Assign keywork for RD_SMSPEC_NETWORK_VAR

[andreas-el]

Issue
Resolves #943

Picks changes from commit OPM/ResInsight@70f31aa

⚠️ Currently untested, awaiting access to Troll to try to replicate Network vectors in summary.
Alif has tested using Troll data. ✅

[andreas-el]

Alif Be wrote:
I think I figure out the problem. wgname is not defined for network keyword
12:23
we need to add network to this line

resdata/lib/resdata/smspec_node.cpp

Line 567 in 4062d1e

this->var_type == RD_SMSPEC_LOCAL_WELL_VAR)

smspec_node.cpp
this->var_type == RD_SMSPEC_LOCAL_WELL_VAR)
https://github.com/[equinor/resdata](https://github.com/equinor/resdata)|equinor/resdataequinor/resdata | Added by GitHub
12:24
tested on troll dataset and seems to work

[eivindjahren]

Looks good, but are we sure that network variables always have well names? In any case, we now need to change the following: https://github.com/equinor/ert/blob/e35e7e07b39534ac3ed83085a87a970b633dcc70/src/ert/config/_read_summary.py#L167-L170

[alifbe]

Looks good, but are we sure that network variables always have well names? In any case, we now need to change the following: https://github.com/equinor/ert/blob/e35e7e07b39534ac3ed83085a87a970b633dcc70/src/ert/config/_read_summary.py#L167-L170

Yes, network vector is always have a node name which is stored in either WGNAMES or NAMES. There are obviously some special vectors which are beginning with N but not network vector.

See this commit OPM/ResInsight@1cf2cd7

[eivindjahren]

We checked and it seems that WGNAMES or NAMES will be present for network variables.

[andreas-el]

Network variables are either juntions or pipes, e.g.
NJOPR, NJWPR, NPGFR.
Ref. Intersect manual p1619 onwards.

[eivindjahren]

Yes, this seems to be correct.