wesleythegeolien / t20-dash_well_viz Goto Github PK

The aim is to try and build some interactive log visualisations using dash in notebooks and demonstrate what is possible (and also what the limitations are) using dash on subsurface data. Ultimately it would be good to try and collate all the examples into a library of templates to share after the Hackathon :swung: I have put together a quick Google Sheet with some initial ideas, so feel free to either pick one that interests you or add something to the list that you think would be fun to make, all ideas welcome.

License: Apache License 2.0

Lasso 13.16% Jupyter Notebook 86.79% Python 0.04% CSS 0.01%

swung-t20

t20-dash_well_viz's Introduction

t20-Dash_Well_Viz

Introduction

This repository is part of the Software Underground's Transform 2020 hackathon.

The aim is to try and build some interactive log visualisations using dash in notebooks and demonstrate what is possible (and also what the limitations are) using dash on subsurface data.

Ultimately it would be good to try and collate all the examples into a library of templates to share after the Hackathon swung.

Ideas

On Google Sheets

Helper module

Some helper code will be in the package dashwellviz. To install, clone this repository and install using pip -e e.g.

$ git clone https://github.com/WesleytheGeolien/t20-dash_well_viz
$ pip install -e t20-dash_well_viz
$ python
>>> import dashwellviz

t20-dash_well_viz's People

Contributors

Stargazers

Watchers

Forkers

dougmcclymont kinverarity1 mharty3 jjpantano st4tic0 esadan dcostalouis bsburnham morrgeo 8556732 mlbelobraydi david-thul abv-hub mariabravosegnini zhang-cugb

t20-dash_well_viz's Issues

generating new curves

I have written down some additional capabilities that can build off https://toddheitmann.github.io/PetroPy/index.html#
This needs more refinement and post to inquiry about possible collaboration.

log.fluid_properties_parameters (read_csv)
generate_fluid_property_curves
edit_fluid_property_curves (optional)
generate_fluid_property_curves
formation_multimineral_model

generate_list_of_curves
generate_list_of_calculations (read_csv)
select_calculation(s)
select_well(s)
map_function_arguments_to_curves
assign_values_for_arguments_for_no_curves
generate_new_curve(s)

Facing issue with CRS conversion for display of well coordinates on the map

I have used the well header info to read Lat Lon for each well, Petrel project suggests that a CRS is GDA 94. I have attempted first to use the Lat/Lon (turned them to decimal from Degree Minutes Seconds) directly from the well header. This shows the wells thousands of Km north of where they are supposed to be.
I wrote a small function to convert the Lat Lon to WGS 84 but the converted numbers also look out of place.

Add functionality to configure log/track display

I've been thinking about how best to let the user select options for log and track display.

I'm thinking that the best way might be on a track by track basis. This way, the user could have control over:

the number of tracks
the log curves displayed in each track
the display properties for those curves (color, line style, log/linear etc)

I think we can nest this inside of some details to avoid the sidebar getting too cluttered.

It would be great to hear some other thoughts toward this idea

Pick a las file and choose a log from it

Continue to work on the well top picking tool

This is a general issue, and will need to be split up into many later on. I think this tool has a lot of value as a manual top picker. This is a fundamental step for a geologist working on interpreting well logs. Any contributions in code or ideas are welcome.

Create tabbed plots: set up tabs so we can display more plots than just the crossplot

https://dash.plotly.com/dash-core-components/tabs

Ressouces

Add documentation to the wiki

Option for logarithmic axes in the cross plot

This can probably be implemented as a checkbox below each of the dropdowns to select which data that is displayed on the axes.

Abstract dash app

The dash app should probably be as simple as possible having layers of abstraction will help keeping it readable.

See branch #32

Data upload to Dash app

provide a way to upload data locally to the dash databoard

Secondary x axis in figure with multiple subplots

@dougmcclymont needing this for filling between logs too.

Plotly docs are not much use: https://plotly.com/python/multiple-axes/

I don't think this helps us: plotly/plotly.py#1553

Very ugly hack solution which would work but I feel tired just thinking about how it would cause trouble later on: https://stackoverflow.com/questions/54232879/secondary-parallel-x-axis-on-plotly-charts-python

selection interactivity between cross plot and log curve trace plot

The idea here is to be able to interactively select and highlight data across the cross plots and log plots.

The input would be a selection via lasso or box selection on the cross plot (the blue circle) and the output would be some indication on the log trace plot which points are selected (the yellow highlights.) It also should work in reverse with selections on the log traces updating the selection on the cross plot.

This is the relevant information from the dash tutorial.
https://dash.plotly.com/interactive-graphing

Deploy our app(s) to the cloud for demo

In their tutorial yesterday, Matt and Brendon showed how easy it was to deploy an app using https://www.pythonanywhere.com/. I think it would be really cool if we could get this working and post the link during the demo.

I have never tried it before, but it looks like it only takes a couple of clicks to get it running