You can subscribe to this list here.
2007 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(3) |
Nov
|
Dec
|
---|---|---|---|---|---|---|---|---|---|---|---|---|
2009 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(4) |
Jun
|
Jul
(5) |
Aug
(17) |
Sep
|
Oct
(1) |
Nov
|
Dec
(1) |
2010 |
Jan
(1) |
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(2) |
Dec
(8) |
2011 |
Jan
|
Feb
|
Mar
(3) |
Apr
|
May
|
Jun
|
Jul
|
Aug
(1) |
Sep
|
Oct
|
Nov
|
Dec
|
2016 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
(1) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
From: Deepak S. <dee...@id...> - 2016-06-27 18:21:10
|
Installing SVN ============================================== windows ---------- Get the subversion client tortoiseSVN (http://tortoisesvn.tigris.org/) linux ------ Install Subversion client software (http://subversion.tigris.org/) Ubuntu: http://subversion.tigris.org/getting.html#ubuntu $ sudo apt-get install subversion For more information on setup please check out https://help.ubuntu.com/community/Subversion You try the command line client or a GUI client like rapidSVN http://www.rapidsvn.org/index.php/Main_Page *NIX ----- Check out http://subversion.tigris.org/getting.html for binaries or source balls. you can try the command line client or a gui client like RapidSVN More info: --------- More info can be found at the PAGC CODE section on Source Forge https://sourceforge.net/svn/?group_id=66325 Accessing the Sourceforge repository using SVN =============================================== Change history before this publication ======================================== https://pagc.svn.sourceforge.net/svnroot/pagc/trunk Name: trunk URL: https://pagc.svn.sourceforge.net/svnroot/pagc/trunk Repository: <None> Repository UUID: <None> Revision: 6 Node Kind: directory Last Changed Author: putler Last Changed Rev: 6 Last Changed Date: 2006-12-15 21:18 GMT Text Last Updated: 2006-12-15 21:18 GMT Properties Last Updated: 2006-12-15 21:18 GMT Checksum: <None> Changes made to SVN =================== * directory structure changed.... Read ....svn/pagc/README.txt for new directory layout * old pagc project files moved to svn/pagc/pagc * added new directories ofr different pagc components * pagc library sits alone. Install dependencies UBUNTU ==================== question: runtimes versus dev headers * The gcc C compiler: sudo apt-get install gcc * Apache httpd webserver: sudo apt-get install apache2 * Berkeley db: sudo apt-get install libdb4.4-dev * Fastcgi Apache module: sudo apt-get install libapache2-mod-fastcgi * Fastcgi library: sudo apt-get install libfcgi-dev * pthreads? * For the brave. One liner: sudo apt-get install gcc apache2 libapache2-mod-fastcgi libfcgi-dev Install Dependencies Windows ============================ The gcc C compiler, autotools, and utilities they use are not provided. These should be readily available for your system if not already installed. MINGW versions for Windows are available from http://www.mingw.org. Apache httpd webserver. Berkeley db. Fastcgi Apache module. Note : The software also will function as a CGI program. In this case Fastcgi is not needed. Fastcgi library (included in the Fastcgi developer's kit). This too is not needed if the responder is deployed as a CGI program. variables to consider ========================== Many ways to do this, however these are sugested (ubuntu at least) Web Users and groups -------------------- $APACHE_OWNER = www-data $APACHE_GROUP = www-data web directories --------------- ${CGIIDEST} = /usr/lib/cgi-bin $PAGCCGI = /usr/lib/cgi-bin/geocode_response (you might have to make this directory manualy) $INSTALLDIR = /var/www/geocode/build directory structure ------------------- /var/www/geocode/build streets parcels standard Install PAGC library ======================= install ---------- tar -xzvf pagc-0.1.4.tar.gz cd pagc-0.1.4 chown ${APACHE_OWNER}:${APACHE_GROUP} * ./configure make make install Confirm install --------------- ls /usr/local/lib/*pagc* /usr/local/lib/libpagc.a /usr/local/lib/libpagc.la /usr/local/lib/libpagc.so /usr/local/lib/libpagc.so.1.0.0 /usr/local/lib/libpagc.so.1 remove ------------- Install geocode_response ========================= install ---------- tar -xzvf geocode_response-1.0.0.tar.gz cd geocode_response-1.0.0 chown ${APACHE_OWNER}:${APACHE_GROUP} * ./configure make template install -c -m 755 $APACHE_OWNER -g $APACHE_GROUP ./geocode_response ${CGIIDEST}/geocode_response Ubuntu standard install -c -m 755 www-data -g www-data ./geocode_response /usr/lib/cgi-bin/geocode_response Confirm install --------------- ls /usr/lib/cgi-bin/geocode_response/ /usr/lib/cgi-bin/geocode_response/geocode_response remove ----------- make uninstall Install pagc_build_schema =========================== install --------- tar -xzvf pagc_build_schema-1.0.0.tar.gz cd pagc_build_schema-1.0.0 chown ${APACHE_OWNER}:${APACHE_GROUP} * ./configure make template -------- install -c -m 755 -o $APACHE_OWNER -g $APACHE_GROUP ./pagc_build_schema $INSTALLDIR/pagc_build_schema install -c -m 755 -o $APACHE_OWNER -g $APACHE_GROUP ./build_streets.sh $INSTALLDIR/build_streets.sh install -c -m 755 -o $APACHE_OWNER -g $APACHE_GROUP ./build_streets.sh $INSTALLDIR/ build_parcels.sh ubuntu example: --------- install -c -m 755 -o www-data -g www-data ./pagc_build_schema /var/www/geocode/build/pagc_build_schema install -c -m 755 -o www-data -g www-data ./build_streets.sh /var/www/geocode/build/build_streets.sh install -c -m 755 -o www-data -g www-data ./build_streets.sh /var/www/geocode/build/build_parcels.sh confirm install ----------------- ls /var/www/geocode/build/ build_parcels.sh build_streets.sh pagc_build_schema remove ------- Install Standardization files =============================== Files: FILENAME=rules.txt FILENAME=lexicon.csv FILENAME=gazeteer.csv Source locations: ./svnroot/pagc/data/standard Destinations: $INSTALLDIR=${PARENT}/build $INSTALLDIR=${PARENT}/standard Template: install -c -m 644 -o www-data -g $APACHE_GROUP ./${FILENAME} $INSTALLDIR/$FILENAME Ubuntu Example: (from the build_schema directory) install -c -m 644 -o www-data -g www-data ../data/standard/rules.txt /var/www/geocode/build/rules.txt install -c -m 644 -o www-data -g www-data ../data/lexicon.csv /var/www/geocode/build/lexicon.csv install -c -m 644 -o www-data -g www-data ../data/gazeteer.csv /var/www/geocode/build/gazeteer.csv install -c -m 644 -o www-data -g www-data ../data/standard/rules.txt /var/www/geocode/standard/rules.txt install -c -m 644 -o www-data -g www-data ../data/lexicon.csv /var/www/geocode/standard/lexicon.csv install -c -m 644 -o www-data -g www-data ../data/gazeteer.csv /var/www/geocode/standard/gazeteer.csv install -c -m 644 -o www-data -g www-data ../data/standard/rules.txt /usr/local/share/rules.txt install -c -m 644 -o www-data -g www-data ../data/lexicon.csv /usr/local/share/lexicon.csv install -c -m 644 -o www-data -g www-data ../data/gazeteer.csv /usr/local/share/gazeteer.csv Install Schema tables ===================== Sources FILENAME=streets.dbf FILENAME=parcels.dbf INSTALLDIR=${PARENT}/streets INSTALLDIR=${PARENT}/parcels template: install -c -m 644 -o $APACHE_OWNER -g $APACHE_GROUP ./${FILENAME} $INSTALLDIR/$FILENAME ubuntu example: install -c -m 644 -o www-data -g www-data ./streets.dbf /var/www/geocode/streets.dbf install -c -m 644 -o www-data -g www-data ./parcels.dbf /var/www/geocode/parcels.dbf |
From: Joseph Y. <jo...@in...> - 2011-08-30 17:01:36
|
Hi all, per this message: https://sourceforge.net/mailarchive/message.php?msg_id=23314763 I too have ran into the issue of not having parcel data available to use*. Is the only option still to have to recompile the code? The message linked above speaks of a .conf or .xml setup, I think even a simpler approach could be taken - if the data directory isn't defined or can't be found then consider the data not being used, especially for parcels. If I've understood the documentation and the migration of directory structures, the code already traverses through a list of searching for where (all of) the data is located anyway. And it would be much easier to document a FAQ on "why isn't my parcel (or enter X) data not being used? A: is your data on PATH Y, than is it so say, John Q. Public - recompile the source. I'm new to the whole "figure out how to geocode on your own" due to an app I'm building and not being able to afford a commercial geocoding service yet. For example Google services will set you back $10K up front per year. I've contacted local government officials for data and one county came back with $1000 for just their data set - there are 99 other counties in that state...that $10K is looking pretty attractive, but my pocket lint keeps it real. So, keeping at it, I've discovered PAGC and am thankful beyond words for what the developers have given to the community, I just wished that I could wrap my head around the project faster as a whole. The documentation is fantastic and staring at it long enough finally starts to seep in (curse of the developer writing the documentation - you all know what you're doing so the docs makes since to ya - I'll give an example a little later in this conversation and completely plan to help with writing docs if it's OK). In the process of keeping it real - I've located Parcel and Address information for the state of Indiana (state I'm trying to setup geocoding for first). The location of this data should anyone else be interested is: http://inmap.indiana.edu/dload_page/infrastructure.html (Address points) http://inmap.indiana.edu/dload_page/demographics.html(Parcels) The data is provided in .geodatabase format so you'll be required to have the appropriate software to use it, or in time perhaps the API will be mature enough for users to have handy opensource tools to convert the files. The problem with the above data files is that parcels has no address data at all, so it's pretty much useless in the form it is now. Would you all think that the following is a good course of action to "join" the address points and parcels. What I'm thinking is importing the data into PostGres so that I can query the parcels data for each address point on whether or not the point is inside of the parcel boundries. If it is, then merge the parcel and address information so that the parcel data now has an address. Below are the .dbf data that each data set contains. I also imagine that I'll have to work on the Address Points so that Suffic, Add_type, Prefix and Zipcode contains data (all of those columns are blank in the data set). Here's the data header from the addresses data set (.dbf): OBJECTID,N,9,0 SOURCEID,C,40 ORIGIN,C,130 LOADDATE,D ADDRESS,C,60 NUMBER,C,1 SUFFIX,C,6 ADD_TYPE,C,254 PREFIX,C,10 STREET,C,60 CITY,C,60 COUNTY,C,60 STATE,C,60 ZIPCODE,N,9,0 Of which the Suffix, Add_type, prefix, county and zipcode fields are always blank (zipcode is actually "0"). And for the parcels data set: OBJECTID,N,9,0 IDPARCEL,C,40 SOURCEID,C,40 ORIGIN,C,130 LOADDATE,D PARCELID,C,100 CNTY_ID,N,4,0 LENGTH,N,19,11 AREA,N,19,11 Shape_Leng,N,19,11 Shape_Area,N,19,11 there isn't anything in the two data sets that can be used to link the two together other than to try a geospacial contains or is within kind of lookup. I've tried to read and understand the topics.html doc on how to go about setting up a .dbf to provide the data parser with where data is located at, but not being versed in .dbf formats or databases I'm not connecting the dots on what the parser needs and how to tell it what field/column the data is at/labeled. If someone could point me in a previous post or to a clearer description of what to do many thanks would be delivered. Sorry for such a long post, but I'll capture the output and make a write up on the wiki if that is alright. Sincerely, Joe |
