? .svn ? Rich_Document.php ? SolrPhpClient.2008-11-25.zip ? apachesolr_338534_highlighting.diff ? apachesolr_drupal_http_request.diff ? apachesolr_drupal_http_request_2.diff ? highlighting.diff ? highlighting_body.diff ? minimal-spelling-303937-27.patch ? new_php_client_and_solr_trunk.diff ? spatch.diff ? splunk.acquiapipe.net ? SolrPhpClient/.svn ? SolrPhpClient/LICENSE ? SolrPhpClient/phpdocs ? SolrPhpClient/Apache/.svn ? SolrPhpClient/Apache/Solr/.svn ? SolrPhpClient/Apache/Solr/RichDocument.php ? SolrPhpClient/Apache/Solr/do.php ? SolrPhpClient/Apache/Solr/Service/.svn ? contrib/.svn ? contrib/apachesolr_attachments/.svn ? contrib/apachesolr_image/.svn ? contrib/apachesolr_lang/.svn ? contrib/apachesolr_mlt/.svn ? contrib/apachesolr_multisitesearch/.svn ? contrib/apachesolr_nodeaccess/.svn ? contrib/apachesolr_nodeaccess/tests/.svn ? tests/.svn Index: apachesolr_search.module =================================================================== RCS file: /cvs/drupal-contrib/contributions/modules/apachesolr/apachesolr_search.module,v retrieving revision 1.1.2.6.2.45 diff -u -r1.1.2.6.2.45 apachesolr_search.module --- apachesolr_search.module 11 Dec 2008 20:57:47 -0000 1.1.2.6.2.45 +++ apachesolr_search.module 14 Dec 2008 11:16:07 -0000 @@ -106,6 +106,11 @@ // We default to getting snippets from the body. $hl_fl = is_null($params['hl.fl']) ? 'body' : $params['hl.fl']; + if (variable_get('apachesolr_search_spellcheck', FALSE)) { + //Add new parameter to the search request + $params['spellcheck.q'] = $query->get_query_basic(); + $params['spellcheck'] = 'true'; + } // TODO: This adds all of the possible facets to the query. Not all // of these facets have their blocks enabled, so the list should be @@ -211,9 +216,9 @@ $response = $solr->search($query->get_query_basic(), $params['start'], $params['rows'], $params); // The response is cached so that it is accessible to the blocks and anything // else that needs it beyond the initial search. + $total = $response->response->numFound; apachesolr_static_response_cache($response); apachesolr_has_searched(TRUE); - $total = $response->response->numFound; pager_query("SELECT %d", $params['rows'], 0, NULL, $total); if ($total > 0) { $extra = array(); @@ -441,6 +446,70 @@ } } +/** + * Implementation of hook_form_[form_id]_alter(). + * + * This adds spelling suggestions to the search form. + */ +function apachesolr_search_form_search_form_alter(&$form, $form_state) { + + if (($form['module']['#value'] == 'apachesolr_search') && variable_get('apachesolr_search_spellcheck', FALSE) && apachesolr_has_searched() && ($response = apachesolr_static_response_cache())) { + //Get spellchecker suggestions into an array. + $suggestions = get_object_vars($response->spellcheck->suggestions); + + if ($suggestions) { + //Prepage original query and replace wrong words. + $query = apachesolr_drupal_query(); + $new_keywords = $query->get_query_basic(); + + foreach($suggestions as $word => $value) { + $new_keywords = substr_replace($new_keywords, $value->suggestion[0], $value->startOffset, $value->endOffset - $value->startOffset); + } + + $form['basic']['suggestion'] = array( + '#prefix' => '
', + '#suffix' => '
', + '#type' => 'item', + '#title' => t('Did you mean'), + '#value' => l($new_keywords, 'search/'. arg(1) .'/'. $new_keywords), + ); + } + } +} + +/** + * Implementation of hook_form_[form_id]_alter(). + * + * This adds spelling suggestions to the search form. + */ +function apachesolr_search_form_apachesolr_settings_alter(&$form, $form_state) { + $form['apachesolr_search_spellcheck'] = array( + '#type' => 'checkbox', + '#title' => t('Enable Spellchecker and suggestions'), + '#default_value' => variable_get('apachesolr_search_spellcheck', FALSE), + '#description' => t('Enable spellchecker and get word suggestions. Also known as the "Did you mean" feature'), + ); + + $form['#submit'][] = 'apachesolr_search_build_specheck'; + // Move buttons to the bottom. + $buttons = $form['buttons']; + unset($form['buttons']); + $form['buttons'] = $buttons; +} + +function apachesolr_search_build_specheck() { + try { + $solr = apachesolr_get_solr(); + $params['spellcheck'] = 'true'; + $params['spellcheck.build'] = 'true'; + $params['echoParams'] = 'all'; + $response = $solr->search('solr', 0, 0, $params); + } + catch (Exception $e) { + watchdog('Apache Solr', $e->getMessage(), NULL, WATCHDOG_ERROR); + } +} + function apachesolr_search_get_username($facet) { if ($facet == 0) { return variable_get('anonymous', t('Anonymous')); Index: SolrPhpClient/ChangeLog =================================================================== RCS file: /cvs/drupal-contrib/contributions/modules/apachesolr/SolrPhpClient/Attic/ChangeLog,v retrieving revision 1.1.4.2 diff -u -r1.1.4.2 ChangeLog --- SolrPhpClient/ChangeLog 25 Oct 2008 14:47:29 -0000 1.1.4.2 +++ SolrPhpClient/ChangeLog 14 Dec 2008 11:16:08 -0000 @@ -1,3 +1,33 @@ +2008-11-24 18:14 djimenez + + * Pieter Berkel requested the ability to set field and document + boost values when adding documents to Solr: + + https://issues.apache.org/jira/browse/SOLR-341?focusedCommentId=12650095#action_12650095 + + Added the ability to do so in the Apache_Solr_Document class and + updated the XML fragment generation code in Apache_Solr_Service to + include the new boost values when available. + +2008-11-14 22:08 djimenez + + * Apache/Solr/Document.php: Rich Robinson raised an issue on the + Solr issue tracker where he found that documents with values = + false were creating prematurely exiting foreach loops. I tracked + it down to the valid() implementation of the Iterator iterface, + and after fussing with it decided to just implement + IteratorAggregator and use the pre-existing SPL ArrayIterator + class. Simplifies the code and now all document values are looped + even when some are false. + +2008-09-02 14:33 djimenez + + * Apache/Solr/Document.php, Apache/Solr/Response.php, + Apache/Solr/Service.php, Apache/Solr/Service/Balancer.php: At the + request of several people using these classes I've changed the + members and functions that were private to protected so that they + can be overridden / used in subclasses + 2008-08-26 14:50 dwolfe * Apache/Solr/Service/Balancer.php: Changed default value for @@ -19,11 +49,6 @@ to decide if we should retry the call. If the ping timesout then we select another read service (if one is available) and continue -2008-04-11 20:46 djimenez - - * Apache/Solr/Service/Balancer.php: doubled timeouts on balancer so - I'll see less exceptions during high load - 2008-01-26 01:03 djimenez * Apache/Solr/Response.php: HTTP 100 Continue response can also be @@ -50,6 +75,12 @@ build is not compiled with this options and we never would have found it ourselves. +2008-01-23 02:53 djimenez + + * Apache/Solr/Service.php, ChangeLog: 5 seconds seemed like too + long a default timeout, set to 2. Added a Changelog that I'm + packaging with the distribution. + 2008-01-14 19:12 djimenez * Apache/Solr/Response.php, Apache/Solr/Service.php, Solr.clss: @@ -76,11 +107,6 @@ balancer so that it will try other services during a communication error on a call. -2007-10-02 16:54 djimenez - - * Apache/Solr/Response.php: Missed changing the package / - subpackage pair. refs #29 - 2007-10-02 02:52 djimenez * Apache/Solr/Service.php: Got rid of the const defined defaults Index: SolrPhpClient/Apache/Solr/Document.php =================================================================== RCS file: /cvs/drupal-contrib/contributions/modules/apachesolr/SolrPhpClient/Apache/Solr/Document.php,v retrieving revision 1.1.4.4 diff -u -r1.1.4.4 Document.php --- SolrPhpClient/Apache/Solr/Document.php 25 Oct 2008 16:49:25 -0000 1.1.4.4 +++ SolrPhpClient/Apache/Solr/Document.php 14 Dec 2008 11:16:08 -0000 @@ -1,5 +1,4 @@ * ... * $document->title = 'Something'; @@ -40,11 +39,225 @@ * } * */ -class Apache_Solr_Document implements Iterator +class Apache_Solr_Document implements IteratorAggregate { + /** + * Document boost value + * + * @var float + */ + protected $_documentBoost = false; + + /** + * Document field values, indexed by name + * + * @var array + */ protected $_fields = array(); /** + * Document field boost values, indexed by name + * + * @var array array of floats + */ + protected $_fieldBoosts = array(); + + /** + * Clear all boosts and fields from this document + */ + public function clear() + { + $this->_documentBoost = false; + + $this->_fields = array(); + $this->_fieldBoosts = array(); + } + + /** + * Get current document boost + * + * @return mixed will be false for default, or else a float + */ + public function getBoost() + { + return $this->_documentBoost; + } + + /** + * Set document boost factor + * + * @param mixed $boost Use false for default boost, else cast to float + */ + public function setBoost($boost) + { + if ($boost !