Do not use the same module name as 3rd party libs

Development Beta Testing
1

Hi.

It takes a while since I was able to understand that my python versions (2 or 3) set of libs error was not related to my local installation (python 2 and python 3).

The cause was just a poor (or lazy) name for a local script. AI gives me a trick for tracking this issue down:
$ ls | grep networkx.py
Sure, a bad idea to use this name when the script will call networkx…

import networkx
import matplotlib.pyplot as plt
...
etc.

So, a local path and naming issue. This can be a problem on addons.

1 Like
2

Just for more details. I am looking at any graphical view for a simple representation of distributions for main surnames into a small village (around 200 houses and farms in 1819).

I get some indexes which will match the timeline properties of an individual or family record (names and properties with dates and locations). In fact often a simple name, but sometimes with the spouse and the occupation. So, linked with farms, lands, etc. records. The second set of documents is only pages (5) with the list of families and a house number (census of 1819).

This might not work for big cities. So, I only try to look at a view at a glance. Here, the first try:

1819_Vendenheim_Patronymes_rues_quartiers
1819_Vendenheim_Patronymes_rues_quartiers1366×655 166 KB

it is in french but only simple textual labels (rue = street).

The lacks will be clearly on local dataset. Even with most sources available on line, I had to check the next censuses for a robust streets and surnames sequence (based on house numbers). Once done, to make the links will be very quick, with a surnames list and some python set of libs.

Maybe wasting time was when I made my mistake with the local module name issue. And I just thought that it could be a common strange issue with some addons and 3rd party libs or with any bundle gramps packages.

I could improve a little bit the result but this will be difficult to generate a global addon for gramps. Even with AI support, we need to make our local lists according to available sources. As OCR wasn’t efficient enough (i.e. worse than myself). It was an experimentation : a selection of the village and use a small set of “clear” surnames. So, not a “real” life test, but it was fun.