From: Stephen W. <wo...@sw...> - 2011-03-11 18:49:00
|
Hi all, Walter fixed the issue I reported about zipcodes causing failures. Thanks for the awesome support. The fix was checked into branches/datastore/pagclib/score.c and you only need to rebuild pagclib and reinstall it. There is no need to reload your data. Also, I don't think this bug impacts everyone, it is more an issue if you load data separating right and left sides into two separate records, but I expect there are cases in the normal right-left side data where only one side is populated that might cause a problem anyway I have loaded all of Tiger 2010 into an instance here: http://imaptools.com:8080/geocode/ I have loaded all of Tiger 2010 using the sqlite datastore option which creates a 27 GB database for the data for the 3300 Census counties and equivalents. Please give it a try and let me know if you find any issues with it. Enjoy, -Steve |
From: Stephen W. <wo...@sw...> - 2011-03-08 20:16:01
|
Hi all, As promised: http://imaptools.com:8080/geocode/ OK, so as I noted before there is still a bug in PAGC that Walter and I need to find a fix related to including the Postal Code. It causes a high geocode failure rate, but PAGC finds things fine without the Postal Code so Just leave it out for now. Or try with it and remove it if it fails. Here are some stats: It took about 5-6 days to do the Tiger join and load it into PAGC. I think part of this time is because I have a Postgresql autovacuum process running which was creating a lot of disk contention. Which, BTW, are still running: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 21975 postgres 20 0 125m 4320 692 S 1 0.1 132:37.13 postgres 21976 postgres 20 0 72952 3556 464 S 1 0.1 766:16.90 postgres and impacts on the response of anything try to do disk I/O, including the URL above. woodbri@mappy:~$ ls /u/data/tgr2010-geo-v2/streets/ -lh total 28G -rw-r--r-- 1 woodbri woodbri 255M 2011-03-08 03:20 build_log.err -rw-r--r-- 1 woodbri woodbri 286K 2011-03-08 03:20 last_append_log.err -rw-r--r-- 1 woodbri woodbri 307M 2011-03-08 03:20 Streets.ix3 -rw-r--r-- 1 woodbri woodbri 27G 2011-03-08 03:20 Streets.pagc So the SQLite database is 27 GB and the Streets.ix3 is 307 MB woodbri@mappy:/u/data/tgr2010-geo-v2/streets$ sqlite3 Streets.pagc SQLite version 3.5.9 Enter ".help" for instructions sqlite> .schema CREATE TABLE ix0 ( pagc_key text, pagc_data integer ); CREATE TABLE ix1 ( pagc_key text, pagc_data integer ); CREATE TABLE ix2 ( pagc_key text, pagc_data integer ); CREATE TABLE ix5 ( pagc_key integer, pagc_data blob ); CREATE TABLE pagc_cfg ( pagc_key text, pagc_data blob ); CREATE TABLE pgx ( pagc_key integer, pagc_data text ); CREATE INDEX 'x_ix0' ON 'ix0' ( pagc_key ); CREATE INDEX 'x_ix1' ON 'ix1' ( pagc_key ); CREATE INDEX 'x_ix2' ON 'ix2' ( pagc_key ); CREATE INDEX 'x_ix5' ON 'ix5' ( pagc_key ); CREATE INDEX 'x_pgx' ON 'pgx' ( pagc_key ); sqlite> select count(*) from ix0; 100792158 sqlite> select count(*) from pgx; 50583769 Yikes! a whole lot of something in those tables! -Steve |
From: Stephen W. <wo...@sw...> - 2011-03-05 01:43:25
|
Hi all, I just wanted to give an update on my experiences working with the datastore branch in SVN. First for those not on the devel list, I have merged single line parseaddress branch into the datastore branch, so all the new features are now integrated into branches/datastore. I say all, but there may be some missing code from the landmarks branch as I found geocode_response on datastore had not been merged so I merged that with my own changes. If you find stuff missing or broken please let us know and Walter I and will try to attend to it. Ok next, I'm really loving the new datastore branch. I build mine using shapefiles as input and sqlite as the datastore and plain files for the cache. I explain how I built stuff below. So I'm trying to build PAGC and a geocode response that will have the whole US loaded into it. Working with Tiger that mean you have to load one county and then append to that about 3300 more counties. Needless to say this is time consuming especially if your process blows up along the way! Ah, more than once! One of the very cool benefits of using SQLite is that it is an ACID compliant database which means that it protects the database at all costs and keeps it in a stable usable state, even when it crashes! So some 200 appends and 4 hours into the load I ran across a file that caused pagc_build_schema to SEGV and crash. I'm now thinking I have to spend 4 hrs to rebuild the database up to the file before the crash just so I can debug it. But wait, I say, why not see what happens if I just rerun the command that crashed. So I start up the debugger, rerun the command, and it appears to be hung. :( I ^C out of it and there is a long delay and it eventually returns, and I exit the debugger, look at the file system again and notice the sqlite journal file is no longer there and the database is slightly smaller than before. Oh, the big delay, like it was hung, was SQLite using the journal file to ROLLBACK the transaction that was open when it crashed, restoring the database back to a usable state. Cool, I get back into the debugger and rerun the command and it runs and SEGVs like before and I can now see what is broken. The shapefile had a shape that was too big for a fixed size buffer, I up the buffer size and rebuild PAGC. Now when I rerun the command it does the rollback again because I crashed it again and then completes without a failure and the database is still in good shape! So I restart my script appending to the existing database and another 6 hours into loading data, my tiger join process barfs and causes pagc_build_schema to die again. I little more debugging, some bug fixes and we are off and running again! I currently have 8.4 GB SQLite database which contains 836 or 3300 counties and the load continues. This resiliency is what makes SQLite such an awesome database and it is very fast and has zero management so it is ideal for deploying to production sites. Walter has done a great job with the datastore branch. I look forward to check out some of the other options that he has built into the code. I'm giving my config a major workout as we speak loading data and will post the geocode_response URL when its ready for some testing. OK, here is how I built stuff cd branches/datastore/ds ./configure --with-norm=s --with-raw=d --with-sch=d --with-cache=f make sudo make install sudo ldconfig cd ../pagclib ./configure make sudo make install sudo ldconfig cd ../parseaddress ./configure make sudo make install sudo ldconfig cd ../geocode_response ## edit data_cap.h as required ./configure --with-parseaddress # add other options you want make sudo make install sudo ldconfig sudo cp geocode_response /usr/lib/cgi-bin/geocode_response This is great stuff, I hope you enjoy it as much as I am - Well maybe I enjoy this just a little too much ;) -Steve |
From: Treichel, K. (MDOR) <Ken...@st...> - 2010-12-10 23:14:54
|
Kent Treichel responded; I posted the original answer. The values for the US come from the USPS. They have many non-exclusive datasets, (http://www.usps.com/ncsc/) which are licensable. For about a hundred bucks, you can get the ZIPs with a lat long centroid. You can also license the data yourself in most cases. For ZIP codes specifically, there are 42,000 unique ZIPS and about 80,000 place names associated with those ZIPs. Each ZIP has a preferred name and up to 12 additional place names. In Minnesota, the ZIPs with the greatest number of names are mostly rural areas where post offices might have closed and been aggregated. If your address is in Canada, I have no idea how to get that data. The downside to USPS place names is that they have their own spelling convention which doesn't change or does, but only when they want to change it. There is also a abbreviated name which is limited to 13 characters, I think, for fitting on an envelope. The veracity of the previous statements has not been checked by the USPS. -----Original Message----- From: The PAGC users mailing list [mailto:pag...@li...] Sent: Wednesday, December 08, 2010 8:19 PM To: pag...@li...; pag...@li... Subject: Re: [Pagc-users] PlaceName_USPS tags On 12/2/2010 1:01 PM, The PAGC users mailing list wrote: > I have a reasonable guess and am prepared to be proven wrong. > PlaceName would be the city/town/civil division in which the point is > located. PlaceName_USPS would be the name used by the US Postal > Service to deliver the mail, or more correctly, one or more names > associated with a particular ZIP code and postal sub station. This > is why suburbs are sometimes not listed as the city for mailing. I > can provide examples if this is less than helpful. Not sure who responded here. I understood what PlaceName_USPS meant, that was not really my question(s): Where do the values come from? Since I loaded the data and I did not define these tags in my data or schema, my questions is why are they here as opposed to just using PlaceName. Where did the values come from? Is there some hidden table of USPS names somewhere? What happens if my address is in Canada? Is there a mapping of all 45,000+ USPS zipcodes to they USPS preferred city names? Inquiring minds want to know :) This might be more of a question for the dev list so cross posting there. Sorry for the duplicate copies. Best regards, -Steve > -----Original Message----- From: The PAGC users mailing list > [mailto:pag...@li...] Sent: Wednesday, December > 01, 2010 9:18 PM To: pag...@li... Subject: > [Pagc-users] PlaceName_USPS tags > > Hi, > > I have noticed PlaceName_USPS tags in the xml and I noticed that > there are both PlaceName and PlaceName_USPS defined in the XSD. The > Documentation was less the forth coming with answers as to how these > differ and why one is used over the other? Also, if we have > PlaceName_USPS data in the reference data how that might get linked > to that tag. > > Sorry to bombard the list tonight! > > Thanks, -Steve ------------------------------------------------------------------------------ This SF Dev2Dev email is sponsored by: WikiLeaks The End of the Free Internet http://p.sf.net/sfu/therealnews-com _______________________________________________ Pagc-users mailing list Pag...@li... https://lists.sourceforge.net/lists/listinfo/pagc-users |