== false) + { + $this->_documentBoost = (float) $boost; + } + else + { + $this->_documentBoost = false; + } + } + + /** + * Add a value to a multi-valued field + * + * NOTE: the solr XML format allows you to specify boosts + * PER value even though the underlying Lucene implementation + * only allows a boost per field. To remedy this, the final + * field boost value will be the product of all specified boosts + * on field values - this is similar to SolrJ's functionality. + * + * + * $doc = new Apache_Solr_Document(); + * + * $doc->addField('foo', 'bar', 2.0); + * $doc->addField('foo', 'baz', 3.0); + * + * // resultant field boost will be 6! + * echo $doc->getFieldBoost('foo'); + * + * + * @param string $key + * @param mixed $value + * @param float $boost + */ + public function addField($key, $value, $boost = false) + { + if (!isset($this->_fields[$key])) + { + $this->_fields[$key] = array(); + } + + if (!isset($this->_fieldBoosts[$key])) + { + $this->setFieldBoost($key, $boost); + } + else if ($boost !== false) + { + if ($this->_fieldBoosts[$key] !== false) + { + $this->_fieldBoosts[$key] *= (float) $boost; + } + else + { + $this->_fieldBoosts[$key] = (float) $boost; + } + } + + if (!is_array($this->_fields[$key])) + { + $this->_fields[$key] = array($this->_fields[$key]); + } + + $this->_fields[$key][] = $value; + } + + /** + * Handle the array manipulation for a multi-valued field + * + * @param string $key + * @param string $value + * + * @deprecated Use addField(...) instead + */ + public function setMultiValue($key, $value, $boost = false) + { + $this->addField($key, $value, $boost); + } + + /** + * Get field information + * + * @param string $key + * @return mixed associative array of info if field exists, false otherwise + */ + public function getField($key) + { + if (isset($this->_fields[$key])) + { + return array( + 'name' => $key, + 'value' => $this->_fields[$key], + 'boost' => $this->_fieldBoosts[$key] + ); + } + + return false; + } + + /** + * Set a field value. Multi-valued fields should be set as arrays + * or instead use the addField(...) function which will automatically + * make sure the field is an array. + * + * @param string $key + * @param mixed $value + * @param float $boost + */ + public function setField($key, $value, $boost = false) + { + $this->_fields[$key] = $value; + $this->setFieldBoost($key, $boost); + } + + public function getFieldBoost($key) + { + return $this->_fieldBoosts[$key]; + } + + public function setFieldBoost($key, $boost) + { + //@note:JacobSingh changed this because of problem w/ multivalued fields + if ($boost !== false && $boost !== null) + { + $this->_fieldBoosts[$key] = (float) $boost; + } + else + { + $this->_fieldBoosts = false; + } + } + + /** + * Get the names of all fields in this document + * + * @return array + */ + public function getFieldNames() + { + return array_keys($this->_fields); + } + + /** + * Get the values of all fields in this document + * + * @return array + */ + public function getFieldValues() + { + return array_values($this->_fields); + } + + /** + * IteratorAggregate implementation function. Allows usage: + * + * + * foreach ($document as $key => $value) + * { + * ... + * } + * + */ + public function getIterator() + { + $arrayObject = new ArrayObject($this->_fields); + + return $arrayObject->getIterator(); + } + + /** * Magic get for field values * * @param string $key @@ -52,149 +265,54 @@ */ public function __get($key) { - return $this->_fields[$key]; + return $this->_fields[$key]; + } + + /** + * Magic set for field values. Multi-valued fields should be set as arrays + * or instead use the setMultiValue(...) function which will automatically + * make sure the field is an array. + * + * @param string $key + * @param mixed $value + */ + public function __set($key, $value) + { + $this->_fields[$key] = $value; + + if (!isset($this->_fieldBoosts[$key])) + { + $this->_fieldBoosts[$key] = false; + } + } + + /** + * Magic isset for fields values. Do not call directly. Allows usage: + * + * + * isset($document->some_field); + * + * + * @param string $key + * @return boolean + */ + public function __isset($key) + { + return isset($this->_fields[$key]); } - /** - * Magic set for field values. Multi-valued fields should be set as arrays - * or instead use the setMultiValue(...) function which will automatically - * make sure the field is an array. - * - * @param string $key - * @param mixed $value - */ - public function __set($key, $value) - { - $this->_fields[$key] = $value; - } - - /** - * Magic isset for fields values. Do no call directly. Allows usage: - * - * - * isset($document->some_field); - * - * - * @param string $key - * @return boolean - */ - public function __isset($key) - { - return isset($this->_fields[$key]); - } - - /** - * Magic unset for field values. Do no call directly. Allows usage: - * - * - * unset($document->some_field); - * - * - * @param string $key - */ - public function __unset($key) - { - unset($this->_fields[$key]); - } - - /** - * Handle the array manipulation for a multi-valued field - * - * @param string $key - * @param string $value - */ - public function setMultiValue($key, $value) - { - if (!isset($this->_fields[$key])) - { - $this->_fields[$key] = array(); - } - - if (!is_array($this->_fields[$key])) - { - $this->_fields[$key] = array($this->_fields[$key]); - } - - $this->_fields[$key][] = $value; - } - - /** - * Get the names of all fields in this document - * - * @return array - */ - public function getFieldNames() - { - return array_keys($this->_fields); - } - - /** - * Iterator implementation function, proxies to _fields. Allows usage: - * - * - * foreach ($document as $key => $value) - * { - * ... - * } - * - */ - public function rewind() { - reset($this->_fields); - } - - /** - * Iterator implementation function, proxies to _fields. Allows usage: - * - * - * foreach ($document as $key => $value) - * { - * ... - * } - * - */ - public function current() { - return current($this->_fields); - } - - /** - * Iterator implementation function, proxies to _fields. Allows usage: - * - * - * foreach ($document as $key => $value) - * { - * ... - * } - * - */ - public function key() { - return key($this->_fields); - } - - /** - * Iterator implementation function, proxies to _fields. Allows usage: - * - * - * foreach ($document as $key => $value) - * { - * ... - * } - * - */ - public function next() { - return next($this->_fields); - } - - /** - * Iterator implementation function, proxies to _fields. Allows usage: - * - * - * foreach ($document as $key => $value) - * { - * ... - * } - * - */ - public function valid() { - return current($this->_fields) !