensity tick labels

cc8fc82

bokeh plot

d884273

color peaks based on annotation

113aa4c

ion mobility spectrum plot with plotly

af2aa6a

function to ensure spectra are in list format

fad3337

spectrum ion mobility plot with matplotlib

bbde6f2

no custom color for ion mobility peaks

904292f

ion mobility default y-label

9f1c3b5

improve matplotlib and plotly IM plots

52f8f89

spectrum ion mobility plot with bokeh

c15998f

docstrings

11a78c8

cleanup

b70d934

axelwalter requested a review from jcharkow June 3, 2024 13:03

axelwalter added 2 commits June 4, 2024 16:40

add MSExperiment plotting

08cc65a

- input is pyopenms.MSExperiment().get_df(long=True) - relative intensity support - can bin peaks to reduce complexity and increase performance - binning done with pandas internally, simpler and less dependencies then datashader and holoviews

add 3D peak map matplotlib

f7de02a

jcharkow requested changes Jun 5, 2024

View reviewed changes

demo streamlit app

7f77fa2

diverse improvements

762b88f

- separate MSExperiment 2D and 3D plotting - separate MSSpectrum ion mobility plotting - Spectrum plot x-axis adjustment to not cut off peaks and annotations - fix legend always visible in Bokeh spectrum plot

axelwalter requested a review from jcharkow June 6, 2024 09:23

axelwalter added 2 commits June 6, 2024 13:11

update demo app

ff21e36

add help text to parameter widgets

5a6840d

jcharkow approved these changes Jun 6, 2024

View reviewed changes

axelwalter merged commit b2591bf into OpenMS:main Jun 10, 2024

Conversation

axelwalter commented May 15, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jcharkow left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

axelwalter commented Jun 5, 2024

Uh oh!

jcharkow left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

axelwalter commented May 15, 2024 •

edited

Loading