From: The P. u. m. l. <pag...@li...> - 2010-12-09 09:45:20
|
Hi Steve -- for some reason I'm not on the pagc-users list so I didn't see your original post. I can answer this question. The MetroGIS data contains a separate field for the USPS place name (and this is the reason for the CHAR_ALT construct for CITY). It is a matter for the response rather than the request and serves to increase the likelihood of a correct match on the place name field. The query however would have just one place name. The xsd is just defining the format of the response. Thanks, <= walter => > On 12/2/2010 1:01 PM, The PAGC users mailing list wrote: >> I have a reasonable guess and am prepared to be proven wrong. >> PlaceName would be the city/town/civil division in which the point is >> located. PlaceName_USPS would be the name used by the US Postal >> Service to deliver the mail, or more correctly, one or more names >> associated with a particular ZIP code and postal sub station. This >> is why suburbs are sometimes not listed as the city for mailing. I >> can provide examples if this is less than helpful. > > Not sure who responded here. I understood what PlaceName_USPS meant, > that was not really my question(s): > > Where do the values come from? > > Since I loaded the data and I did not define these tags in my data or > schema, my questions is why are they here as opposed to just using > PlaceName. > > Where did the values come from? > > Is there some hidden table of USPS names somewhere? > What happens if my address is in Canada? > > Is there a mapping of all 45,000+ USPS zipcodes to they USPS preferred > city names? > > Inquiring minds want to know :) This might be more of a question for the > dev list so cross posting there. Sorry for the duplicate copies. > > Best regards, > -Steve > >> -----Original Message----- From: The PAGC users mailing list >> [mailto:pag...@li...] Sent: Wednesday, December >> 01, 2010 9:18 PM To: pag...@li... Subject: >> [Pagc-users] PlaceName_USPS tags >> >> Hi, >> >> I have noticed PlaceName_USPS tags in the xml and I noticed that >> there are both PlaceName and PlaceName_USPS defined in the XSD. The >> Documentation was less the forth coming with answers as to how these >> differ and why one is used over the other? Also, if we have >> PlaceName_USPS data in the reference data how that might get linked >> to that tag. >> >> Sorry to bombard the list tonight! >> >> Thanks, -Steve > > ------------------------------------------------------------------------------ > This SF Dev2Dev email is sponsored by: > > WikiLeaks The End of the Free Internet > http://p.sf.net/sfu/therealnews-com > _______________________________________________ > Pagc-devel mailing list > Pag...@li... > https://lists.sourceforge.net/lists/listinfo/pagc-devel > |
From: The P. u. m. l. <pag...@li...> - 2010-12-09 02:43:56
|
On 12/2/2010 1:01 PM, The PAGC users mailing list wrote: > I have a reasonable guess and am prepared to be proven wrong. > PlaceName would be the city/town/civil division in which the point is > located. PlaceName_USPS would be the name used by the US Postal > Service to deliver the mail, or more correctly, one or more names > associated with a particular ZIP code and postal sub station. This > is why suburbs are sometimes not listed as the city for mailing. I > can provide examples if this is less than helpful. Not sure who responded here. I understood what PlaceName_USPS meant, that was not really my question(s): Where do the values come from? Since I loaded the data and I did not define these tags in my data or schema, my questions is why are they here as opposed to just using PlaceName. Where did the values come from? Is there some hidden table of USPS names somewhere? What happens if my address is in Canada? Is there a mapping of all 45,000+ USPS zipcodes to they USPS preferred city names? Inquiring minds want to know :) This might be more of a question for the dev list so cross posting there. Sorry for the duplicate copies. Best regards, -Steve > -----Original Message----- From: The PAGC users mailing list > [mailto:pag...@li...] Sent: Wednesday, December > 01, 2010 9:18 PM To: pag...@li... Subject: > [Pagc-users] PlaceName_USPS tags > > Hi, > > I have noticed PlaceName_USPS tags in the xml and I noticed that > there are both PlaceName and PlaceName_USPS defined in the XSD. The > Documentation was less the forth coming with answers as to how these > differ and why one is used over the other? Also, if we have > PlaceName_USPS data in the reference data how that might get linked > to that tag. > > Sorry to bombard the list tonight! > > Thanks, -Steve |
From: The P. u. m. l. <pag...@li...> - 2010-12-02 18:02:25
|
I have a reasonable guess and am prepared to be proven wrong. PlaceName would be the city/town/civil division in which the point is located. PlaceName_USPS would be the name used by the US Postal Service to deliver the mail, or more correctly, one or more names associated with a particular ZIP code and postal sub station. This is why suburbs are sometimes not listed as the city for mailing. I can provide examples if this is less than helpful. -----Original Message----- From: The PAGC users mailing list [mailto:pag...@li...] Sent: Wednesday, December 01, 2010 9:18 PM To: pag...@li... Subject: [Pagc-users] PlaceName_USPS tags Hi, I have noticed PlaceName_USPS tags in the xml and I noticed that there are both PlaceName and PlaceName_USPS defined in the XSD. The Documentation was less the forth coming with answers as to how these differ and why one is used over the other? Also, if we have PlaceName_USPS data in the reference data how that might get linked to that tag. Sorry to bombard the list tonight! Thanks, -Steve ------------------------------------------------------------------------------ Increase Visibility of Your 3D Game App & Earn a Chance To Win $500! Tap into the largest installed PC base & get more eyes on your game by optimizing for Intel(R) Graphics Technology. Get started today with the Intel(R) Software Partner Program. Five $500 cash prizes are up for grabs. http://p.sf.net/sfu/intelisp-dev2dev _______________________________________________ Pagc-users mailing list Pag...@li... https://lists.sourceforge.net/lists/listinfo/pagc-users |
From: The P. u. m. l. <pag...@li...> - 2010-12-02 03:18:30
|
Hi, I have noticed PlaceName_USPS tags in the xml and I noticed that there are both PlaceName and PlaceName_USPS defined in the XSD. The Documentation was less the forth coming with answers as to how these differ and why one is used over the other? Also, if we have PlaceName_USPS data in the reference data how that might get linked to that tag. Sorry to bombard the list tonight! Thanks, -Steve |
From: The P. u. m. l. <pag...@li...> - 2010-12-02 02:39:04
|
Hi all, I have gotten feed back from a few people on this. and have been able to fix most of the problems reported. http://imaptools.com:8080/parseaddress/ Please try and trip it up and forward to me addresses that are not parsed correctly. I am going to start the process of integrating this into PAGC geocode_response. Thanks, -Steve W |
From: The P. u. m. l. <pag...@li...> - 2010-12-02 02:26:39
|
Hi all, I am working on loading all of Tiger 2009 into PAGC. I have made some progress with the most awesome support from Walter to help me debug issues and fix bugs. So here is a link to my instance that only has Massachusetts loaded at the moment. This was loaded by appending each county to the existing instance. Tiger has something like 3300 counties for the whole thing. So I only have about 3290 more counties to load :) http://imaptools.com:8080/geocode/ Feel free to play with it and report any issues back to me. The current data is all loaded as single sided records as opposed to left-right sided records. That will come later when I have more time to tweak my tiger join utility so it creates left-right sided records. This means that in the output you may find multiple records with the same addressIdentifier tag because there is one for each address range on that street. IE: one or more for the left side and one or more for the right side. Walter, Here is a problem that I can not explain. These two queries are identical except the first include the zipcode and the second does not. The second does report the zipcode in the results, but if you add it to the input it fails to find the record. http://imaptools.com:8080/cgi-bin/gc-tiger?methodName=GeocodeRequest&Version=1.1&InterpolationOffset=10.0&CompleteAddressNumber=101&CompleteStreetName=Newbury+St&PlaceName=Boston&StateName=MA&ZipCode=02116&CountryCode=US&ResponseFormat=XML http://imaptools.com:8080/cgi-bin/gc-tiger?methodName=GeocodeRequest&Version=1.1&InterpolationOffset=10.0&CompleteAddressNumber=101&CompleteStreetName=Newbury+St&PlaceName=Boston&StateName=MA&ZipCode=&CountryCode=US&ResponseFormat=XML Thanks, -Steve Woodbridge |
From: The P. u. m. l. <pag...@li...> - 2010-12-01 20:55:33
|
Hi Walter and all, I'm running the landmark branch and I'm having trouble getting geocode response to work as documented, or as I interpret the documentation, which is likely wrong ;) My directory structure is: woodbri@mappy:/u/data/tgr2009-geo-v1$ ls * testgc standard: featwords.csv gazeteer.csv lexicon.csv tgr2pagc-single.dbf streets: build_log.err last_append_log.err Streets.ix2 Streets.pgc __db.001 Streets.ix0 Streets.ix3 Streets.pgx __db.002 Streets.ix1 Streets.ix5 and for deployment I'll put geocode_response in cgi-bin along with symbolic links to it like gc-tiger, gc-navteq, gc-whatever and these names will be configured in apache to work with different data directories. My data_cap.h file looks like: #define NUM_LANDMARK_SCHEMAS 1 #define NUM_MATCH_SCHEMAS 1 #define HAVE_SITE_INTERPOLATED //#define HAVE_SITE_ADDRESS_PRECISE //#define HAVE_SITE_LANDMARK #ifdef HAVE_SITE_INTERPOLATED PAGC_DATA_CAP interp_data_cap = { "Streets" , "streets" , "PAGC_STREETS_PATH" , "Streets" , "INTERPOLATED" } ; #endif So for test and debug purposes I have a simple script I can run from the command line like: --- this should be Strategy 2 in the initalization code --- #!/bin/sh #export PAGC_STREETS_PATH=/u/data/tgr2009-geo-v1/streets export PAGC_DATA_PATH=/u/data/tgr2009-geo-v1 #export PAGC_STAND_PATH=/u/data/tgr2009-geo-v1/standard #cd /u/data/tgr2009-geo-v1/ /home/woodbri/work/pagc/pagc/branches/landmarks/geocode_response/geocode_response_f woodbri@mappy:~/work/pagc/pagc/branches/landmarks/geocode_response$ /u/data/tgr2009-geo-v1/testgc Could not open site interpolate schema Can't open: streets/Streets.pgc ERROR: Could not restore the schema streets/Streets --- OR this should be strategy 2 in the initialization code --- #!