== false; - } + /** + * Magic unset for field values. Do not call directly. Allows usage: + * + * + * unset($document->some_field); + * + * + * @param string $key + */ + public function __unset($key) + { + unset($this->_fields[$key]); + unset($this->_fieldBoosts[$key]); + } } \ No newline at end of file Index: SolrPhpClient/Apache/Solr/Response.php =================================================================== RCS file: /cvs/drupal-contrib/contributions/modules/apachesolr/SolrPhpClient/Apache/Solr/Response.php,v retrieving revision 1.1.2.1.2.6 diff -u -r1.1.2.1.2.6 Response.php --- SolrPhpClient/Apache/Solr/Response.php 25 Oct 2008 16:49:25 -0000 1.1.2.1.2.6 +++ SolrPhpClient/Apache/Solr/Response.php 14 Dec 2008 11:16:09 -0000 @@ -1,5 +1,4 @@ 0) - { - //look at the first headers for the HTTP status code - //and message (errors are usually returned this way) - // - //HTTP 100 Continue response can also be returned before - //the REAL status header, so we need look until we find - //the last header starting with HTTP - // - //the spec: http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html#sec10.1 - // - //Thanks to Daniel Andersson for pointing out this oversight - while (isset($httpHeaders[0]) && substr($httpHeaders[0], 0, 4) == 'HTTP') - { - $parts = split(' ', substr($httpHeaders[0], 9), 2); - - $status = $parts[0]; - $statusMessage = trim($parts[1]); - - array_shift($httpHeaders); - } - - //Look for the Content-Type response header and determine type - //and encoding from it (if possible - such as 'Content-Type: text/plain; charset=UTF-8') - foreach ($httpHeaders as $header) - { - if (strncasecmp($header, 'Content-Type:', 13) == 0) - { - //split content type value into two parts if possible - $parts = split(';', substr($header, 13), 2); - - $type = trim($parts[0]); - - if ($parts[1]) - { - //split the encoding section again to get the value - $parts = split('=', $parts[1], 2); - - if ($parts[1]) - { - $encoding = trim($parts[1]); - } - } - - break; - } - } - } - - $this->_rawResponse = $rawResponse; - $this->_type = $type; - $this->_encoding = $encoding; - $this->_httpStatus = $status; - $this->_httpStatusMessage = $statusMessage; - $this->_createDocuments = (bool) $createDocuments; - $this->_collapseSingleValueArrays = (bool) $collapseSingleValueArrays; - } - - /** - * Get the HTTP status code - * - * @return integer - */ - public function getHttpStatus() - { - return $this->_httpStatus; - } - - /** - * Get the HTTP status message of the response - * - * @return string - */ - public function getHttpStatusMessage() - { - return $this->_httpStatusMessage; - } - - /** - * Get content type of this Solr response - * - * @return string - */ - public function getType() - { - return $this->_type; - } - - /** - * Get character encoding of this response. Should usually be utf-8, but just in case - * - * @return string - */ - public function getEncoding() - { - return $this->_encoding; - } - - /** - * Get the raw response as it was given to this object - * - * @return string - */ - public function getRawResponse() - { - return $this->_rawResponse; - } - - /** - * Magic get to expose the parsed data and to lazily load it - * - * @param unknown_type $key - * @return unknown - */ - public function __get($key) - { - if (!$this->_isParsed) - { - $this->_parseData(); - $this->_isParsed = true; - } - - if (isset($this->_parsedData->$key)) - { - return $this->_parsedData->$key; - } - - return null; - } - - /** - * Parse the raw response into the parsed_data array for access - */ - protected function _parseData() - { - //An alternative would be to use Zend_Json::decode(...) - $data = json_decode($this->_rawResponse); - - //if we're configured to collapse single valued arrays or to convert them to Apache_Solr_Document objects - //and we have response documents, then try to collapse the values and / or convert them now - if (($this->_createDocuments || $this->_collapseSingleValueArrays) && isset($data->response) && is_array($data->response->docs)) - { - $documents = array(); - - foreach ($data->response->docs as $originalDocument) - { - if ($this->_createDocuments) - { - $document = new Apache_Solr_Document(); - } - else - { - $document = $originalDocument; - } - - foreach ($originalDocument as $key => $value) - { - //If a result is an array with only a single - //value then its nice to be able to access - //it as if it were always a single value - if ($this->_collapseSingleValueArrays && is_array($value) && count($value) <= 1) - { - $value = array_shift($value); - } - - $document->$key = $value; - } + /** + * Holds the raw response used in construction + * + * @var string + */ + protected $_rawResponse; + + /** + * Parsed values from the passed in http headers + * + * @var string + */ + protected $_httpStatus, $_httpStatusMessage, $_type, $_encoding; + + /** + * Whether the raw response has been parsed + * + * @var boolean + */ + protected $_isParsed = false; + + /** + * Parsed representation of the data + * + * @var mixed + */ + protected $_parsedData; + + /** + * Data parsing flags. Determines what extra processing should be done + * after the data is initially converted to a data structure. + * + * @var boolean + */ + protected $_createDocuments = true, + $_collapseSingleValueArrays = true; + + /** + * Constructor. Takes the raw HTTP response body and the exploded HTTP headers + * + * @param string $rawResponse + * @param array $httpHeaders + * @param boolean $createDocuments Whether to convert the documents json_decoded as stdClass instances to Apache_Solr_Document instances + * @param boolean $collapseSingleValueArrays Whether to make multivalued fields appear as single values + */ + public function __construct($rawResponse, $httpHeaders = array(), $createDocuments = true, $collapseSingleValueArrays = true) + { + //Assume 0, 'Communication Error', utf-8, and text/plain + $status = 0; + $statusMessage = 'Communication Error'; + $type = 'text/plain'; + $encoding = 'UTF-8'; + + //iterate through headers for real status, type, and encoding + if (is_array($httpHeaders) && count($httpHeaders) > 0) + { + //look at the first headers for the HTTP status code + //and message (errors are usually returned this way) + // + //HTTP 100 Continue response can also be returned before + //the REAL status header, so we need look until we find + //the last header starting with HTTP + // + //the spec: http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html#sec10.1 + // + //Thanks to Daniel Andersson for pointing out this oversight + while (isset($httpHeaders[0]) && substr($httpHeaders[0], 0, 4) == 'HTTP') + { + $parts = split(' ', substr($httpHeaders[0], 9), 2); + + $status = $parts[0]; + $statusMessage = trim($parts[1]); + + array_shift($httpHeaders); + } + + //Look for the Content-Type response header and determine type + //and encoding from it (if possible - such as 'Content-Type: text/plain; charset=UTF-8') + foreach ($httpHeaders as $header) + { + if (strncasecmp($header, 'Content-Type:', 13) == 0) + { + //split content type value into two parts if possible + $parts = split(';', substr($header, 13), 2); + + $type = trim($parts[0]); + + if ($parts[1]) + { + //split the encoding section again to get the value + $parts = split('=', $parts[1], 2); + + if ($parts[1]) + { + $encoding = trim($parts[1]); + } + } + + break; + } + } + } + + $this->_rawResponse = $rawResponse; + $this->_type = $type; + $this->_encoding = $encoding; + $this->_httpStatus = $status; + $this->_httpStatusMessage = $statusMessage; + $this->_createDocuments = (bool) $createDocuments; + $this->_collapseSingleValueArrays = (bool) $collapseSingleValueArrays; + } + + /** + * Get the HTTP status code + * + * @return integer + */ + public function getHttpStatus() + { + return $this->_httpStatus; + } + + /** + * Get the HTTP status message of the response + * + * @return string + */ + public function getHttpStatusMessage() + { + return $this->_httpStatusMessage; + } + + /** + * Get content type of this Solr response + * + * @return string + */ + public function getType() + { + return $this->_type; + } + + /** + * Get character encoding of this response. Should usually be utf-8, but just in case + * + * @return string + */ + public function getEncoding() + { + return $this->_encoding; + } + + /** + * Get the raw response as it was given to this object + * + * @return string + */ + public function getRawResponse() + { + return $this->_rawResponse; + } + + /** + * Magic get to expose the parsed data and to lazily load it + * + * @param unknown_type $key + * @return unknown + */ + public function __get($key) + { + if (!