Charts II: NGA - 2700 Charts

[recri]

Originally Posted by HappySeagull

oooh!that was fast!nicely done!

Yes, well, I know that it reads the text form of the catalogs, and that it thinks it understands everything it reads, and it appears to be right in the few cases I've checked, but I don't know what it missed or messed up on. A good start.

Oh, and it isn't distinguishing between base charts and insets in an entirely consistent way, though that may be the catalog's problem. So sometimes you get a list of the additional plans, and sometimes you get a list of everything on the page.

[HappySeagull]

Note to self:when done,change the chart status to "doable in Phase 1" if nothing else is wrong or no kap will be made...(sigh)...

[z.stadler]

Originally Posted by HappySeagull

Note to self:when done,change the chart status to "doable in Phase 1" if nothing else is wrong or no kap will be made...(sigh)...

It will be valuable to keep the chart characterization, so perhaps "Doable in phase 1 - Special PLYs needed"

[bcn]

Great work Recri!

For Region 5 this yields another 288 charts to do....

There is none for Region 6. Is this correct?

For the categories I would prefer to have their origin traceable:
- "inset rectangular"
- "inset not rectangular"
- "composed rectangular"
- "composed not rectangular"

And probably it will be advisable to create a additional parameter:
- number of PLYs: 4, 5, 6, 8(?) or "other" (see 53283 with round corner)

The charts in the category "special PLYs needed" will have to add as well at least 4 PLY's for each "indent" or "extrusion".

Hubert

[nohal]

Originally Posted by bcn

Great work Recri!

Yes, thanks a lot

For Region 5 this yields another 288 charts to do....

There is none for Region 6. Is this correct?

Unfortunately it's impossible to convert the region 6 PDF catalog to text for some reason. I will try to stop by a friend with an Acrobat license and do something about it when time allows.
If someone is interested in the lists of the charts with insets as produced from our current data (the ones we knew have them from the catalogs + the ones identified based on the status set by the users), they are at https://docs.google.com/leaf?id=0B5c...ZDY0NWZk&hl=en
I will merge this info with the data recri produced later.

For the categories I would prefer to have their origin traceable:
- "inset rectangular"
- "inset not rectangular"
- "composed rectangular"
- "composed not rectangular"

And probably it will be advisable to create a additional parameter:
- number of PLYs: 4, 5, 6, 8(?) or "other" (see 53283 with round corner)

The charts in the category "special PLYs needed" will have to add as well at least 4 PLY's for each "indent" or "extrusion".

Hubert

I will add this as we progress further.

Right now it looks like the process will have a couple of consequent steps:
1. We have to complete the list of the insets (Google Docs look ok for this task)
2. We have to define the extents of the insets (new PHP/JavaScript hell, but luckily not as complicated as the calibration)
3. We have to create the images on the server (pretty easy, will take time)
4. We have to calibrate the rectangular insets (extension of the current logic and forms, not too difficult)
5. We have to calibrate the non-rectangular stuff (We have to invent something smart, I have a few ideas already)

All this comes when we are done with the charts we can handle already and after I have a solution for defining the chart polygons ready (should be really soon now).

Pavel

[paulr]

Originally Posted by nohal

Unfortunately it's impossible to convert the region 6 PDF catalog to text for some reason. I will try to stop by a friend with an Acrobat license and do something about it when time allows.

Pavel

I can save nga6.pdf to html, with apparently decent layout and accuracy, straight from Acrobat.
Let me know if you'd like me to upload the output somewhere.

Paul

[cagney]

Nohal & recri

Here is Region 6 converted from pdf to text.
nga6.txt.doc

Thomas

[recri]

Originally Posted by paulr

I can save nga6.pdf to html, with apparently decent layout and accuracy, straight from Acrobat.
Let me know if you'd like me to upload the output somewhere.

Paul

My conversions were making column first tables with no gaps in the chart number list to indicate where the plans insert into the list.

But one last try with pdftohtml (from the Ubuntu popplerutils package) has produced a row first table break out, and I've parsed it with minor hassles.

https://docs.google.com/spreadsheet/...JuOXc&hl=en_US
is the new version, I don't have time to look it over this morning.

[recri]

A few fixes from yesterday and this morning:


92356 - NW REF point outside main chart N and W edges, changed status to non-rectangular
37403 - moved NE REF point from inset chart frame
97469 - SW corner lat/lng corrected
53283 - NE corner lat corrected
24509 - SW corner lng corrected
18722 - NE corner lng corrected
25723 - SW corner lng corrected
24404 - SW corner lng corrected
81155 - NE corner lng corrected
26100 - added NW REF point
97481 - NE corner lat corrected
71400 - NE corner lat corrected to graticule

26340 corrected SW corner coordinates
26284 supplied main chart N and E coordinates
23101 corrected SW corner lng from E to W
62302 corrected NE corner lng

[wdkester]

Chaart 26284 is a bit odd in that the right hand margin of the main chart is not contiguous and the coordinates in the upper right corner do not belong to the main chart;however, if you assume that the eastern margins of both charts are in a straight line (have the same X) then it's possible to get, what would be, the NE coordinates of the main chart. It's currently marked partly doable. Since we don't know how non rectangular charts will be handled and I may not be seeing things correctly, this one is questionable.

[recri]

Originally Posted by wdkester

Chaart 26284 is a bit odd in that the right hand margin of the main chart is not contiguous and the coordinates in the upper right corner do not belong to the main chart;however, if you assume that the eastern margins of both charts are in a straight line (have the same X) then it's possible to get, what would be, the NE coordinates of the main chart. It's currently marked partly doable. Since we don't know how non rectangular charts will be handled and I may not be seeing things correctly, this one is questionable.

I've done several that way, but you need to look out for corner panels which don't line up. I've seen a few where the corner panel lacks the graticule frame of the main panel so its corner is offset out.

I mark them as partly doable with rectangular/non-rectangular insets if the chart rectangle lines up square in opencpn, otherwise they end up as composed of rectangular/non-rectangular charts.

[recri]

Is the corner cut-out extender still running?

I show 95 charts with corner cut-outs extended last night, but only 80 this morning, but I haven't seen any extended corner charts in regions 2, 5, or 8 though they all have corners not visible charts.

[bcn]

Hi Recri,

with respect to zone 5 the situation is that all the extended corner files are worked through resulting in 17 charts that will need an even greater margin to catch the corners.

So those are due for the next loop.

Hubert

[recri]

Originally Posted by bcn

Hi Recri,

with respect to zone 5 the situation is that all the extended corner files are worked through resulting in 17 charts that will need an even greater margin to catch the corners.

So those are due for the next loop.

Hubert

Thank you, Hubert, perhaps that's what happened in regions 2 and 8, too.

[HappySeagull]

..yes,same here...just a few that were still out of bounds,so are reset as such for Nohal to find.
kick myself I didn't fill in every detail when the corners were easier to read..it's a bit more nuisance getting em off the nga zoomify site...