/bin/sh export PAGC_STREETS_PATH=/u/data/tgr2009-geo-v1/streets export PAGC_DATA_PATH=/u/data/tgr2009-geo-v1 export PAGC_STAND_PATH=/u/data/tgr2009-geo-v1/standard #cd /u/data/tgr2009-geo-v1/ /home/woodbri/work/pagc/pagc/branches/landmarks/geocode_response/geocode_response_f woodbri@mappy:~/work/pagc/pagc/branches/landmarks/geocode_response$ /u/data/tgr2009-geo-v1/testgc Could not open site interpolate schema Can't open: /u/data/tgr2009-geo-v1/streets.pgc ERROR: Could not restore the schema /u/data/tgr2009-geo-v1/streets --- OR this is strategy 1 in the initialization code and it appears to work --- #!/bin/sh #export PAGC_STREETS_PATH=/u/data/tgr2009-geo-v1/streets #export PAGC_DATA_PATH=/u/data/tgr2009-geo-v1 #export PAGC_STAND_PATH=/u/data/tgr2009-geo-v1/standard cd /u/data/tgr2009-geo-v1/ /home/woodbri/work/pagc/pagc/branches/landmarks/geocode_response/geocode_response_f woodbri@mappy:~/work/pagc/pagc/branches/landmarks/geocode_response$ /u/data/tgr2009-geo-v1/testgc Content-type: text/xml <?xml version="1.0" encoding="UTF-8" ?> <GeocodeService xmlns="http://www.metrogis.org/geocode" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.metrogis.org/geocode GeocodeService.xsd" xmlns:gml="http://www.opengis.net/gml"> <ResponseHeader /> <Response Version="1.1" RequestID=""> <GeocodeResponse><GeocodeResponseList numberOfGeocodedAddresses="0"> </GeocodeResponseList> </GeocodeResponse> <responseFaultList numberOfFaults="2"> <fault> <faultcode>Server</faultcode> <faultstring>Server process error</faultstring> <detail>No request method received</detail> </fault> <fault> <faultcode>Server</faultcode> <faultstring>Server process error</faultstring> <detail>retrieve_CGI_request: Could not retrieve</detail> </fault></responseFaultList> </Response></GeocodeService> |
From: The P. u. m. l. <pag...@li...> - 2010-11-11 04:01:30
|
Dan and all, I made some more changes. http://imaptools.com:8080/parseaddress/ this works better for most cases, send me cases that fail. As I have indicated, this kind of parsing is very ambiguous at best. I have made the parsing more aggressive about finding a city name, but now it does not like to parse a strings without a city, like: num street postcode It works in a lot of cases but not always. You can click the links on the test page to run those cases or enter your own test cases. Once I get the correct versions of Berekly DB loaded so I can compile PAGC, I can start to look at integrating this. Walter, Any thoughts or guidance on integration would be appreciated. The test code looks like this: stH = (HHash *) calloc(1, sizeof(struct hsearch_data)); err = load_state_hash(stH); if (err) { printf("Error: Failed to load state hash (%d)\n", err); exit(1); } result = parseaddress(stH, subject, &err); if (result) { printf(" Num: %s\n", result->num); printf(" Street: %s\n", result->street); printf(" Address1: %s\n", result->address1); printf(" City: %s\n", result->city); printf(" State: %s\n", result->st); printf(" Zipcode: %s\n", result->zip); printf(" Zip plus: %s\n", result->zipplus); printf(" Country: %s\n", result->cc); printf(" Latitude: %.6f\n", result->lat); printf("Longitude: %.6f\n", result->lon); } else printf("Input failed to parse!\n"); free_address(result); free_state_hash(stH); Thanks, -Steve On 11/1/2010 2:58 PM, Dan Putler wrote: > Hi Steve, > > Yep, Trail is the name of the town. I dug up an address from Trail just > because it seemed like one that could cause problems. Another one that > came to mind, which is for a hypothetical example (the street does not > actually exist) is: > > 123 St Olaf St > St Paul, MN > > Dan > > On 11/01/2010 11:18 AM, Stephen Woodbridge wrote: >> On 11/1/2010 1:22 PM, Dan Putler wrote: >> >>> Hi Steve, >>> >>> This is great! I've put in some Vancouver addresses as well, and they >>> are parsed correctly, down to recognizing that if the "state" equals BC, >>> then county equals Canada. >>> >>> One thing that threw it off, which is not surprising, is the following >>> address that I took directly from a tourism site for Trail, BC (it is >>> for "The Corporation of the City of Trail"). The address is: >>> >>> 1394 Pine Ave Trail, BC V1R 4E6 >>> >>> Your parser produces the following: >>> >>> Num: 1394 >>> Street: Pine >>> Address1: 1394 Pine >>> City: Ave Trail >>> State: BC >>> Zipcode: V1R 4E6 >>> Zip plus: (null) >>> Country: CA >>> Latitude: 0.000000 >>> Longitude: 0.000000 >>> >> OK, is Trail the name of the city? I use a HUGE regex to detect suffix >> street types, but not sure what this one is doing. I'll add it to the >> queue to look at. This is probably just a bug. >> >> >>> which is interesting since I thought it would have problems with the >>> address, I just didn't know what they would be. I ran the same address >>> through the Google, Yahoo, and Bing parsers, and they all handled it. My >>> guess is that they all take advantage of the their lexicons in parsing >>> the address string. Would it be possible to use the lexicon to determine >>> if the third token in the address matches a street type value, and if it >>> does, make it the street type, and if it doesn't, make it part of the >>> street name? This would correctly pickup "Ave", and would likely result >>> in "Trail" being picked up correctly as the city (which is tricky, since >>> "Trail" is also a valid street type). >>> >>> The next thing I looked at was apartment/unit/suite issues. Canada Post >>> wants apartment/unit/suite appended in front of the street number, with >>> a "-" character separating the unit number from the street number. For >>> example 395-2771 Main St, Vancouver, BC, where 395 would be the unit >>> number, and 2771 is the street number. This seems like it would be >>> fairly easy to parse, and I think most Canadians tend to use this >>> address format. The USPS address standard of this address is 2771 Main >>> St Unit 395, Vancouver, BC (see >>> http://pe.usps.com/text/pub28/pub28c2.html ). I don't think Americans >>> tend to use this addresses form left to their own devices, and most web >>> and paper forms that collect address information tend to want to ask for >>> apartment/unit/suite identifiers as a second address line, which allows >>> them to easily put addresses into the USPS standard. >>> >> I really do not try to recognize unit numbers. based on the various ways >> that this is treated it can be before the number like in Canada, or >> appended to the address line, etc. If you read the section on this in >> pub28 you can see it is very complex. The code I have would be pretty >> easy to add recognizer for this kind of things. One problem with this is >> that in the mid-west there is a US house numbering scheme called block >> address which is pn1-n1-pn2-n2-sn2 which I do recognize and in fact fits >> your example "395-2771" which is why that is put into the number field. >> Obvious like the behavior could be conditional based on the country, but >> it is not the case at the moment. >> >> >>> Num: 395-2771 >>> Street: Main St >>> Address1: 395-2771 Main St >>> City: Vancouver >>> State: BC >>> Zipcode: (null) >>> Zip plus: (null) >>> Country: CA >>> Latitude: 0.000000 >>> Longitude: 0.000000 >>> >>> >> >> >>> My thoughts at this point. All in all, I think you've done great work on >>> this Steve. >>> >> Thanks, it is a start. I appreciate your feedback and examples that are >> broken are useful and help me improve this. >> >> -Steve >> >> >>> Dan >>> >>> On 11/01/2010 08:51 AM, Stephen Woodbridge wrote: >>> >>>> Hi all, >>>> >>>> I have been working on a single line address parser that we can >>>> hopefully integrate into PAGC in the near future. I also plan to make a >>>> version available that is can be called via an Ajax request to get the >>>> parsed results back into a web client so that the client can give the >>>> user an opportunity to fix the parsing if the geocoding fails. The code >>>> is written in C but I have written a simple PHP wrapper so you can play >>>> with it in it current state. >>>> >>>> http://imaptools.com:8080/parseaddress/ >>>> >>>> Feedback is welcome. >>>> >>>> I'll try to put together more information and collect my thoughts on >>>> this shortly but I have a few other tasks that come first and mostly >>>> want to get this out for some feedback and to give the community a heads >>>> up that this is coming. >>>> >>>> Thanks, >>>> -Steve >>>> >>>> ------------------------------------------------------------------------------ >>>> Nokia and AT&T present the 2010 Calling All Innovators-North America contest >>>> Create new apps& games for the Nokia N8 for consumers in U.S. and Canada >>>> $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing >>>> Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store >>>> http://p.sf.net/sfu/nokia-dev2dev >>>> _______________________________________________ >>>> Pagc-devel mailing list >>>> Pag...@li... >>>> https://lists.sourceforge.net/lists/listinfo/pagc-devel >>>> >>>> >>> >>> ------------------------------------------------------------------------------ >>> Nokia and AT&T present the 2010 Calling All Innovators-North America contest >>> Create new apps& games for the Nokia N8 for consumers in U.S. and Canada >>> $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing >>> Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store >>> http://p.sf.net/sfu/nokia-dev2dev >>> _______________________________________________ >>> Pagc-devel mailing list >>> Pag...@li... >>> https://lists.sourceforge.net/lists/listinfo/pagc-devel >>> >> >> ------------------------------------------------------------------------------ >> Nokia and AT&T present the 2010 Calling All Innovators-North America contest >> Create new apps& games for the Nokia N8 for consumers in U.S. and Canada >> $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing >> Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store >> http://p.sf.net/sfu/nokia-dev2dev >> _______________________________________________ >> Pagc-devel mailing list >> Pag...@li... >> https://lists.sourceforge.net/lists/listinfo/pagc-devel >> > > > ------------------------------------------------------------------------------ > Nokia and AT&T present the 2010 Calling All Innovators-North America contest > Create new apps& games for the Nokia N8 for consumers in U.S. and Canada > $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing > Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store > http://p.sf.net/sfu/nokia-dev2dev > _______________________________________________ > Pagc-devel mailing list > Pag...@li... > https://lists.sourceforge.net/lists/listinfo/pagc-devel |
From: The P. u. m. l. <pag...@li...> - 2010-11-01 15:51:28
|