$this->_isParsed) + { + $this->_parseData(); + $this->_isParsed = true; + } + + if (isset($this->_parsedData->$key)) + { + return $this->_parsedData->$key; + } + + return null; + } + + /** + * Parse the raw response into the parsed_data array for access + */ + protected function _parseData() + { + //An alternative would be to use Zend_Json::decode(...) + $data = json_decode($this->_rawResponse); + + //if we're configured to collapse single valued arrays or to convert them to Apache_Solr_Document objects + //and we have response documents, then try to collapse the values and / or convert them now + if (($this->_createDocuments || $this->_collapseSingleValueArrays) && isset($data->response) && is_array($data->response->docs)) + { + $documents = array(); + + foreach ($data->response->docs as $originalDocument) + { + if ($this->_createDocuments) + { + $document = new Apache_Solr_Document(); + } + else + { + $document = $originalDocument; + } + + foreach ($originalDocument as $key => $value) + { + //If a result is an array with only a single + //value then its nice to be able to access + //it as if it were always a single value + if ($this->_collapseSingleValueArrays && is_array($value) && count($value) <= 1) + { + $value = array_shift($value); + } + + $document->$key = $value; + } - $documents[] = $document; - } + $documents[] = $document; + } - $data->response->docs = $documents; - } + $data->response->docs = $documents; + } - $this->_parsedData = $data; - } + $this->_parsedData = $data; + } } \ No newline at end of file Index: SolrPhpClient/Apache/Solr/Service.php =================================================================== RCS file: /cvs/drupal-contrib/contributions/modules/apachesolr/SolrPhpClient/Apache/Solr/Service.php,v retrieving revision 1.1.2.1.2.10 diff -u -r1.1.2.1.2.10 Service.php --- SolrPhpClient/Apache/Solr/Service.php 18 Nov 2008 16:40:12 -0000 1.1.2.1.2.10 +++ SolrPhpClient/Apache/Solr/Service.php 14 Dec 2008 11:16:09 -0000 @@ -1,5 +1,4 @@ */ +//@note:JacobSingh changed this so class location was not hardcoded, but relative require_once(dirname(__FILE__) . '/Document.php'); require_once(dirname(__FILE__) . '/Response.php'); @@ -35,24 +35,24 @@ * * if ($solr->ping()) * { - * $solr->deleteByQuery('*:*'); //deletes ALL documents - be careful :) + * $solr->deleteByQuery('*:*'); //deletes ALL documents - be careful :) * - * $document = new Apache_Solr_Document(); - * $document->id = uniqid(); //or something else suitably unique + * $document = new Apache_Solr_Document(); + * $document->id = uniqid(); //or something else suitably unique * - * $document->title = 'Some Title'; - * $document->content = 'Some content for this wonderful document. Blah blah blah.'; + * $document->title = 'Some Title'; + * $document->content = 'Some content for this wonderful document. Blah blah blah.'; * - * $solr->addDocument($document); //if you're going to be adding documents in bulk using addDocuments - * //with an array of documents is faster + * $solr->addDocument($document); //if you're going to be adding documents in bulk using addDocuments + * //with an array of documents is faster * - * $solr->commit(); //commit to see the deletes and the document - * $solr->optimize(); //merges multiple segments into one + * $solr->commit(); //commit to see the deletes and the document + * $solr->optimize(); //merges multiple segments into one * - * //and the one we all care about, search! - * //any other common or custom parameters to the request handler can go in the - * //optional 4th array argument. - * $solr->search('content:blah', 0, 10, array('sort' => 'timestamp desc')); + * //and the one we all care about, search! + * //any other common or custom parameters to the request handler can go in the + * //optional 4th array argument. + * $solr->search('content:blah', 0, 10, array('sort' => 'timestamp desc')); * } * ... * @@ -62,825 +62,848 @@ */ class Apache_Solr_Service { - /** - * Response version we support - */ - const SOLR_VERSION = '1.2'; - - /** - * Response writer we support - * - * @todo Solr 1.3 release may change this to SerializedPHP or PHP implementation - */ - const SOLR_WRITER = 'json'; - - /** - * NamedList Treatment constants - */ - const NAMED_LIST_FLAT = 'flat'; - const NAMED_LIST_MAP = 'map'; - - /** - * Servlet mappings - */ - const PING_SERVLET = 'admin/ping'; - const UPDATE_SERVLET = 'update'; - const SEARCH_SERVLET = 'select'; - const THREADS_SERVLET = 'admin/threads'; - - /** - * Server identification strings - * - * @var string - */ - protected $_host, $_port, $_path; - - /** - * Whether {@link Apache_Solr_Response} objects should create {@link Apache_Solr_Document}s in - * the returned parsed data - * - * @var boolean - */ - protected $_createDocuments = true; - - /** - * Whether {@link Apache_Solr_Response} objects should have multivalue fields with only a single value - * collapsed to appear as a single value would. - * - * @var boolean - */ - protected $_collapseSingleValueArrays = true; - - /** - * How NamedLists should be formatted in the output. This specifically effects facet counts. Valid values - * are {@link Apache_Solr_Service::NAMED_LIST_MAP} (default) or {@link Apache_Solr_Service::NAMED_LIST_FLAT}. - * - * @var string - */ - protected $_namedListTreatment = self::NAMED_LIST_MAP; - - /** - * Query delimiters. Someone might want to be able to change - * these (to use & instead of & for example), so I've provided them. - * - * @var string - */ - protected $_queryDelimiter = '?', $_queryStringDelimiter = '&'; - - /** - * Constructed servlet full path URLs - * - * @var string - */ - protected $_updateUrl, $_searchUrl, $_threadsUrl; - - /** - * Keep track of whether our URLs have been constructed - * - * @var boolean - */ - protected $_urlsInited = false; - - /** - * Stream context for posting - * - * @var resource - */ - protected $_postContext; - - /** - * Escape a value for special query characters such as ':', '(', ')', '*', '?', etc. - * - * NOTE: inside a phrase fewer characters need escaped, use {@link Apache_Solr_Service::escapePhrase()} instead - * - * @param string $value - * @return string - */ - static public function escape($value) - { - //list taken from http://lucene.apache.org/java/docs/queryparsersyntax.html#Escaping%20Special%20Characters - $pattern = '/(\+|-|&&|\|\||!|\(|\)|\{|}|\[|]|\^|"|~|\*|\?|:|\\\)/'; - $replace = '\\\$1'; - - return preg_replace($pattern, $replace, $value); - } - - /** - * Escape a value meant to be contained in a phrase for special query characters - * - * @param string $value - * @return string - */ - static public function escapePhrase($value) - { - $pattern = '/("|\\\)/'; - $replace = '\\\$1'; - - return preg_replace($pattern, $replace, $value); - } - - /** - * Convenience function for creating phrase syntax from a value - * - * @param string $value - * @return string - */ - static public function phrase($value) - { - return '"' . self::escapePhrase($value) . '"'; - } - - /** - * Constructor. All parameters are optional and will take on default values - * if not specified. - * - * @param string $host - * @param string $port - * @param string $path - */ - public function __construct($host = 'localhost', $port = 8180, $path = '/solr/') - { - $this->setHost($host); - $this->setPort($port); - $this->setPath($path); - - $this->_initUrls(); - - //set up the stream context for posting with file_get_contents - $contextOpts = array( - 'http' => array( - 'method' => 'POST', - 'header' => "Content-Type: text/xml; charset=UTF-8\r\n" //php.net example showed \r\n at the end - ) - ); - - $this->_postContext = stream_context_create($contextOpts); - } - - /** - * Return a valid http URL given this server's host, port and path and a provided servlet name - * - * @param string $servlet - * @return string - */ - protected function _constructUrl($servlet, $params = array()) - { - if (count($params)) - { - //escape all parameters appropriately for inclusion in the query string - $escapedParams = array(); - - foreach ($params as $key => $value) - { - $escapedParams[] = urlencode($key) . '=' . urlencode($value); - } - - $queryString = $this->_queryDelimiter . implode($this->_queryStringDelimiter, $escapedParams); - } - else - { - $queryString = ''; - } - - return 'http://' . $this->_host . ':' . $this->_port . $this->_path . $servlet . $queryString; - } - - /** - * Construct the Full URLs for the three servlets we reference - */ - protected function _initUrls() - { - //Initialize our full servlet URLs now that we have server information - $this->_updateUrl = $this->_constructUrl(self::UPDATE_SERVLET, array('wt' => self::SOLR_WRITER )); - $this->_searchUrl = $this->_constructUrl(self::SEARCH_SERVLET); - $this->_threadsUrl = $this->_constructUrl(self::THREADS_SERVLET, array('wt' => self::SOLR_WRITER )); - - $this->_urlsInited = true; - } - - /** - * Central method for making a get operation against this Solr Server - * - * @param string $url - * @param float $timeout