# This patch file was generated by NetBeans IDE # This patch can be applied using context Tools: Apply Diff Patch action on respective folder. # It uses platform neutral UTF-8 encoding. # Above lines and this line are ignored by the patching process. Index: millennium.module --- millennium.module Base (1.13.2.33.2.2.2.109) +++ millennium.module Locally Modified (Based On 1.13.2.33.2.2.2.109) @@ -2548,7 +2548,7 @@ } // Check Drupal's cache $cached_data = cache_get($cid); - if ($cached_data) { + if ($cached_data->data && $cached_data->expire > time()) { $cache[$cid] = $cached_data->data; return $cache[$cid]; } @@ -2564,7 +2564,20 @@ $fetched_html = $result->data; } - static $code_2_human = array( + $items = millennium_parse_holdings_from_html($fetched_html); + // Set static cache + $cache[$cid] = $items; + // Cache for one day + cache_set($cid, $items, 'cache', time()+(24*3600)); + return $items; +} + +/** + * Recieves html that contains a holdings table, and returns an array describing + * its items. + */ +function millennium_parse_holdings_from_html($html) { + static $code_2_human = array( '1' => 'location', '$' => 'location', '!' => 'message', @@ -2574,14 +2587,16 @@ '#' => 'copy', 'y' => '????', // TODO: find out what field y is! ); + $items = array(); + /* -- EXAMPLE 1 -- -   CEM GENERAL -      -  QA76.76.P37 M49 2007       + +   DISPONIBLE -- EXAMPLE 2 -- @@ -2613,38 +2628,38 @@  BP365 .E8  1970 Show Nearby Items On Shelf  CHECK SHELVES + -- EXAMPLE 5 -- + http://catalog.waukeganpl.org/record=b1292762&searchscope=0 + + +  Children's Magazine + +  MAGAZINE  no.26 Nov 2008  Nearby on shelf +  DUE 08-31-09BILLED + */ - $ok = preg_match_all('/(.*?<\/tr>(?:.*?<\/tr>|))/si', $fetched_html, $holding_matches, PREG_SET_ORDER); - if ($ok) { + if (preg_match_all('/(.*?<\/tr>(?:.*?<\/tr>|))/si', $html, $holding_matches, PREG_SET_ORDER)) { foreach ($holding_matches as $holding) { - $ok = preg_match_all('/(?: *  *|)(?:|)(.*?)(?:<\/|(?: *  *|)(?:|)(.*?)(?=<\/td|... matches. + foreach ($matches as $match) { + $name = $code_2_human[$match[1]]; + $value = trim(decode_entities(str_replace(" ", " ", strip_tags($match[2])))); + $item[$name] = $value; + } + // There can also be a URL assigned to the item: look for it. + $ok = preg_match('/colspan=".">.*?([^<]+).*?([^<]+) item parsing + $holdings_tests = array( + array( + ' +   CEM GENERAL + +  QA76.76.P37 M49 2007       + +   DISPONIBLE + ', + array ( + 'location' => 'CEM GENERAL', + 'classnumber' => 'QA76.76.P37 M49 2007', + 'classvolume' => '', + 'copy' => '', + 'message' => '', + 'status' => 'DISPONIBLE', + ) + ), + array( + ' + -- EXAMPLE 2 -- + + +   MTY Biblioteca Digital + +  RC346 .A3 2005b       + +   DISPONIBLE + + Consultar el servicio a través de la Biblioteca Digital + + ', + array ( + 'location' => 'MTY Biblioteca Digital', + 'classnumber' => 'RC346 .A3 2005b', + 'classvolume' => '', + 'copy' => '', + 'message' => '', + 'status' => 'DISPONIBLE', + 'link' => + array ( + 'url' => 'http://xxxxxxxxxxxxxxxxxxx/resourceTOC.aspx?resourceID=54', + 'title' => 'Consultar el servicio a través de la Biblioteca Digital', + ), + ) + ), + array(' + -- EXAMPLE 3 -- + + +   CCM HEMEROTECA + + +  F1201 N49  15: 1992 Ene.-Mar.    +   DISPONIBLE + ', + array ( + 'location' => 'CCM HEMEROTECA', + 'classnumber' => 'F1201 N49', + 'classvolume' => '15: 1992 Ene.-Mar.', + 'copy' => '', + 'status' => 'DISPONIBLE', + ) + ), + array( + ' + -- EXAMPLE 4 -- + + +  DEN Main + + +  BP365 .E8  1970 Show Nearby Items On Shelf +  CHECK SHELVES + ', + array ( + 'location' => 'DEN Main', + 'classnumber' => 'BP365 .E8', + 'copy' => '', + 'classvolume' => '1970', + 'message' => '', + '????' => 'Show Nearby Items On Shelf', + 'status' => 'CHECK SHELVES', + ) + ), + array( + ' + -- EXAMPLE 5 -- + http://catalog.waukeganpl.org/record=b1292762&searchscope=0 + + +  Childrens Magazine + +  MAGAZINE  no.26 Nov 2008  Nearby on shelf +  DUE 08-31-09BILLED + ', + array ( + 'location' => 'Childrens Magazine', + 'classnumber' => 'MAGAZINE', + 'classvolume' => 'no.26 Nov 2008', + 'copy' => '', + 'status' => 'DUE 08-31-09BILLED', + ) + ) + ); + foreach ($holdings_tests as $testnum => $test) { + $items = millennium_parse_holdings_from_html($test[0]); + $parsed = $items[0]; + $should_be = $test[1]; + $this->assertEqual($parsed, $should_be, t("Holdings parser test %testnum")); + } + // Test the get_real_baseurl parser // Bad URLs $bad_urls = explode('|', 'http:/example.com|example.com|ftp://example.com');