Hi all, I have been working on a single line address parser that we can hopefully integrate into PAGC in the near future. I also plan to make a version available that is can be called via an Ajax request to get the parsed results back into a web client so that the client can give the user an opportunity to fix the parsing if the geocoding fails. The code is written in C but I have written a simple PHP wrapper so you can play with it in it current state. http://imaptools.com:8080/parseaddress/ Feedback is welcome. I'll try to put together more information and collect my thoughts on this shortly but I have a few other tasks that come first and mostly want to get this out for some feedback and to give the community a heads up that this is coming. Thanks, -Steve |
From: The P. u. m. l. <pag...@li...> - 2010-01-03 16:06:07
|
Hey Folks, Just wanting to probe the community to see what platforms and binaries may be of interest. Of course currently there are three install methods * source * windows installer * osgeo4w installer (http://trac.osgeo.org/osgeo4w/) (Cygwin based) I am interested in what other platforms may be of interest. I am currently a Ubuntu user so will be exploring the Personal package Archives (PPA) to provide Ubuntu Aptitude support. There is already good support for all the dependencies on that platform. I have also worked with various other distros and could try my hand at RPM's or other packaging schemes. let me know what you are your clients might like to have access to. Cheers |
From: The P. u. m. l. <pag...@li...> - 2009-12-01 06:30:46
|
Hey Folks, First off a special thanks to Walter Sinclair for pointing me in the right direction, without which I would still be in the land of C getting lost. Thanks to Stephen for referring me to the mapserver mapscript SWIG interface for an example, I learned that geospatial projects are still "special" because the solutions for geo are never the same as "hello world" solutions. A big thanks to Dan Putler who encouraged me to keep going with trying to find a pythonic solution for exposing the pagc library. Knowing there are people that want to use a solution is sometimes what is needed to keep ploughing ahead. I figured out a solution, quite clean infact, using pythons distutils. Distutils is a python toolkit used when people want to distribute their code. Using Distutils allowed me to move away from a manual process of creating a wrapper, compiling the resultant file and then linking libraries. Three different stages where something can go wrong. Distutils, generaly relies on a setup.py files initiated from the command line as follows: python setup.py build_ext --inplace Please note that the bellow solution assumes you have a compiled and installed version of libpagc. I do plan to merge the new files back into the SVN trunk for every ones benefit. however, for the record please find the solution bellow. Then I plan to add these bindings to the OSGEO4W platform. OS: Ubuntu 9.10 PAGC: svn revision 138 Python: 2.6.4 ===============setup.py================== #!/usr/bin/env python """ setup.py file for PAGC SWIG interface """ from distutils.core import setup, Extension pagc_client_module = Extension('_pagc_client', sources=['pagc_client.i'], swig_opts=['-I/usr/include','-I/usr/local/include'], library_dirs=['/usr/local/lib','/usr/lib'], libraries=['pagc'] ) setup (name = 'pagc_client', version = '0.1', author = "PAGC Project: Dave Sampson", description = """Simple swig extension to pagc library""", ext_modules = [pagc_client_module], py_modules = ["pagc_client"], ) ================== Good luck. |
From: The P. u. m. l. <pag...@li...> - 2009-08-19 18:43:34
|
I am note sure what the issue was but I went back and rebuilt the index and now I get similar results to Steve's online example... Cheers > -----Original Message----- > From: The PAGC users mailing list > [mailto:pag...@li...] > Sent: Wednesday, August 19, 2009 12:46 > To: pag...@li... > Subject: Re: [Pagc-users] NO MATCH on Ottawa request > > Ahh a Canadian live sample. GREAT.. This helps me with > comparisons and validating what I am doing. > > The response lines up almost perfectly with google map (of > course people are always going to compare.... > > A couple of tests and I am sure now something is amis. > > If I try 381 dovercourt ave in Ottawa on my local machine I > get 10 results that are all over the place and hardly a match > on google. On your service I get 1 result that is a pretty > close google match... > > The resulting goolge matches are not a big thing for me right > now but the fact your service works well quells my fears. > > questions: > 1. did you re-project the RNF file? (I tried mine as WGS84 > but will probably go back to the default) 2. Does projection > matter from an indexing point of view? > 3. For the placenames did you embed that into the attribute > tables of the RNF or did you build a pseudo parcels layer to > encapsulate place names? (data source?) 4. did you combine > all the provincial RNF's together into one or did you just > grab the full RNF? > 5. Could you run the service using a seprate dataset for each > province eg. <dataSource>streets-ontario</dataSource>. Would > this increase performance? > > Cheers > > > -----Original Message----- > > From: The PAGC users mailing list > > [mailto:pag...@li...] > > Sent: Wednesday, August 19, 2009 11:49 > > To: pag...@li... > > Subject: Re: [Pagc-users] NO MATCH on Ottawa request > > > > I have all of Canada loaded into PAGC here. This is based > on StatCan > > RNF with some additional work that I have done with it. > > > > Here is a simple form to access it: > > http://imaptools.com/pagc.html > > > > Here is the response I get for "3 Falk Ave, Ottawa, ON" > > http://imaptools.com/cgi-bin/geocode-CA?CompleteAddressNumber= > > 3&CompleteStreetName=falk+ave&PlaceName=ottawa&StateName=ON&Re > > > sponseFormat=XML&methodName=GeocodeRequest&Version=1.1&CountryCode=CA > > > > If you get an exact match then you do not also get fuzzy > results. So > > for example if you enter: "3 Falk, Ottawa, ON" > > ie: leave off the type you get: > > http://imaptools.com/cgi-bin/geocode-CA?CompleteAddressNumber= > > 3&CompleteStreetName=falk&PlaceName=ottawa&StateName=ON&Respon > > seFormat=XML&methodName=GeocodeRequest&Version=1.1&CountryCode=CA > > > > -Steve > > http://imaptools.com/ > > > > The PAGC users mailing list wrote: > > > > > > > > > Hey Folks, > > > > > > Just wanted to try an Ottawa street name and number match > > > (interpolated) and I feel the response is not correct. > > > > > > 1. I used the Ontario RNF file ( > > > > > > http://www.statcan.gc.ca/bsolc/olc-cel/olc-cel?catno=92-500-G&lang=eng > > > > > > <http://www.statcan.gc.ca/bsolc/olc-cel/olc-cel?catno=92-500-G&lang=en > > > g> ) 2. I used the following geomatry to clip Ontario to issolate > > > Ottawa () > > > > > > POLYGON ((-76.102095409501359 > 45.53588445601013,-75.238960733557604 > > > 45.5348701 > > > 96696446,-75.238960733557604 > 45.091638876617218,-76.102095409501359 > > > 45.091638876 > > > 617218,-76.102095409501359 45.53588445601013)) > > > > > > 3. I recompiled PAGC to work only on INTERPOLATION and removed > > > unnecesary Schemas 4. Build the schema using the INTERNAL > > pagc schemas > > > for Stats Can RNF. > > > 5. I submitted the following request > > > > > > > > > http://127.0.0.1/cgi-bin/geocode_response.exe?methodName=GeocodeReques > > > > > > t&Version=1.1&CompleteAddressNumber=3&CompleteStreetName=Falk+Ave&Plac > > > eName=Anywhere&CountryCode=CA > > > > > > <http://127.0.0.1/cgi-bin/geocode_response.exe?methodName=GeocodeReque > > > > > > st&Version=1.1&CompleteAddressNumber=3&CompleteStreetName=Falk+Ave&Pla > > > ceName=Anywhere&CountryCode=CA> > > > > > > 6. received the following response (see bellow > > > > > > 7. here is the results of the attrrbute table for the road I was > > > expecting to get back. > > > > > > RB_UID,C,80 NAME,C,80 TYPE,C,80 DIRECTION,C,80 > > > ADDR_FM_LE,N,32,3 ADDR_TO_LE,N,32,3 > > ADDR_FM_RG,N,32,3 > > > ADDR_TO_RG,N,32,3 > > > > > > 2055809 Falk AVE > > 7 > > > 21 12 28 > > > > > > 2055844 Falk AVE > > 1 > > > 5 4 10 > > > > > > This was another one I would have expected to be returned > > > > > > 2037911 Falkirk CRES > > 1967 > > > 2031 1960 > 2032 > > > > > > Any ideas why these three segments were missed? > > > > > > Cheers > > > > > > > > > ================API Response====================== <?xml > > version="1.0" > > > encoding="UTF-8" ?> > > > > > > <GeocodeService xmlns="http://www.metrogis.org/geocode" > > > xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" > > > xsi:schemaLocation="http://www.metrogis.org/geocode > > GeocodeService.xsd" > > > xmlns:gml="http://www.opengis.net/gml"> > > > > > > <ResponseHeader /> > > > > > > <Response Version="1.1" RequestID=""> > > > > > > <GeocodeResponse><GeocodeResponseList > numberOfGeocodedAddresses="4"> > > > > > > <GeocodedAddress> > > > > > > <Address> > > > > > > <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> > > > <CompleteStreetName><StreetName>FABLE</StreetName> > > > > > > <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Addres > > > s> > > > > > > <gml:Point><gml:pos>-75.764288 45.274075</gml:pos></gml:Point> > > > > > > <GeocodeMatchCode accuracy="0.788518" matchType="INTERPOLATED" > > > note=""/> > > > > > > > > > <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentif > > > ier></addressIdentifier></source></GeocodedAddress> > > > > > > <GeocodedAddress> > > > > > > <Address> > > > > > > <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> > > > <CompleteStreetName><StreetName>FRANK</StreetName> > > > > > > <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Addres > > > s> > > > > > > <gml:Point><gml:pos>-75.681606 45.417662</gml:pos></gml:Point> > > > > > > <GeocodeMatchCode accuracy="0.782426" matchType="INTERPOLATED" > > > note=""/> > > > > > > > > > <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentif > > > ier></addressIdentifier></source></GeocodedAddress> > > > > > > <GeocodedAddress> > > > > > > <Address> > > > > > > <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> > > > <CompleteStreetName><StreetName>FOULIS</StreetName> > > > > > > <PostType>CRESCENT</PostType></CompleteStreetName></SiteAddress></Addr > > > ess> > > > > > > <gml:Point><gml:pos>-75.893134 45.296146</gml:pos></gml:Point> > > > > > > <GeocodeMatchCode accuracy="0.718459" matchType="INTERPOLATED" > > > note="P"/> > > > > > > > > > <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentif > > > ier></addressIdentifier></source></GeocodedAddress> > > > > > > <GeocodedAddress> > > > > > > <Address> > > > > > > <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> > > > <CompleteStreetName><StreetName>PHYLIS</StreetName> > > > > > > <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Addres > > > s> > > > > > > <gml:Point><gml:pos>-75.771141 45.277526</gml:pos></gml:Point> > > > > > > <GeocodeMatchCode accuracy="0.688865" matchType="INTERPOLATED" > > > note=""/> > > > > > > > > > <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentif > > > ier></addressIdentifier></source></GeocodedAddress> > > > > > > </GeocodeResponseList> > > > > > > </GeocodeResponse> > > > > > > <requestedAddress> > > > > > > <CompleteAddressNumber>3</CompleteAddressNumber> > > > <CompleteStreetName>Falk Ave</CompleteStreetName> > > > <PlaceName>Anywhere</PlaceName></requestedAddress> > > > > > > </Response></GeocodeService> > > > > > > > > > > > > ---------------------------------------------------------------------- > > > -- > > > > > > > > > ---------------------------------------------------------------------- > > > -------- Let Crystal Reports handle the reporting - Free Crystal > > > Reports 2008 30-Day trial. Simplify your report design, > integration > > > and deployment - and focus on what you do best, core application > > > coding. Discover what's new with Crystal Reports now. > > > http://p.sf.net/sfu/bobj-july > > > > > > > > > > > > ---------------------------------------------------------------------- > > > -- > > > > > > _______________________________________________ > > > Pagc-users mailing list > > > Pag...