[Keith]

Yes, an update of the server's PHP version has caused date()/mktime() errors. This will be fixed in the near future. Apologies in the meantime.

[Keith]

Yes the enclosure checks use opposite edges – geomEnclose2 already does this.

[Keith]

Min and max area rules are in microns squared. If the docs say otherwise then they will be corrected…

[Keith]

The Python interface is nearly done, just have to do the multi enclosure check, which is a variant of the existing conditional enclosure check. Hopefully will be finished and tested by 17/4/16.

The interface code recognises rules like this:

out_layer = geomAllowedEncs(layer1, layer2, 10,10,4],[2,28,2],[0,32,2, message)

The rule triplets are of the form [enc1, enc2, max_num_enc1] where enc1 is the smaller enclosure, enc2 is the larger enclosure, and max_num_enc1 is the maximum number of edges that can have the smaller enclosure.

Another question though: is the 'default' rule of 10nm enclosure on all 4 edges an absolute rule, or is it the minimum? i.e. if you have an enclosure of say 20nm on all sides, should this fail?

[Keith]

You need to OR the two flags together.

[Keith]

Not at present. It could probably be done though. I'll see how much is involved.

[Keith]

The optional message has also been added to geomAllowedWidths() now.

[Keith]

OK so some more questions:

1. For the multiple enclosure rule, is it a general case that there is e.g. a larger rule for 2 sides and a smaller rule for the other two sides (28 on 2 sides, 2 on other 2)? Or can the number of sides with larger/smaller rule be other than 2?

2. The example has 1 default rule (10nm all sides) plus 2 other rules (28/2 and 32/0). Can there be e.g. 1 default rule and 1 or more other rules? Or no default rule and 1 or more other rules?

Is this type of rule only for enclosure?

[Keith]

The geomAllowedWidths issue is fixed in 4.4.67. It now correctly obeys the 'horizontal' and 'vertical' flags:

horizontal: tracks running horizontally only are considered. More precisely, the width of objects with two projecting horizontal edges is checked.

vertical: tracks running vertically only are considered. The width of objects with two projecting vertical edges is checked.

none: the minimum width of objects is checked.

[Keith]

If you want the shape to be exactly 0.064 x 0.064 then use the following:

geomArea(layer, 0.004096, 0.004096)
geomWidth(layer, 0.064)

The first flags any shapes that have area smaller or bigger than 0.064 x 0.064.
The second traps shapes where one edge is > 0.064 and one < 0.064 but the area passes the first check.

I will have a look as to why geomAllowedWidths() should treat width/height differently. As with no flags, w/h = 0.064 and h/w = 0.065 should pass.

[Keith]

The issue is that it's seeing the 0.064 as being a valid width and hence the shape passes the rule.

What should it do if there was a track of say width = 0.064 and height 0.100?

Only thing you can do is use the horizontal/vertical flags to differentiate width and height.

About the comment/message, this is a limitation of the way the function is wrapped in Python. I hope to resolve this at some point.

[Keith]

In 4.4.67 I will add the following. If p is a polygon, then

pt = p[0]

will return the Point at index 0 etc.

p[0].x

gives the x coord at index 0

[Keith]

Hmm, this could be improved. Given a polygon p,

pl = p.ptlist()
pts = pointList(pl, p.nPoints())
x0 = pts.at(0).x
y0 = pts.at(0).y

etc

I'll see if I can make it a bit more easy in Python so ptlist() returns a Python list.

[Keith]

OK I see what you are trying to do.

It is not trivial so may take a little while to get a fully tested implementation.

[Keith]

It will be in 4.4.67 to be released on 10th April

[Author]

Posts