Read timeout in seconds - * @return Apache_Solr_Response - * - * @todo implement timeout ability - * @throws Exception If a non 200 response status is returned - */ - protected function _sendRawGet($url, $timeout = FALSE) - { - //$http_response_header is set by file_get_contents - $http_response_header = NULL; - $response = new Apache_Solr_Response(@file_get_contents($url), $http_response_header, $this->_createDocuments, $this->_collapseSingleValueArrays); - - if ($response->getHttpStatus() != 200) - { - throw new Exception('"' . $response->getHttpStatus() . '" Status: ' . $response->getHttpStatusMessage(), $response->getHttpStatus()); - } - - return $response; - } - - /** - * Central method for making a post operation against this Solr Server - * - * @param string $url - * @param string $rawPost - * @param float $timeout Read timeout in seconds - * @param string $contentType - * @return Apache_Solr_Response - * - * @throws Exception If a non 200 response status is returned - */ - protected function _sendRawPost($url, $rawPost, $timeout = FALSE, $contentType = 'text/xml; charset=UTF-8') - { - //ensure content type is correct - stream_context_set_option($this->_postContext, 'http', 'header', 'Content-Type: ' . $contentType); - - //set the read timeout if specified - if ($timeout !== FALSE) - { - stream_context_set_option($this->_postContext, 'http', 'timeout', $timeout); - } - - //set the content - stream_context_set_option($this->_postContext, 'http', 'content', $rawPost); - - //$http_response_header is set by file_get_contents - $response = new Apache_Solr_Response(@file_get_contents($url, false, $this->_postContext), $http_response_header, $this->_createDocuments, $this->_collapseSingleValueArrays); - - if ($response->getHttpStatus() != 200) - { - throw new Exception('"' . $response->getHttpStatus() . '" Status: ' . $response->getHttpStatusMessage(), $response->getHttpStatus()); - } - - return $response; - } - - /** - * Returns the set host - * - * @return string - */ - public function getHost() - { - return $this->_host; - } - - /** - * Set the host used. If empty will fallback to constants - * - * @param string $host - */ - public function setHost($host) - { - //Use the provided host or use the default - if (empty($host)) - { - throw new Exception('Host parameter is empty'); - } - else - { - $this->_host = $host; - } - - if ($this->_urlsInited) - { - $this->_initUrls(); - } - } - - /** - * Get the set port - * - * @return integer - */ - public function getPort() - { - return $this->_port; - } - - /** - * Set the port used. If empty will fallback to constants - * - * @param integer $port - */ - public function setPort($port) - { - //Use the provided port or use the default - $port = (int) $port; - - if ($port <= 0) - { - throw new Exception('Port is not a valid port number'); - } - else - { - $this->_port = $port; - } - - if ($this->_urlsInited) - { - $this->_initUrls(); - } - } - - /** - * Get the set path. - * - * @return string - */ - public function getPath() - { - return $this->_path; - } - - /** - * Set the path used. If empty will fallback to constants - * - * @param string $path - */ - public function setPath($path) - { - $path = trim($path, '/'); - - $this->_path = '/' . $path . '/'; - - if ($this->_urlsInited) - { - $this->_initUrls(); - } - } - - /** - * Set the create documents flag. This determines whether {@link Apache_Solr_Response} objects will - * parse the response and create {@link Apache_Solr_Document} instances in place. - * - * @param unknown_type $createDocuments - */ - public function setCreateDocuments($createDocuments) - { - $this->_createDocuments = (bool) $createDocuments; - } - - /** - * Get the current state of teh create documents flag. - * - * @return boolean - */ - public function getCreateDocuments() - { - return $this->_createDocuments; - } - - /** - * Set the collapse single value arrays flag. - * - * @param boolean $collapseSingleValueArrays - */ - public function setCollapseSingleValueArrays($collapseSingleValueArrays) - { - $this->_collapseSingleValueArrays = (bool) $collapseSingleValueArrays; - } - - /** - * Get the current state of the collapse single value arrays flag. - * - * @return boolean - */ - public function getCollapseSingleValueArrays() - { - return $this->_collapseSingleValueArrays; - } - - /** - * Set how NamedLists should be formatted in the response data. This mainly effects - * the facet counts format. - * - * @param string $namedListTreatment - * @throws Exception If invalid option is set - */ - public function setNamedListTreatmet($namedListTreatment) - { - switch ((string) $namedListTreatment) - { - case Apache_Solr_Service::NAMED_LIST_FLAT: - $this->_namedListTreatment = Apache_Solr_Service::NAMED_LIST_FLAT; - break; - - case Apache_Solr_Service::NAMED_LIST_MAP: - $this->_namedListTreatment = Apache_Solr_Service::NAMED_LIST_MAP; - break; - - default: - throw new Exception('Not a valid named list treatement option'); - } - } - - /** - * Get the current setting for named list treatment. - * - * @return string - */ - public function getNamedListTreatment() - { - return $this->_namedListTreatment; - } - - - /** - * Set the string used to separate the path form the query string. - * Defaulted to '?' - * - * @param string $queryDelimiter - */ - public function setQueryDelimiter($queryDelimiter) - { - $this->_queryDelimiter = $queryDelimiter; - } - - /** - * Set the string used to separate the parameters in thequery string - * Defaulted to '&' - * - * @param string $queryStringDelimiter - */ - public function setQueryStringDelimiter($queryStringDelimiter) - { - $this->_queryStringDelimiter = $queryStringDelimiter; - } - - /** - * Call the /admin/ping servlet, can be used to quickly tell if a connection to the - * server is able to be made. - * - * @param float $timeout maximum time to wait for ping in seconds, -1 for unlimited (default is 2) - * @return float Actual time taken to ping the server, FALSE if timeout occurs - */ - public function ping($timeout = 2) - { - $timeout = (float) $timeout; - - if ($timeout <= 0) - { - $timeout = -1; - } - - $start = microtime(true); - - //to prevent strict errors - $errno = 0; - $errstr = ''; - - //try to connect to the host with timeout - $fp = fsockopen($this->_host, $this->_port, $errno, $errstr, $timeout); - - if ($fp) - { - //If we have a timeout set, then determine the amount of time we have left - //in the request and set the stream timeout for the write operation - if ($timeout > 0) - { - //do the calculation - $writeTimeout = $timeout - (microtime(true) - $start); - - //check if we're out of time - if ($writeTimeout <= 0) - { - fclose($fp); - return false; - } - - //convert to microseconds and set the stream timeout - $writeTimeoutInMicroseconds = (int) $writeTimeout * 1000000; - stream_set_timeout($fp, 0, $writeTimeoutInMicroseconds); - } - - $request = 'HEAD ' . $this->_path . self::PING_SERVLET . ' HTTP/1.1' . "\r\n" . - 'host: ' . $this->_host . "\r\n" . - 'Connection: close' . "\r\n" . - "\r\n"; - - fwrite($fp, $request); - - //check the stream meta data to see if we timed out during the operation - $metaData = stream_get_meta_data($fp); - - if (isset($metaData['timeout']) && $metaData['timeout']) - { - fclose($fp); - return false; - } - - - //if we have a timeout set and have made it this far, determine the amount of time - //still remaining and set the timeout appropriately before the read operation - if ($timeout > 0) - { - //do the calculation - $readTimeout = $timeout - (microtime(true) - $start); - - //check if we've run out of time - if ($readTimeout <= 0) - { - fclose($fp); - return false; - } - - //convert to microseconds and set the stream timeout - $readTimeoutInMicroseconds = $readTimeout * 1000000; - stream_set_timeout($fp, 0, $readTimeoutInMicroseconds); - } - - //at the very least we should get a response header line of - //HTTP/1.1 200 OK - $response = fread($fp, 15); - - //check the stream meta data to see if we timed out during the operation - $metaData = stream_get_meta_data($fp); - fclose($fp); //we're done with the connection - ignore the rest - - if (isset($metaData['timeout']) && $metaData['timeout']) - { - return false; - } - - //finally, check the response header line - if ($response != 'HTTP/1.1 200 OK') - { - return false; - } - - //we made it, return the approximate ping time - return microtime(true) - $start; - } - - //we weren't able to make a connection - return false; - } - - /** - * Call the /admin/threads servlet and retrieve information about all threads in the - * Solr servlet's thread group. Useful for diagnostics. - * - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function threads() - { - return $this->_sendRawGet($this->_threadsUrl); - } - - /** - * Raw Add Method. Takes a raw post body and sends it to the update service. Post body - * should be a complete and well formed "add" xml document. - * - * @param string $rawPost - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function add($rawPost) - { - return $this->_sendRawPost($this->_updateUrl, $rawPost); - } - - /** - * Add a Solr Document to the index - * - * @param Apache_Solr_Document $document - * @param boolean $allowDups - * @param boolean $overwritePending - * @param boolean $overwriteCommitted - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function addDocument(Apache_Solr_Document $document, $allowDups = false, $overwritePending = true, $overwriteCommitted = true) - { - $dupValue = $allowDups ? 