@li... > > > https://lists.sourceforge.net/lists/listinfo/pagc-users > > > > > > -------------------------------------------------------------- > > ---------------- > > Let Crystal Reports handle the reporting - Free Crystal > Reports 2008 > > 30-Day trial. Simplify your report design, integration and > deployment > > - and focus on what you do best, core application coding. Discover > > what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july > > _______________________________________________ > > Pagc-users mailing list > > Pag...@li... > > https://lists.sourceforge.net/lists/listinfo/pagc-users > > > > -------------------------------------------------------------- > ---------------- > Let Crystal Reports handle the reporting - Free Crystal > Reports 2008 30-Day trial. Simplify your report design, > integration and deployment - and focus on what you do best, > core application coding. Discover what's new with Crystal > Reports now. http://p.sf.net/sfu/bobj-july > _______________________________________________ > Pagc-users mailing list > Pag...@li... > https://lists.sourceforge.net/lists/listinfo/pagc-users > |
From: The P. u. m. l. <pag...@li...> - 2009-08-19 16:46:26
|
Ahh a Canadian live sample. GREAT.. This helps me with comparisons and validating what I am doing. The response lines up almost perfectly with google map (of course people are always going to compare.... A couple of tests and I am sure now something is amis. If I try 381 dovercourt ave in Ottawa on my local machine I get 10 results that are all over the place and hardly a match on google. On your service I get 1 result that is a pretty close google match... The resulting goolge matches are not a big thing for me right now but the fact your service works well quells my fears. questions: 1. did you re-project the RNF file? (I tried mine as WGS84 but will probably go back to the default) 2. Does projection matter from an indexing point of view? 3. For the placenames did you embed that into the attribute tables of the RNF or did you build a pseudo parcels layer to encapsulate place names? (data source?) 4. did you combine all the provincial RNF's together into one or did you just grab the full RNF? 5. Could you run the service using a seprate dataset for each province eg. <dataSource>streets-ontario</dataSource>. Would this increase performance? Cheers > -----Original Message----- > From: The PAGC users mailing list > [mailto:pag...@li...] > Sent: Wednesday, August 19, 2009 11:49 > To: pag...@li... > Subject: Re: [Pagc-users] NO MATCH on Ottawa request > > I have all of Canada loaded into PAGC here. This is based on > StatCan RNF with some additional work that I have done with it. > > Here is a simple form to access it: > http://imaptools.com/pagc.html > > Here is the response I get for "3 Falk Ave, Ottawa, ON" > http://imaptools.com/cgi-bin/geocode-CA?CompleteAddressNumber= > 3&CompleteStreetName=falk+ave&PlaceName=ottawa&StateName=ON&Re > sponseFormat=XML&methodName=GeocodeRequest&Version=1.1&CountryCode=CA > > If you get an exact match then you do not also get fuzzy > results. So for example if you enter: "3 Falk, Ottawa, ON" > ie: leave off the type you get: > http://imaptools.com/cgi-bin/geocode-CA?CompleteAddressNumber= > 3&CompleteStreetName=falk&PlaceName=ottawa&StateName=ON&Respon > seFormat=XML&methodName=GeocodeRequest&Version=1.1&CountryCode=CA > > -Steve > http://imaptools.com/ > > The PAGC users mailing list wrote: > > > > > > Hey Folks, > > > > Just wanted to try an Ottawa street name and number match > > (interpolated) and I feel the response is not correct. > > > > 1. I used the Ontario RNF file ( > > > http://www.statcan.gc.ca/bsolc/olc-cel/olc-cel?catno=92-500-G&lang=eng > > > <http://www.statcan.gc.ca/bsolc/olc-cel/olc-cel?catno=92-500-G&lang=en > > g> ) 2. I used the following geomatry to clip Ontario to issolate > > Ottawa () > > > > POLYGON ((-76.102095409501359 45.53588445601013,-75.238960733557604 > > 45.5348701 > > 96696446,-75.238960733557604 45.091638876617218,-76.102095409501359 > > 45.091638876 > > 617218,-76.102095409501359 45.53588445601013)) > > > > 3. I recompiled PAGC to work only on INTERPOLATION and removed > > unnecesary Schemas 4. Build the schema using the INTERNAL > pagc schemas > > for Stats Can RNF. > > 5. I submitted the following request > > > > > http://127.0.0.1/cgi-bin/geocode_response.exe?methodName=GeocodeReques > > > t&Version=1.1&CompleteAddressNumber=3&CompleteStreetName=Falk+Ave&Plac > > eName=Anywhere&CountryCode=CA > > > <http://127.0.0.1/cgi-bin/geocode_response.exe?methodName=GeocodeReque > > > st&Version=1.1&CompleteAddressNumber=3&CompleteStreetName=Falk+Ave&Pla > > ceName=Anywhere&CountryCode=CA> > > > > 6. received the following response (see bellow > > > > 7. here is the results of the attrrbute table for the road I was > > expecting to get back. > > > > RB_UID,C,80 NAME,C,80 TYPE,C,80 DIRECTION,C,80 > > ADDR_FM_LE,N,32,3 ADDR_TO_LE,N,32,3 > ADDR_FM_RG,N,32,3 > > ADDR_TO_RG,N,32,3 > > > > 2055809 Falk AVE > 7 > > 21 12 28 > > > > 2055844 Falk AVE > 1 > > 5 4 10 > > > > This was another one I would have expected to be returned > > > > 2037911 Falkirk CRES > 1967 > > 2031 1960 2032 > > > > Any ideas why these three segments were missed? > > > > Cheers > > > > > > ================API Response====================== <?xml > version="1.0" > > encoding="UTF-8" ?> > > > > <GeocodeService xmlns="http://www.metrogis.org/geocode" > > xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" > > xsi:schemaLocation="http://www.metrogis.org/geocode > GeocodeService.xsd" > > xmlns:gml="http://www.opengis.net/gml"> > > > > <ResponseHeader /> > > > > <Response Version="1.1" RequestID=""> > > > > <GeocodeResponse><GeocodeResponseList numberOfGeocodedAddresses="4"> > > > > <GeocodedAddress> > > > > <Address> > > > > <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> > > <CompleteStreetName><StreetName>FABLE</StreetName> > > > <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Addres > > s> > > > > <gml:Point><gml:pos>-75.764288 45.274075</gml:pos></gml:Point> > > > > <GeocodeMatchCode accuracy="0.788518" matchType="INTERPOLATED" > > note=""/> > > > > > <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentif > > ier></addressIdentifier></source></GeocodedAddress> > > > > <GeocodedAddress> > > > > <Address> > > > > <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> > > <CompleteStreetName><StreetName>FRANK</StreetName> > > > <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Addres > > s> > > > > <gml:Point><gml:pos>-75.681606 45.417662</gml:pos></gml:Point> > > > > <GeocodeMatchCode accuracy="0.782426" matchType="INTERPOLATED" > > note=""/> > > > > > <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentif > > ier></addressIdentifier></source></GeocodedAddress> > > > > <GeocodedAddress> > > > > <Address> > > > > <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> > > <CompleteStreetName><StreetName>FOULIS</StreetName> > > > <PostType>CRESCENT</PostType></CompleteStreetName></SiteAddress></Addr > > ess> > > > > <gml:Point><gml:pos>-75.893134 45.296146</gml:pos></gml:Point> > > > > <GeocodeMatchCode accuracy="0.718459" matchType="INTERPOLATED" > > note="P"/> > > > > > <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentif > > ier></addressIdentifier></source></GeocodedAddress> > > > > <GeocodedAddress> > > > > <Address> > > > > <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> > > <CompleteStreetName><StreetName>PHYLIS</StreetName> > > > <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Addres > > s> > > > > <gml:Point><gml:pos>-75.771141 45.277526</gml:pos></gml:Point> > > > > <GeocodeMatchCode accuracy="0.688865" matchType="INTERPOLATED" > > note=""/> > > > > > <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentif > > ier></addressIdentifier></source></GeocodedAddress> > > > > </GeocodeResponseList> > > > > </GeocodeResponse> > > > > <requestedAddress> > > > > <CompleteAddressNumber>3</CompleteAddressNumber> > > <CompleteStreetName>Falk Ave</CompleteStreetName> > > <PlaceName>Anywhere</PlaceName></requestedAddress> > > > > </Response></GeocodeService> > > > > > > > ---------------------------------------------------------------------- > > -- > > > > > ---------------------------------------------------------------------- > > -------- Let Crystal Reports handle the reporting - Free Crystal > > Reports 2008 30-Day trial. Simplify your report design, integration > > and deployment - and focus on what you do best, core application > > coding. Discover what's new with Crystal Reports now. > > http://p.sf.net/sfu/bobj-july > > > > > > > ---------------------------------------------------------------------- > > -- > > > > _______________________________________________ > > Pagc-users mailing list > > Pag...@li... > > https://lists.sourceforge.net/lists/listinfo/pagc-users > > > -------------------------------------------------------------- > ---------------- > Let Crystal Reports handle the reporting - Free Crystal > Reports 2008 30-Day trial. Simplify your report design, > integration and deployment - and focus on what you do best, > core application coding. Discover what's new with Crystal > Reports now. http://p.sf.net/sfu/bobj-july > _______________________________________________ > Pagc-users mailing list > Pag...@li... > https://lists.sourceforge.net/lists/listinfo/pagc-users > |
From: The P. u. m. l. <pag...@li...> - 2009-08-19 15:49:08
|
