TIE Analysis

Basic ESRI ArcMap's Toolbox to perform Trace Information Extraction (TIE) Analysis.

Installation

Basically, this is an ESRI ArcMap/ArcGis Pro ToolBox (Python2.7 and Python 3.9) running external scripts using Python3.

The install.py script is very swisstopospecific, it basically:

• Create a .condarc in the $HOMEdirectory ( _H:_ )
• Set up a geocover.ini file in the $HOME\local\geocover directory. Check the variables tie_conda_env and proxyfor corectness.
• Create a conda environment in D:\conda\envs\TIE
• Map the drive T: to a conda environment on the LgX

Conda

On the Citrix VDI and the BURAUT laptop, the only Python3 available is the one of the ESRI ArcGis Pro C:\Program Files\ArcGIS\Pro installation.

There are several limitations:

• You cannnot modify the default argispro-py3 conda environment. You may only clone it.
• On BURAUT, you only have a C:\ drive and may only write to C:\LegacySW
• On the VDI, you have C:\ and D:\ and may write almost everywhere.
• Calling any conda env from within ArcGis Pro will pollute the sys-path with the default environment variables.

Creating a conda env from scratch (takes about 2 hours). The command will create a new env D:\conda\envs

H:\code\bitbuckets\arcmap-tie-toolbox> setup_conda_environment.cmd TIE

Use the network env on the LgX server, mapped on _T:_

H:\> activate T:\TIE

Clone the same environment (takes 30 minutes)

H:\code\bitbuckets\arcmap-tie-toolbox>  clone_conda_environment.cmd TIE

---

NOTE

As there are no arcgis and arcpy modules in this environnement, it's labelled as brokenby ArcGis Pro.

Additionnally, you may try to speed up to subsequent install by using mamba:

conda install mamba mamba install ipycytoscape

You may then activate your conda environnement:

C:>argis-pro-py3> deactivate

activate ANNA

Install modules

Linux:

python -m pip install 'geocover-utils @ git+https://bitbucket.org/procrastinatio/geocover-utils.git@0.1.1'

Windows: python -m pip install --trusted bitbucket.org https://bitbucket.org/procrastinatio/geocover-utils/get/master.zip

python -m pip install --trusted bitbucket.org  https://bitbucket.org/procrastinatio/lg-tie-lib/get/packaging.zip

TODO: add the geocover-utils and traces-extraction libraries

Copy the geocover.ini configuration to H:\local\config\geocover

"C:\Program Files\ArcGIS\Pro\bin\Python\Scripts\conda.exe" create -p D:\conda\envs\EMPTY_FROM_YAML -v --override-channels -c conda-forge -y --show-channel-urls python=3.9.18 conda pip setuptools

Python2

Install Python2 deopnedecies

 H:\> C:\Python27\ArcGIS10.8\Scripts\pip.exe install psutil

 H:\> C:\Python27\ArcGISx6410.8\Scripts\pip.exe install psutil

Create a conda with mayavi and arcgis

Create a bare mininal environment for python 3.9

H:\code\arcmap-tie-toolbox> "C:\Program Files\ArcGIS\Pro\bin\Python\Scripts\conda.exe"  create  -p C:\legacySW/EMPTY -v --override-channels -c conda-forge -y --show-channel-urls mamba python=3.9.18

Install the ESRI stuff

H:\code\arcmap-tie-toolbox> "C:\Program Files\ArcGIS\Pro\bin\Python\Scripts\conda.exe"  install -p C:\legacySW/EMPTY -v -c conda-forge  -c esri  -y --show-channel-urls arcgis arcpy

Add the usefull stuff

H:\code\arcmap-tie-toolbox> "C:\Program Files\ArcGIS\Pro\bin\Python\Scripts\conda.exe"  install -p C:\legacySW/EMPTY -v -c conda-forge  -c esri geocube geopandas matplotlib mayavi numpy rasterio scipy "shapely>=2.0.0"  scikit-image dask

Extract the package list (full source url)

H:\code\arcmap-tie-toolbox> "C:\Program Files\ArcGIS\Pro\bin\Python\Scripts\conda.exe" list  --explicit -p C:\legacySW\EMPTY >h:\EMPTY__list_export.yaml

On the VDI recreate the environment with:

"C:\Program Files\ArcGIS\Pro\bin\Python\Scripts\conda.exe" create   -p D:\conda\envs\EMPTY   --file  EMPTY__list_export.yaml

Python3

Modules to install in a fresh env

conda install geocube geopandas matplotlib mayavi numpy rasterio scipy "shapely>=2.0.0"  scikit-image dask

Update in an ArcGis Pro if cloning (in theory)

conda install --name ARGISTIE  --channel conda-forge geocube geopandas mayavi  rasterio "shapely>=2.0.0"  scikit-image 

Build

(c:\LegacySW\build-distribute) H:\code\arcmap-tie-toolbox>conda build --debug  recipe -c esri -c conda-forge

TIE

Using a dask pipeline

Exporting TIE Analysis to GeoJSON or KML

Usage

Select an area (a few square kilometers, around 1:10'000 or bigger, no sheet border!)

GDB structure

Download DEM and create a mosiac

New Mosic file

Hillshade from DEM mosaic (only for demo)

TIE Analysis

Takes about 10-15 minutes for 5 km2

3D View and TIE Analysis

map.geo.admin.ch (3D)

https://dubious.cloud/TIE/data/Kaiseregg/2.0/tie_as_multilines_under.kml 
https://dubious.cloud/TIE/data/Kaiseregg/2.0/tie_as_lines_under.kml