'true' : 'false'; - $pendingValue = $overwritePending ? 'true' : 'false'; - $committedValue = $overwriteCommitted ? 'true' : 'false'; - - $rawPost = ''; - $rawPost .= $this->_documentToXmlFragment($document); - $rawPost .= ''; - - return $this->add($rawPost); - } - - /** - * Add an array of Solr Documents to the index all at once - * - * @param array $documents Should be an array of Apache_Solr_Document instances - * @param boolean $allowDups - * @param boolean $overwritePending - * @param boolean $overwriteCommitted - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function addDocuments($documents, $allowDups = false, $overwritePending = true, $overwriteCommitted = true) - { - $dupValue = $allowDups ? 'true' : 'false'; - $pendingValue = $overwritePending ? 'true' : 'false'; - $committedValue = $overwriteCommitted ? 'true' : 'false'; - - $rawPost = ''; - - foreach ($documents as $document) - { - if ($document instanceof Apache_Solr_Document) - { - $rawPost .= $this->_documentToXmlFragment($document); - } - } - - $rawPost .= ''; - - return $this->add($rawPost); - } - - /** - * Create an XML fragment from a {@link Apache_Solr_Document} instance appropriate for use inside a Solr add call - * - * @return string - */ - protected function _documentToXmlFragment(Apache_Solr_Document $document) - { - $xml = ''; - - foreach ($document as $key => $value) - { - $key = htmlspecialchars($key, ENT_QUOTES, 'UTF-8'); - - if (is_array($value)) - { - foreach ($value as $multivalue) - { - $multivalue = htmlspecialchars($multivalue, ENT_NOQUOTES, 'UTF-8'); - - $xml .= '' . $multivalue . ''; - } - } - else - { - $value = htmlspecialchars($value, ENT_NOQUOTES, 'UTF-8'); - - $xml .= '' . $value . ''; - } - } - - $xml .= ''; - - return $xml; - } - - /** - * Send a commit command. Will be synchronous unless both wait parameters are set to false. - * - * @param boolean $optimize Defaults to true - * @param boolean $waitFlush Defaults to true - * @param boolean $waitSearcher Defaults to true - * @param float $timeout Maximum expected duration (in seconds) of the commit operation on the server (otherwise, will throw a communication exception). Defaults to 1 hour - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function commit($optimize = true, $waitFlush = true, $waitSearcher = true, $timeout = 3600) - { - $optimizeValue = $optimize ? 'true' : 'false'; - $flushValue = $waitFlush ? 'true' : 'false'; - $searcherValue = $waitSearcher ? 'true' : 'false'; - - $rawPost = ''; - - return $this->_sendRawPost($this->_updateUrl, $rawPost, $timeout); - } - - /** - * Raw Delete Method. Takes a raw post body and sends it to the update service. Body should be - * a complete and well formed "delete" xml document - * - * @param string $rawPost Expected to be utf-8 encoded xml document - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function delete($rawPost) - { - return $this->_sendRawPost($this->_updateUrl, $rawPost); - } - - /** - * Create a delete document based on document ID - * - * @param string $id Expected to be utf-8 encoded - * @param boolean $fromPending - * @param boolean $fromCommitted - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function deleteById($id, $fromPending = true, $fromCommitted = true) - { - $pendingValue = $fromPending ? 'true' : 'false'; - $committedValue = $fromCommitted ? 'true' : 'false'; - - //escape special xml characters - $id = htmlspecialchars($id, ENT_NOQUOTES, 'UTF-8'); - - $rawPost = '' . $id . ''; - - return $this->delete($rawPost); - } - - /** - * Create a delete document based on a query and submit it - * - * @param string $rawQuery Expected to be utf-8 encoded - * @param boolean $fromPending - * @param boolean $fromCommitted - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function deleteByQuery($rawQuery, $fromPending = true, $fromCommitted = true) - { - $pendingValue = $fromPending ? 'true' : 'false'; - $committedValue = $fromCommitted ? 'true' : 'false'; - - // escape special xml characters - $rawQuery = htmlspecialchars($rawQuery, ENT_NOQUOTES, 'UTF-8'); - - $rawPost = '' . $rawQuery . ''; - - return $this->delete($rawPost); - } - - /** - * Send an optimize command. Will be synchronous unless both wait parameters are set - * to false. - * - * @param boolean $waitFlush - * @param boolean $waitSearcher - * @param float $timeout Maximum expected duration of the commit operation on the server (otherwise, will throw a communication exception) - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function optimize($waitFlush = true, $waitSearcher = true, $timeout = 3600) - { - $flushValue = $waitFlush ? 'true' : 'false'; - $searcherValue = $waitSearcher ? 'true' : 'false'; - - $rawPost = ''; - - return $this->_sendRawPost($this->_updateUrl, $rawPost, $timeout); - } - - /** - * Simple Search interface - * - * @param string $query The raw query string - * @param int $offset The starting offset for result documents - * @param int $limit The maximum number of result documents to return - * @param array $params key / value pairs for other query parameters (see Solr documentation), use arrays for parameter keys used more than once (e.g. facet.field) - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function search($query, $offset = 0, $limit = 10, $params = array()) - { - if (!is_array($params)) - { - $params = array(); - } - - // construct our full parameters - // sending the version is important in case the format changes - $params['version'] = self::SOLR_VERSION; - - // common parameters in this interface - $params['wt'] = self::SOLR_WRITER; - $params['json.nl'] = $this->_namedListTreatment; - - $params['q'] = $query; - $params['start'] = $offset; - $params['rows'] = $limit; - - // use http_build_query to encode our arguments because its faster - // than urlencoding all the parts ourselves in a loop - $queryString = http_build_query($params, null, $this->_queryStringDelimiter); - - // because http_build_query treats arrays differently than we want to, correct the query - // string by changing foo[#]=bar (# being an actual number) parameter strings to just - // multiple foo=bar strings. This regex should always work since '=' will be urlencoded - // anywhere else the regex isn't expecting it - $queryString = preg_replace('/%5B(?:[0-9]|[1-9][0-9]+)%5D=/', '=', $queryString); + /** + * Response version we support + */ + const SOLR_VERSION = '1.2'; + + /** + * Response writer we support + * + * @todo Solr 1.3 release may change this to SerializedPHP or PHP implementation + */ + const SOLR_WRITER = 'json'; + + /** + * NamedList Treatment constants + */ + const NAMED_LIST_FLAT = 'flat'; + const NAMED_LIST_MAP = 'map'; + + /** + * Servlet mappings + */ + const PING_SERVLET = 'admin/ping'; + const UPDATE_SERVLET = 'update'; + const SEARCH_SERVLET = 'select'; + const THREADS_SERVLET = 'admin/threads'; + + /** + * Server identification strings + * + * @var string + */ + protected $_host, $_port, $_path; + + /** + * Whether {@link Apache_Solr_Response} objects should create {@link Apache_Solr_Document}s