This is a set of scripts, some generic and some very specific
to some cadastral data sources.  The shape-utils directory contains
small, mostly generic scripts that take an .osm file as input and
output a processed .osm file:

merge-segments.py reads a .osm file with shapes, such as building
outlines/footprints, made up of any number of simple 2-node segments.
It tries to merge these segments into complete closed lines in the most
practical way.  The outline way is given the tags of every segment way
it uses, selecting the tag value that appeared most often if the values
vary.

join-lines.py merges nodes that are less than 2cm apart into one and
merges nodes into ways if they're closer than 2cm, or 20cm in case of
a node that is at the end of a line (which often indicates that a
closed shape hasn't been formed because of unintentional overshoot or
undershoot).

dedupe-segments.py filters out ways that contain the exact same set of
of nodes leaving only one.

merge-building-addrs.py is a version of the script at
https://github.com/balrog-kun/osm-addr-tools that uses a single input
.osm file containing address nodes and building outlines and puts the
address tags on the outlines, removing the nodes, depending on various
conditions.  Also allows the node to be some distance off the building
wall.

egib.py applies the rules of the basic Polish cadastral data "K-1"
manual (this is a long story) to convert some attributes into more
usual OSM tags.  This script will almost surely need modification for
any dataset other than the one small county it was originaly made
for, even though most of the rules should be the same throughout
Polish cadastre.

These scripts together do (in smaller steps) most of what convert.py
does all at once at the cost of it being a hard to read ball of
spaghetti code.  They may be use in conjuction with ogr2osm.py to
convert data from various sources such as local governments' geodesy
offices, which often only have (or make available) data produced by
automatic vectorisation of old paper cadastral maps meaning that the
files in reality describe a picture of the map rather than a geo-
database.  In this case they'd be used in the following order, after
some reconfiguration:

ogr2osm
dedupe-segments
join-lines
merged-segments
egib
JOSM manual corrections/clean-up
JOSM download of existing data
JOSM manual conflation
merge-building-addrs

Or they can be used individually.

The convert.py script takes a DXF file as exported from the ewmapa
program and does the first 5 processing steps and some more.  It
probably needs a new config file for a new dataset.  Various
config files used with data from various Polish counties are
included.