I have all of Canada loaded into PAGC here. This is based on StatCan RNF with some additional work that I have done with it. Here is a simple form to access it: http://imaptools.com/pagc.html Here is the response I get for "3 Falk Ave, Ottawa, ON" http://imaptools.com/cgi-bin/geocode-CA?CompleteAddressNumber=3&CompleteStreetName=falk+ave&PlaceName=ottawa&StateName=ON&ResponseFormat=XML&methodName=GeocodeRequest&Version=1.1&CountryCode=CA If you get an exact match then you do not also get fuzzy results. So for example if you enter: "3 Falk, Ottawa, ON" ie: leave off the type you get: http://imaptools.com/cgi-bin/geocode-CA?CompleteAddressNumber=3&CompleteStreetName=falk&PlaceName=ottawa&StateName=ON&ResponseFormat=XML&methodName=GeocodeRequest&Version=1.1&CountryCode=CA -Steve http://imaptools.com/ The PAGC users mailing list wrote: > > > Hey Folks, > > Just wanted to try an Ottawa street name and number match (interpolated) > and I feel the response is not correct. > > 1. I used the Ontario RNF file ( > http://www.statcan.gc.ca/bsolc/olc-cel/olc-cel?catno=92-500-G&lang=eng > <http://www.statcan.gc.ca/bsolc/olc-cel/olc-cel?catno=92-500-G&lang=eng> ) > 2. I used the following geomatry to clip Ontario to issolate Ottawa () > > POLYGON ((-76.102095409501359 45.53588445601013,-75.238960733557604 > 45.5348701 > 96696446,-75.238960733557604 45.091638876617218,-76.102095409501359 > 45.091638876 > 617218,-76.102095409501359 45.53588445601013)) > > 3. I recompiled PAGC to work only on INTERPOLATION and removed > unnecesary Schemas > 4. Build the schema using the INTERNAL pagc schemas for Stats Can RNF. > 5. I submitted the following request > > http://127.0.0.1/cgi-bin/geocode_response.exe?methodName=GeocodeRequest&Version=1.1&CompleteAddressNumber=3&CompleteStreetName=Falk+Ave&PlaceName=Anywhere&CountryCode=CA > <http://127.0.0.1/cgi-bin/geocode_response.exe?methodName=GeocodeRequest&Version=1.1&CompleteAddressNumber=3&CompleteStreetName=Falk+Ave&PlaceName=Anywhere&CountryCode=CA> > > 6. received the following response (see bellow > > 7. here is the results of the attrrbute table for the road I was > expecting to get back. > > RB_UID,C,80 NAME,C,80 TYPE,C,80 DIRECTION,C,80 > ADDR_FM_LE,N,32,3 ADDR_TO_LE,N,32,3 ADDR_FM_RG,N,32,3 > ADDR_TO_RG,N,32,3 > > 2055809 Falk AVE 7 > 21 12 28 > > 2055844 Falk AVE 1 > 5 4 10 > > This was another one I would have expected to be returned > > 2037911 Falkirk CRES 1967 > 2031 1960 2032 > > Any ideas why these three segments were missed? > > Cheers > > > ================API Response====================== > <?xml version="1.0" encoding="UTF-8" ?> > > <GeocodeService xmlns="http://www.metrogis.org/geocode" > xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" > xsi:schemaLocation="http://www.metrogis.org/geocode GeocodeService.xsd" > xmlns:gml="http://www.opengis.net/gml"> > > <ResponseHeader /> > > <Response Version="1.1" RequestID=""> > > <GeocodeResponse><GeocodeResponseList numberOfGeocodedAddresses="4"> > > <GeocodedAddress> > > <Address> > > <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> > <CompleteStreetName><StreetName>FABLE</StreetName> > <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Address> > > <gml:Point><gml:pos>-75.764288 45.274075</gml:pos></gml:Point> > > <GeocodeMatchCode accuracy="0.788518" matchType="INTERPOLATED" note=""/> > > <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentifier></addressIdentifier></source></GeocodedAddress> > > <GeocodedAddress> > > <Address> > > <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> > <CompleteStreetName><StreetName>FRANK</StreetName> > <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Address> > > <gml:Point><gml:pos>-75.681606 45.417662</gml:pos></gml:Point> > > <GeocodeMatchCode accuracy="0.782426" matchType="INTERPOLATED" note=""/> > > <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentifier></addressIdentifier></source></GeocodedAddress> > > <GeocodedAddress> > > <Address> > > <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> > <CompleteStreetName><StreetName>FOULIS</StreetName> > <PostType>CRESCENT</PostType></CompleteStreetName></SiteAddress></Address> > > <gml:Point><gml:pos>-75.893134 45.296146</gml:pos></gml:Point> > > <GeocodeMatchCode accuracy="0.718459" matchType="INTERPOLATED" note="P"/> > > <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentifier></addressIdentifier></source></GeocodedAddress> > > <GeocodedAddress> > > <Address> > > <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> > <CompleteStreetName><StreetName>PHYLIS</StreetName> > <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Address> > > <gml:Point><gml:pos>-75.771141 45.277526</gml:pos></gml:Point> > > <GeocodeMatchCode accuracy="0.688865" matchType="INTERPOLATED" note=""/> > > <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentifier></addressIdentifier></source></GeocodedAddress> > > </GeocodeResponseList> > > </GeocodeResponse> > > <requestedAddress> > > <CompleteAddressNumber>3</CompleteAddressNumber> > <CompleteStreetName>Falk Ave</CompleteStreetName> > <PlaceName>Anywhere</PlaceName></requestedAddress> > > </Response></GeocodeService> > > > ------------------------------------------------------------------------ > > ------------------------------------------------------------------------------ > Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day > trial. Simplify your report design, integration and deployment - and focus on > what you do best, core application coding. Discover what's new with > Crystal Reports now. http://p.sf.net/sfu/bobj-july > > > ------------------------------------------------------------------------ > > _______________________________________________ > Pagc-users mailing list > Pag...@li... > https://lists.sourceforge.net/lists/listinfo/pagc-users |
From: The P. u. m. l. <pag...@li...> - 2009-08-19 15:40:45
|
Hey Folks, Just wanted to try an Ottawa street name and number match (interpolated) and I feel the response is not correct. 1. I used the Ontario RNF file (http://www.statcan.gc.ca/bsolc/olc-cel/olc-cel?catno=92-500-G&lang=eng) 2. I used the following geomatry to clip Ontario to issolate Ottawa () POLYGON ((-76.102095409501359 45.53588445601013,-75.238960733557604 45.5348701 96696446,-75.238960733557604 45.091638876617218,-76.102095409501359 45.091638876 617218,-76.102095409501359 45.53588445601013)) 3. I recompiled PAGC to work only on INTERPOLATION and removed unnecesary Schemas 4. Build the schema using the INTERNAL pagc schemas for Stats Can RNF. 5. I submitted the following request http://127.0.0.1/cgi-bin/geocode_response.exe?methodName=GeocodeRequest& Version=1.1&CompleteAddressNumber=3&CompleteStreetName=Falk+Ave&PlaceNam e=Anywhere&CountryCode=CA 6. received the following response (see bellow 7. here is the results of the attrrbute table for the road I was expecting to get back. RB_UID,C,80 NAME,C,80 TYPE,C,80 DIRECTION,C,80 ADDR_FM_LE,N,32,3 ADDR_TO_LE,N,32,3 ADDR_FM_RG,N,32,3 ADDR_TO_RG,N,32,3 2055809 Falk AVE 7 21 12 28 2055844 Falk AVE 1 5 4 10 This was another one I would have expected to be returned 2037911 Falkirk CRES 1967 2031 1960 2032 Any ideas why these three segments were missed? Cheers ================API Response====================== <?xml version="1.0" encoding="UTF-8" ?> <GeocodeService xmlns="http://www.metrogis.org/geocode" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.metrogis.org/geocode GeocodeService.xsd" xmlns:gml="http://www.opengis.net/gml"> <ResponseHeader /> <Response Version="1.1" RequestID=""> <GeocodeResponse><GeocodeResponseList numberOfGeocodedAddresses="4"> <GeocodedAddress> <Address> <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> <CompleteStreetName><StreetName>FABLE</StreetName> <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Address> <gml:Point><gml:pos>-75.764288 45.274075</gml:pos></gml:Point> <GeocodeMatchCode accuracy="0.788518" matchType="INTERPOLATED" note=""/> <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentifie r></addressIdentifier></source></GeocodedAddress> <GeocodedAddress> <Address> <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> <CompleteStreetName><StreetName>FRANK</StreetName> <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Address> <gml:Point><gml:pos>-75.681606 45.417662</gml:pos></gml:Point> <GeocodeMatchCode accuracy="0.782426" matchType="INTERPOLATED" note=""/> <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentifie r></addressIdentifier></source></GeocodedAddress> <GeocodedAddress> <Address> <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> <CompleteStreetName><StreetName>FOULIS</StreetName> <PostType>CRESCENT</PostType></CompleteStreetName></SiteAddress></Addres s> <gml:Point><gml:pos>-75.893134 45.296146</gml:pos></gml:Point> <GeocodeMatchCode accuracy="0.718459" matchType="INTERPOLATED" note="P"/> <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentifie r></addressIdentifier></source></GeocodedAddress> <GeocodedAddress> <Address> <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> <CompleteStreetName><StreetName>PHYLIS</StreetName> <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Address> <gml:Point><gml:pos>-75.771141 45.277526</gml:pos></gml:Point> <GeocodeMatchCode accuracy="0.688865" matchType="INTERPOLATED" note=""/> <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentifie r></addressIdentifier></source></GeocodedAddress> </GeocodeResponseList> </GeocodeResponse> <requestedAddress> <CompleteAddressNumber>3</CompleteAddressNumber> <CompleteStreetName>Falk Ave</CompleteStreetName> <PlaceName>Anywhere</PlaceName></requestedAddress> </Response></GeocodeService> |
From: The P. u. m. l. <pag...@li...> - 2009-08-19 15:39:33
|