in + * the returned parsed data + * + * @var boolean + */ + protected $_createDocuments = true; + + /** + * Whether {@link Apache_Solr_Response} objects should have multivalue fields with only a single value + * collapsed to appear as a single value would. + * + * @var boolean + */ + protected $_collapseSingleValueArrays = true; + + /** + * How NamedLists should be formatted in the output. This specifically effects facet counts. Valid values + * are {@link Apache_Solr_Service::NAMED_LIST_MAP} (default) or {@link Apache_Solr_Service::NAMED_LIST_FLAT}. + * + * @var string + */ + protected $_namedListTreatment = self::NAMED_LIST_MAP; + + /** + * Query delimiters. Someone might want to be able to change + * these (to use & instead of & for example), so I've provided them. + * + * @var string + */ + protected $_queryDelimiter = '?', $_queryStringDelimiter = '&'; + + /** + * Constructed servlet full path URLs + * + * @var string + */ + protected $_updateUrl, $_searchUrl, $_threadsUrl; + + /** + * Keep track of whether our URLs have been constructed + * + * @var boolean + */ + protected $_urlsInited = false; + + /** + * Stream context for posting + * + * @var resource + */ + protected $_postContext; + + /** + * Escape a value for special query characters such as ':', '(', ')', '*', '?', etc. + * + * NOTE: inside a phrase fewer characters need escaped, use {@link Apache_Solr_Service::escapePhrase()} instead + * + * @param string $value + * @return string + */ + static public function escape($value) + { + //list taken from http://lucene.apache.org/java/docs/queryparsersyntax.html#Escaping%20Special%20Characters + $pattern = '/(\+|-|&&|\|\||!|\(|\)|\{|}|\[|]|\^|"|~|\*|\?|:|\\\)/'; + $replace = '\\\$1'; + + return preg_replace($pattern, $replace, $value); + } + + /** + * Escape a value meant to be contained in a phrase for special query characters + * + * @param string $value + * @return string + */ + static public function escapePhrase($value) + { + $pattern = '/("|\\\)/'; + $replace = '\\\$1'; + + return preg_replace($pattern, $replace, $value); + } + + /** + * Convenience function for creating phrase syntax from a value + * + * @param string $value + * @return string + */ + static public function phrase($value) + { + return '"' . self::escapePhrase($value) . '"'; + } + + /** + * Constructor. All parameters are optional and will take on default values + * if not specified. + * + * @param string $host + * @param string $port + * @param string $path + */ + public function __construct($host = 'localhost', $port = 8180, $path = '/solr/') + { + $this->setHost($host); + $this->setPort($port); + $this->setPath($path); + + $this->_initUrls(); + + //set up the stream context for posting with file_get_contents + $contextOpts = array( + 'http' => array( + 'method' => 'POST', + 'header' => "Content-Type: text/xml; charset=UTF-8\r\n" //php.net example showed \r\n at the end + ) + ); + + $this->_postContext = stream_context_create($contextOpts); + } + + /** + * Return a valid http URL given this server's host, port and path and a provided servlet name + * + * @param string $servlet + * @return string + */ + protected function _constructUrl($servlet, $params = array()) + { + if (count($params)) + { + //escape all parameters appropriately for inclusion in the query string + $escapedParams = array(); + + foreach ($params as $key => $value) + { + $escapedParams[] = urlencode($key) . '=' . urlencode($value); + } + + $queryString = $this->_queryDelimiter . implode($this->_queryStringDelimiter, $escapedParams); + } + else + { + $queryString = ''; + } + + return 'http://' . $this->_host . ':' . $this->_port . $this->_path . $servlet . $queryString; + } + + /** + * Construct the Full URLs for the three servlets we reference + */ + protected function _initUrls() + { + //Initialize our full servlet URLs now that we have server information + $this->_updateUrl = $this->_constructUrl(self::UPDATE_SERVLET, array('wt' => self::SOLR_WRITER )); + $this->_searchUrl = $this->_constructUrl(self::SEARCH_SERVLET); + $this->_threadsUrl = $this->_constructUrl(self::THREADS_SERVLET, array('wt' => self::SOLR_WRITER )); + + $this->_urlsInited = true; + } + + /** + * Central method for making a get operation against this Solr Server + * + * @param string $url + * @param float $timeout Read timeout in seconds + * @return Apache_Solr_Response + * + * @todo implement timeout ability + * @throws Exception If a non 200 response status is returned + */ + protected function _sendRawGet($url, $timeout = FALSE) + { + //$http_response_header is set by file_get_contents + $response = new Apache_Solr_Response(@file_get_contents($url), $http_response_header, $this->_createDocuments, $this->_collapseSingleValueArrays); + + if ($response->getHttpStatus() != 200) + { + throw new Exception('"' . $response->getHttpStatus() . '" Status: ' . $response->getHttpStatusMessage(), $response->getHttpStatus()); + } + + return $response; + } + + /** + * Central method for making a post operation against this Solr Server + * + * @param string $url + * @param string $rawPost + * @param float $timeout Read timeout in seconds + * @param string $contentType + * @return Apache_Solr_Response + * + * @throws Exception If a non 200 response status is returned + */ + protected function _sendRawPost($url, $rawPost, $timeout = FALSE, $contentType = 'text/xml; charset=UTF-8') + { + //ensure content type is correct + stream_context_set_option($this->_postContext, 'http', 'header', 'Content-Type: ' . $contentType); + + //set the read timeout if specified + if ($timeout !== FALSE) + { + stream_context_set_option($this->_postContext, 'http', 'timeout', $timeout); + } + + //set the content + stream_context_set_option($this->_postContext, 'http', 'content', $rawPost); + + //$http_response_header is set by file_get_contents + $response = new Apache_Solr_Response(@file_get_contents($url, false, $this->_postContext), $http_response_header, $this->_createDocuments, $this->_collapseSingleValueArrays); + + if ($response->getHttpStatus() != 200) + { + throw new Exception('"' . $response->getHttpStatus() . '" Status: ' . $response->getHttpStatusMessage(), $response->getHttpStatus()); + } + + return $response; + } + + /** + * Returns the set host + * + * @return string + */ + public function getHost() + { + return $this->_host; + } + + /** + * Set the host used. If empty will fallback to constants + * + * @param string $host + */ + public function setHost($host) + { + //Use the provided host or use the default + if (empty($host)) + { + throw new Exception('Host parameter is empty'); + } + else + { + $this->_host = $host; + } + + if ($this->_urlsInited) + { + $this->_initUrls(); + } + } + + /** + * Get the set port + * + * @return integer + */ + public function getPort() + { + return $this->_port; + } + + /** + * Set the port used. If empty will fallback to constants + * + * @param integer $port + */ + public function setPort($port) + { + //Use the provided port or use the default + $port = (int) $port; + + if ($port <= 0) + { + throw new Exception('Port is not a valid port number'); + } + else + { + $this->_port = $port; + } + + if ($this->_urlsInited) + { + $this->_initUrls(); + } + } + + /** + * Get the set path. + * + * @return string + */ + public function getPath() + { + return $this->_path; + } + + /** + * Set the path used. If empty will fallback to constants + * + * @param string $path + */ + public function setPath($path) + { + $path = trim($path, '/'); + + $this->_path = '/' . $path . '/'; + + if ($this->_urlsInited) + { + $this->_initUrls(); + } + } + + /** + * Set the create documents flag. This determines whether {@link Apache_Solr_Response} objects will + * parse the response and create {@link Apache_Solr_Document} instances in place. + * + * @param unknown_type $createDocuments + */ + public function setCreateDocuments($createDocuments) + { + $this->_createDocuments = (bool) $createDocuments; + } + + /** + * Get the current state of teh create documents flag. + * + * @return boolean + */ + public function getCreateDocuments() + { + return $this->_createDocuments; + } + + /** + * Set the collapse single value arrays flag. + * + * @param boolean $collapseSingleValueArrays + */ + public function