Re: [NEW] geo/libpostal

2018-07-30 Thread Frederic Cambus
On Mon, Jul 09, 2018 at 12:07:51AM +0200, Frederic Cambus wrote:

> Finally, here is a new tarball fixing and improving a couple of things:
> 
> - Set AUTOCONF_VERSION and AUTOMAKE_VERSION so it builds
> - Add a _libpostal group so users added to that group can easily fetch
>   data files
> - The data files are now hosted on GitHub, update MESSAGE so users fetch
>   the files from there directly, instead of using libpostal_data which
>   doesn't work anymore and spits out lots of errors
> - Do not install libpostal_data as it is currently broken
> - Mention in MESSAGE how much space is needed for the data files
> 
> New tarball attached, and diff for infrastructure/db/user.list below.

Ping. Anyone willing to look at this?

It can easily be tested with the bundled address_parser program:

> Berkeley, CA, USA

Result:

{
  "city": "berkeley",
  "state": "ca",
  "country": "usa"
}



Re: [NEW] geo/libpostal

2018-07-08 Thread Frederic Cambus
On Mon, Mar 05, 2018 at 06:08:12PM +, Stuart Henderson wrote:

> > C library for parsing/normalizing street addresses around the world, powered
> > by statistical NLP and open geo data.
> > 
> > This library helps convert the free-form addresses that humans use into 
> > clean
> > normalized forms suitable for machine comparison and full-text indexing.
> > 
> > Comments? OK?
> 
> Build fails early, AUTOCONF_VERSION etc aren't set when calling bootstrap.sh.
> 
> I'd also create ${LOCALSTATEDIR}/db/libpostal via @sample in PLIST, and
> it should have an @extra or @extraunexec so that pkg_delete -c removes
> the downloaded files.
> 
> Maybe provide an @newuser to own the state directory and change the command
> line in MESSAGE to use it, you just know people are going to run the fetcher
> as root otherwise.

Finally, here is a new tarball fixing and improving a couple of things:

- Set AUTOCONF_VERSION and AUTOMAKE_VERSION so it builds
- Add a _libpostal group so users added to that group can easily fetch
  data files
- The data files are now hosted on GitHub, update MESSAGE so users fetch
  the files from there directly, instead of using libpostal_data which
  doesn't work anymore and spits out lots of errors
- Do not install libpostal_data as it is currently broken
- Mention in MESSAGE how much space is needed for the data files

New tarball attached, and diff for infrastructure/db/user.list below.

Comments? OK?

Index: infrastructure/db/user.list
===
RCS file: /cvs/ports/infrastructure/db/user.list,v
retrieving revision 1.318
diff -u -p -r1.318 user.list
--- infrastructure/db/user.list 15 Jun 2018 09:29:03 -  1.318
+++ infrastructure/db/user.list 8 Jul 2018 19:39:37 -
@@ -321,3 +321,4 @@ id  usergroup   port options
 810 _sabnzbd   _sabnzbdnews/sabnzbd
 811_openvpnusers   net/openvpn_bsdauth
 812 _mactelnet _mactelnet  net/mac-telnet
+813_libpostal  geo/libpostal


libpostal.tar.gz
Description: application/tar-gz


Re: [NEW] geo/libpostal

2018-03-05 Thread Stuart Henderson
On 2018/03/05 14:50, Frederic Cambus wrote:
> Hi ports@,
> 
> Here is a new port: geo/libpostal
> 
> From DESCR:
> 
> C library for parsing/normalizing street addresses around the world, powered
> by statistical NLP and open geo data.
> 
> This library helps convert the free-form addresses that humans use into clean
> normalized forms suitable for machine comparison and full-text indexing.
> 
> Comments? OK?

Build fails early, AUTOCONF_VERSION etc aren't set when calling bootstrap.sh.

I'd also create ${LOCALSTATEDIR}/db/libpostal via @sample in PLIST, and
it should have an @extra or @extraunexec so that pkg_delete -c removes
the downloaded files.

Maybe provide an @newuser to own the state directory and change the command
line in MESSAGE to use it, you just know people are going to run the fetcher
as root otherwise.



[NEW] geo/libpostal

2018-03-05 Thread Frederic Cambus
Hi ports@,

Here is a new port: geo/libpostal

>From DESCR:

C library for parsing/normalizing street addresses around the world, powered
by statistical NLP and open geo data.

This library helps convert the free-form addresses that humans use into clean
normalized forms suitable for machine comparison and full-text indexing.

Comments? OK?


libpostal.tar.gz
Description: application/tar-gz