Hey Folks, Just wanted to try an Ottawa street name and number match (interpolated) and I feel the response is not correct. 1. I used the Ontario RNF file (http://www.statcan.gc.ca/bsolc/olc-cel/olc-cel?catno=92-500-G&lang=eng) 2. I used the following geomatry to clip Ontario to issolate Ottawa () POLYGON ((-76.102095409501359 45.53588445601013,-75.238960733557604 45.5348701 96696446,-75.238960733557604 45.091638876617218,-76.102095409501359 45.091638876 617218,-76.102095409501359 45.53588445601013)) 3. I recompiled PAGC to work only on INTERPOLATION and removed unnecesary Schemas 4. Build the schema using the INTERNAL pagc schemas for Stats Can RNF. 5. I submitted the following request http://127.0.0.1/cgi-bin/geocode_response.exe?methodName=GeocodeRequest& Version=1.1&CompleteAddressNumber=3&CompleteStreetName=Falk+Ave&PlaceNam e=Anywhere&CountryCode=CA 6. received the following response (see bellow 7. here is the results of the attrrbute table for the road I was expecting to get back. RB_UID,C,80 NAME,C,80 TYPE,C,80 DIRECTION,C,80 ADDR_FM_LE,N,32,3 ADDR_TO_LE,N,32,3 ADDR_FM_RG,N,32,3 ADDR_TO_RG,N,32,3 2055809 Falk AVE 7 21 12 28 2055844 Falk AVE 1 5 4 10 This was another one I would have expected to be returned 2037911 Falkirk CRES 1967 2031 1960 2032 Any ideas why these three segments were missed? Cheers ================API Response====================== <?xml version="1.0" encoding="UTF-8" ?> <GeocodeService xmlns="http://www.metrogis.org/geocode" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.metrogis.org/geocode GeocodeService.xsd" xmlns:gml="http://www.opengis.net/gml"> <ResponseHeader /> <Response Version="1.1" RequestID=""> <GeocodeResponse><GeocodeResponseList numberOfGeocodedAddresses="4"> <GeocodedAddress> <Address> <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> <CompleteStreetName><StreetName>FABLE</StreetName> <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Address> <gml:Point><gml:pos>-75.764288 45.274075</gml:pos></gml:Point> <GeocodeMatchCode accuracy="0.788518" matchType="INTERPOLATED" note=""/> <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentifie r></addressIdentifier></source></GeocodedAddress> <GeocodedAddress> <Address> <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> <CompleteStreetName><StreetName>FRANK</StreetName> <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Address> <gml:Point><gml:pos>-75.681606 45.417662</gml:pos></gml:Point> <GeocodeMatchCode accuracy="0.782426" matchType="INTERPOLATED" note=""/> <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentifie r></addressIdentifier></source></GeocodedAddress> <GeocodedAddress> <Address> <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> <CompleteStreetName><StreetName>FOULIS</StreetName> <PostType>CRESCENT</PostType></CompleteStreetName></SiteAddress></Addres s> <gml:Point><gml:pos>-75.893134 45.296146</gml:pos></gml:Point> <GeocodeMatchCode accuracy="0.718459" matchType="INTERPOLATED" note="P"/> <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentifie r></addressIdentifier></source></GeocodedAddress> <GeocodedAddress> <Address> <SiteAddress><CompleteAddressNumber>3</CompleteAddressNumber> <CompleteStreetName><StreetName>PHYLIS</StreetName> <PostType>STREET</PostType></CompleteStreetName></SiteAddress></Address> <gml:Point><gml:pos>-75.771141 45.277526</gml:pos></gml:Point> <GeocodeMatchCode accuracy="0.688865" matchType="INTERPOLATED" note=""/> <source><dataSource>pagc_data_ottawa_wgs84</dataSource><addressIdentifie r></addressIdentifier></source></GeocodedAddress> </GeocodeResponseList> </GeocodeResponse> <requestedAddress> <CompleteAddressNumber>3</CompleteAddressNumber> <CompleteStreetName>Falk Ave</CompleteStreetName> <PlaceName>Anywhere</PlaceName></requestedAddress> </Response></GeocodeService> |
From: The P. u. m. l. <pag...@li...> - 2009-08-13 20:04:16
|
Hey Folks, I think I finaly came across my issue. I am not running pagc with a site match (aka parcels) schema because I have no parcels info. I came across the following section in the docs: http://www.pagcgeo.org/docs/html/geocode_service.html#customization-of-t he-geocoder-software Aparently I need to remove the fourth line bellow in data_cap.h #define NUM_LANDMARK_SCHEMAS 1 #define NUM_MATCH_SCHEMAS 1 #define HAVE_SITE_INTERPOLATED #define HAVE_SITE_ADDRESS_PRECISE Then re-compile. May I suggest is future versions that this not be dependent on recompiling the code. Perhaps a PAGC.conf file (maybe even an xml file) could be used to define what schemas will be used. PAGC should be universaly accessible to a broad range of users. Even though I can ./configure Make Make install I would much rather not too. For instance, you don't have to recompile open office to use another spreadsheet, or recompile mpaserver to pump out a new map. Just a thought. In the mean time I will try my hand at recompiling, this is scary though. Cheers |
From: The P. u. m. l. <pag...@li...> - 2009-08-12 18:59:19
|
I was able to build with the internal schemas without passing -s switch and it completed. Now when I run the geocode_response.exe from within osge4w I get the following access log entry. 127.0.0.1 - - [12/Aug/2009:14:34:01 -0400] "GET /cgi-bin/geocode_response.exe HTTP/1.1" 500 538 127.0.0.1 - - [12/Aug/2009:14:34:56 -0400] "GET /cgi-bin/geocode_response.exe?methodName=GeocodeRequest&Version=1.1&Requ estStrategy=Interpolated&CompleteAddressNumber=3&CompleteStreetName=falk +ave&CountryCode=CA&ResponseFormat=XML HTTP/1.1" 500 538 And the following apache error log entry [Wed Aug 12 14:34:01 2009] [error] [client 127.0.0.1] Premature end of script headers: geocode_response.exe [Wed Aug 12 14:34:01 2009] [error] [client 127.0.0.1] Could not open site match schema\r\r [Wed Aug 12 14:34:01 2009] [error] [client 127.0.0.1] \r [Wed Aug 12 14:34:01 2009] [error] [client 127.0.0.1] Can't open: C:\\OSGeo4W\\bin\\parcels\\parcels_all7_points_lat83_point.pgc\r [Wed Aug 12 14:34:01 2009] [error] [client 127.0.0.1] \r [Wed Aug 12 14:34:01 2009] [error] [client 127.0.0.1] ERROR: Could not restore the schema C:\\OSGeo4W\\bin\\parcels\\parcels_all7_points_lat83_point\r [Wed Aug 12 14:34:56 2009] [error] [client 127.0.0.1] Premature end of script headers: geocode_response.exe [Wed Aug 12 14:34:56 2009] [error] [client 127.0.0.1] Could not open site match schema\r\r [Wed Aug 12 14:34:56 2009] [error] [client 127.0.0.1] \r [Wed Aug 12 14:34:56 2009] [error] [client 127.0.0.1] Can't open: C:\\OSGeo4W\\bin\\parcels\\parcels_all7_points_lat83_point.pgc\r [Wed Aug 12 14:34:56 2009] [error] [client 127.0.0.1] \r [Wed Aug 12 14:34:56 2009] [error] [client 127.0.0.1] ERROR: Could not restore the schema C:\\OSGeo4W\\bin\\parcels\\parcels_all7_points_lat83_point\r I do not have a parcels file for many reasons. I am have passed RequestStrategy=Interpolated to ensure that no parcels are required for either the BOTH or PRECISE strategies... I feel like I am close but missing something small. cheers > -----Original Message----- > From: The PAGC users mailing list > [mailto:pag...@li...] > Sent: Wednesday, August 12, 2009 14:35 > To: pag...@li... > Subject: Re: [Pagc-users] Gocode_response without parcels....? > > You need to build a schema.dbf file that reflects your street > data set. > For example I created this one: > > $ dbfdump ca-schema2.dbf > FLRNFRE::::: > FLSTATS::::: > FLZBLNK::::: > CITY:CHAR_SINGLE:AC4::: > HOUSE:NUMBER_INTERVAL_LEFT_RIGHT:L_REFADDR:L_NREFADDR:R_REFADD > R:R_NREFADDR > PROV:CHAR_SINGLE:AC2::: > STREET:CHAR_SINGLE:NAME::: > SUFTYP:CHAR_SINGLE:TYPE::: > SUFDIR:CHAR_SINGLE:DIRECTION::: > > which reflects my shapefiles that I wanted to load. You will > likely have to build one for your data. > > -Steve W > > The PAGC users mailing list wrote: > > > > > > Hey Folks, > > > > Getting a little further with setting up the new version of PAGC. > > > > I do not have aprcels shape file set as I know of no free versions > > here in Canada. > > > > So I did not build that particular schema. > > > > The webservice is returning the following errors from error.log > > > > [Wed Aug 12 13:27:12 2009] [error] [client 127.0.0.1] > Premature end of > > script headers: geocode_response.exe [Wed Aug 12 13:27:12 2009] > > [error] [client 127.0.0.1] Could not open site match > schema\r\r [Wed > > Aug 12 13:27:12 2009] [error] [client 127.0.0.1] \r [Wed Aug 12 > > 13:27:12 2009] [error] [client 127.0.0.1] Can't open: > > C:\\OSGeo4W\\bin\\parcels\\parcels_all7_points_lat83_point.pgc\r > > > > [Wed Aug 12 13:27:12 2009] [error] [client 127.0.0.1] \r > [Wed Aug 12 > > 13:27:12 2009] [error] [client 127.0.0.1] ERROR: Could not > restore the > > schema C:\\OSGeo4W\\bin\\parcels\\parcels_all7_points_lat83_point\r > > > > It looks like it is expecting a parcels schema, however > there is not > > going to be one. > > > > How Can I make sure it skips this so the geocode_response actualy > > returns something > > > > Cheers > > > > > > > ---------------------------------------------------------------------- > > -- > > > > > ---------------------------------------------------------------------- > > -------- Let Crystal Reports handle the reporting - Free Crystal > > Reports 2008 30-Day trial. Simplify your report design, integration > > and deployment - and focus on what you do best, core application > > coding. Discover what's new with Crystal Reports now. > > http://p.sf.net/sfu/bobj-july > > > > > > > ---------------------------------------------------------------------- > > -- > > > > _______________________________________________ > > Pagc-users mailing list > > Pag...@li... > > https://lists.sourceforge.net/lists/listinfo/pagc-users > > > -------------------------------------------------------------- > ---------------- > Let Crystal Reports handle the reporting - Free Crystal > Reports 2008 30-Day trial. Simplify your report design, > integration and deployment - and focus on what you do best, > core application coding. Discover what's new with Crystal > Reports now. http://p.sf.net/sfu/bobj-july > _______________________________________________ > Pagc-users mailing list > Pag...@li... > https://lists.sourceforge.net/lists/listinfo/pagc-users > |
From: The P. u. m. l. <pag...@li...> - 2009-08-12 18:34:26
|
You need to build a schema.dbf file that reflects your street data set. For example I created this one: $ dbfdump ca-schema2.dbf FLRNFRE::::: FLSTATS::::: FLZBLNK::::: CITY:CHAR_SINGLE:AC4::: HOUSE:NUMBER_INTERVAL_LEFT_RIGHT:L_REFADDR:L_NREFADDR:R_REFADDR:R_NREFADDR PROV:CHAR_SINGLE:AC2::: STREET:CHAR_SINGLE:NAME::: SUFTYP:CHAR_SINGLE:TYPE::: SUFDIR:CHAR_SINGLE:DIRECTION::: which reflects my shapefiles that I wanted to load. You will likely have to build one for your data. -Steve W The PAGC users mailing list wrote: > > > Hey Folks, > > Getting a little further with setting up the new version of PAGC. > > I do not have aprcels shape file set as I know of no free versions here > in Canada. > > So I did not build that particular schema. > > The webservice is returning the following errors from error.log > > [Wed Aug 12 13:27:12 2009] [error] [client 127.0.0.1] Premature end of > script headers: geocode_response.exe > [Wed Aug 12 13:27:12 2009] [error] [client 127.0.0.1] Could not open > site match schema\r\r > [Wed Aug 12 13:27:12 2009] [error] [client 127.0.0.1] \r > [Wed Aug 12 13:27:12 2009] [error] [client 127.0.0.1] Can't open: > C:\\OSGeo4W\\bin\\parcels\\parcels_all7_points_lat83_point.pgc\r > > [Wed Aug 12 13:27:12 2009] [error] [client 127.0.0.1] \r > [Wed Aug 12 13:27:12 2009] [error] [client 127.0.0.1] ERROR: Could not > restore the schema > C:\\OSGeo4W\\bin\\parcels\\parcels_all7_points_lat83_point\r > > It looks like it is expecting a parcels schema, however there is not > going to be one. > > How Can I make sure it skips this so the geocode_response actualy > returns something > > Cheers > > > ------------------------------------------------------------------------ > > ------------------------------------------------------------------------------ > Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day > trial. Simplify your report design, integration and deployment - and focus on > what you do best, core application coding. Discover what's new with > Crystal Reports now. http://p.sf.net/sfu/bobj-july > > > ------------------------------------------------------------------------ > > _______________________________________________ > Pagc-users mailing list > Pag...@li... > https://lists.sourceforge.net/lists/listinfo/pagc-users |