Address Standardization using yahoo & google

[Blake]

In light of hit or miss performance of the USPS address standardization and its limited scope to the US, couldn't yahoo or google APIs be used for a similar functionality?

I easily modified the geocoding files and update script to insert 9 digit zip codes from yahoo and it works great (below), but it would be nice to have that full address standardization functionality built in. This might also be of more use to the non-US crowd.

From civicrm/CRM/Utils/Geocode/Yahoo.php

Code: [Select]

133            $values['geo_code_1'] = $ret['Latitude' ];
134            $values['geo_code_2'] = $ret['Longitude'];
              $values['postal_code'] = $ret['Zip'];
136            return true;

and from civicrm/bin/UpdateAddress.php

Code: [Select]

156            $address->geo_code_1 = $params['geo_code_1'];
157            $address->geo_code_2 = $params['geo_code_2'];
              $address->postal_code = $params['postal_code'];
158            $address->save( );


[robinhood]

Agree that this would be a good addition especially for non-US addresses.  But what is hit or miss about the USPS address standardization?  Works perfectly here.

[Blake]

Getting approval from USPS to use the service isn't automatic, as detailed by Paul Fisher in http://civicrm.org/node/627

[TallDavid]

RobinHood, are you successfully using the USPS Address Standardization with the USPS Production (not the testing) server?  I've received USPS authorization to use the Production system, but, can't get CiviCRM to work on these transactions.  If you have been successful using the USPS Production server, what version of CiviCRM are you using?

[noah]

I easily modified the geocoding files and update script to insert 9 digit zip codes from yahoo and it works great (below), but it would be nice to have that full address standardization functionality built in. This might also be of more use to the non-US crowd.

I made these modifications and they work fine (well, the first one works fine -- I haven't had occasion to test the second). I've been wanting this functionality for so long, it's great! Here's my slightly changed version, which tries to be polite to existing postal code data (even if the existing data is incorrect). It also splits the postal code into postal_code and postal_code_suffix:

Code: [Select]

133            $values['geo_code_1'] = $ret['Latitude' ];
134            $values['geo_code_2'] = $ret['Longitude'];
               $current_pc = CRM_Utils_Array::value( 'postal_code', $values );
               // if a postal code was already entered, don't change it, except to make it more precise
               if (!$current_pc or (strpos($ret['Zip'], $current_pc) === 0)) {
                 $values['postal_code'] = $ret['Zip'];
                 /* the following logic to split the string was borrowed from CRM/Core/BAO/Address.php
                     -- CRM_Core_BAO_Address::fixAddress. This is actually the function that calls the geocoding script
                     to begin with, but the zipcode business takes place before geocoding gets called.
                 */
                 if ( preg_match('/^(\d{4,5})[+-](\d{4})$/',
                           $ret['Zip'],
                           $match) ) {
                   $values['postal_code']        = $match[1];
                   $values['postal_code_suffix'] = $match[2];
                 }
               }
136            return true;

you could put a similar check in civicrm/bin/UpdateAddress.php, modifying Blake's code --

Code: [Select]

               if (!$dao->postal_code or (strpos($params['postal_code'], $dao->postal_code) === 0))and split the postal code similarly.

[noah]

Here's a version for Google (CRM/Utils/Geocode/Google.php):

Code: [Select]

143             if ( $val[0] && $val[1] ) {
144                 $values['geo_code_1'] = $val[1];
145                 $values['geo_code_2'] = $val[0];
                    //ZIP CODE CODE BELOW ADDED BY NOAH MILLER
                    $zipobj = $xml->Response->Placemark->AddressDetails->Country->AdministrativeArea->SubAdministrativeArea->Locality->PostalCode->PostalCodeNumber;
                    $zip = (string)$zipobj;
                    $current_pc = CRM_Utils_Array::value( 'postal_code', $values );
                    // if a postal code was already entered, don't change it, except to make it more precise
                    if (!$current_pc or (strpos($zip, $current_pc) === 0)) {
                      $values['postal_code'] = $zip;
                      /* the following logic to split the string was borrowed from CRM/Core/BAO/Address.php
                          -- CRM_Core_BAO_Address::fixAddress. This is actually the function that calls the geocoding script
                          to begin with, but the zipcode business takes place before geocoding gets called.
                      */
                      if ( preg_match('/^(\d{4,5})[+-](\d{4})$/',
                                $zip['Zip'],
                                $match) ) {
                        $values['postal_code']        = $match[1];
                        $values['postal_code_suffix'] = $match[2];
                      }
                    }
146                 return true;
165             }