setCollapseSingleValueArrays($collapseSingleValueArrays) + { + $this->_collapseSingleValueArrays = (bool) $collapseSingleValueArrays; + } + + /** + * Get the current state of the collapse single value arrays flag. + * + * @return boolean + */ + public function getCollapseSingleValueArrays() + { + return $this->_collapseSingleValueArrays; + } + + /** + * Set how NamedLists should be formatted in the response data. This mainly effects + * the facet counts format. + * + * @param string $namedListTreatment + * @throws Exception If invalid option is set + */ + public function setNamedListTreatmet($namedListTreatment) + { + switch ((string) $namedListTreatment) + { + case Apache_Solr_Service::NAMED_LIST_FLAT: + $this->_namedListTreatment = Apache_Solr_Service::NAMED_LIST_FLAT; + break; + + case Apache_Solr_Service::NAMED_LIST_MAP: + $this->_namedListTreatment = Apache_Solr_Service::NAMED_LIST_MAP; + break; + + default: + throw new Exception('Not a valid named list treatement option'); + } + } + + /** + * Get the current setting for named list treatment. + * + * @return string + */ + public function getNamedListTreatment() + { + return $this->_namedListTreatment; + } + + + /** + * Set the string used to separate the path form the query string. + * Defaulted to '?' + * + * @param string $queryDelimiter + */ + public function setQueryDelimiter($queryDelimiter) + { + $this->_queryDelimiter = $queryDelimiter; + } + + /** + * Set the string used to separate the parameters in thequery string + * Defaulted to '&' + * + * @param string $queryStringDelimiter + */ + public function setQueryStringDelimiter($queryStringDelimiter) + { + $this->_queryStringDelimiter = $queryStringDelimiter; + } + + /** + * Call the /admin/ping servlet, can be used to quickly tell if a connection to the + * server is able to be made. + * + * @param float $timeout maximum time to wait for ping in seconds, -1 for unlimited (default is 2) + * @return float Actual time taken to ping the server, FALSE if timeout occurs + */ + public function ping($timeout = 2) + { + $timeout = (float) $timeout; + + if ($timeout <= 0) + { + $timeout = -1; + } + + $start = microtime(true); + + //to prevent strict errors + $errno = 0; + $errstr = ''; + + //try to connect to the host with timeout + $fp = fsockopen($this->_host, $this->_port, $errno, $errstr, $timeout); + + if ($fp) + { + //If we have a timeout set, then determine the amount of time we have left + //in the request and set the stream timeout for the write operation + if ($timeout > 0) + { + //do the calculation + $writeTimeout = $timeout - (microtime(true) - $start); + + //check if we're out of time + if ($writeTimeout <= 0) + { + fclose($fp); + return false; + } + + //convert to microseconds and set the stream timeout + $writeTimeoutInMicroseconds = (int) $writeTimeout * 1000000; + stream_set_timeout($fp, 0, $writeTimeoutInMicroseconds); + } + + $request = 'HEAD ' . $this->_path . self::PING_SERVLET . ' HTTP/1.1' . "\r\n" . + 'host: ' . $this->_host . "\r\n" . + 'Connection: close' . "\r\n" . + "\r\n"; + + fwrite($fp, $request); + + //check the stream meta data to see if we timed out during the operation + $metaData = stream_get_meta_data($fp); + + if (isset($metaData['timeout']) && $metaData['timeout']) + { + fclose($fp); + return false; + } + + + //if we have a timeout set and have made it this far, determine the amount of time + //still remaining and set the timeout appropriately before the read operation + if ($timeout > 0) + { + //do the calculation + $readTimeout = $timeout - (microtime(true) - $start); + + //check if we've run out of time + if ($readTimeout <= 0) + { + fclose($fp); + return false; + } + + //convert to microseconds and set the stream timeout + $readTimeoutInMicroseconds = $readTimeout * 1000000; + stream_set_timeout($fp, 0, $readTimeoutInMicroseconds); + } + + //at the very least we should get a response header line of + //HTTP/1.1 200 OK + $response = fread($fp, 15); + + //check the stream meta data to see if we timed out during the operation + $metaData = stream_get_meta_data($fp); + fclose($fp); //we're done with the connection - ignore the rest + + if (isset($metaData['timeout']) && $metaData['timeout']) + { + return false; + } + + //finally, check the response header line + if ($response != 'HTTP/1.1 200 OK') + { + return false; + } + + //we made it, return the approximate ping time + return microtime(true) - $start; + } + + //we weren't able to make a connection + return false; + } + + /** + * Call the /admin/threads servlet and retrieve information about all threads in the + * Solr servlet's thread group. Useful for diagnostics. + * + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function threads() + { + return $this->_sendRawGet($this->_threadsUrl); + } + + /** + * Raw Add Method. Takes a raw post body and sends it to the update service. Post body + * should be a complete and well formed "add" xml document. + * + * @param string $rawPost + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function add($rawPost) + { + return $this->_sendRawPost($this->_updateUrl, $rawPost); + } + + /** + * Add a Solr Document to the index + * + * @param Apache_Solr_Document $document + * @param boolean $allowDups + * @param boolean $overwritePending + * @param boolean $overwriteCommitted + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function addDocument(Apache_Solr_Document $document, $allowDups = false, $overwritePending = true, $overwriteCommitted = true) + { + $dupValue = $allowDups ? 'true' : 'false'; + $pendingValue = $overwritePending ? 'true' : 'false'; + $committedValue = $overwriteCommitted ? 'true' : 'false'; + + $rawPost = ''; + $rawPost .= $this->_documentToXmlFragment($document); + $rawPost .= ''; + + return $this->add($rawPost); + } + + /** + * Add an array of Solr Documents to the index all at once + * + * @param array $documents Should be an array of Apache_Solr_Document instances + * @param boolean $allowDups + * @param boolean $overwritePending + * @param boolean $overwriteCommitted + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function addDocuments($documents, $allowDups = false, $overwritePending = true, $overwriteCommitted = true) + { + $dupValue = $allowDups ? 'true' : 'false'; + $pendingValue = $overwritePending ? 'true' : 'false'; + $committedValue = $overwriteCommitted ? 'true' : 'false'; + + $rawPost = ''; + + foreach ($documents as $document) + { + if ($document instanceof Apache_Solr_Document) + { + $rawPost .= $this->_documentToXmlFragment($document); + } + } + + $rawPost .= ''; + + return $this->add($rawPost); + } + + /** + * Create an XML fragment from a {@link Apache_Solr_Document} instance appropriate for use inside a Solr add call + * + * @return string + */ + protected function _documentToXmlFragment(Apache_Solr_Document $document) + { + $xml = 'getBoost() !== false) + { + $xml .= ' boost="' . $document->getBoost() . '"'; + } + + $xml .= '>'; + + foreach ($document as $key => $value) + { + $key = htmlspecialchars($key, ENT_QUOTES, 'UTF-8'); + $fieldBoost = $document->getFieldBoost($key); + + if (is_array($value)) + { + foreach ($value as $multivalue) + { + $xml .= '' . "\n"; + } + } + else + { + $xml .= '' . "\n"; + } + } + + $xml .= '' . "\n"; + + return $xml; + } + + /** + * Send a commit command. Will be synchronous unless both wait parameters are set to false. + * + * @param boolean $optimize Defaults to true + * @param boolean $waitFlush Defaults to true + * @param boolean $waitSearcher Defaults to true + * @param float $timeout Maximum expected duration (in seconds) of the commit operation on the server (otherwise, will throw a communication exception). Defaults to 1 hour + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function commit($optimize = true, $waitFlush = true, $waitSearcher = true, $timeout = 3600) + { + $optimizeValue = $optimize ? 'true' : 'false'; + $flushValue = $waitFlush ? 'true' : 'false'; + $searcherValue = $waitSearcher ? 'true' : 'false'; + + $rawPost = ''; + + return $this->_sendRawPost($this->_updateUrl, $rawPost, $timeout); + } + + /** + * Raw Delete Method. Takes a raw post body and sends it to the update service. Body should be + * a complete and well formed "delete" xml document + * + * @param string $rawPost Expected to be utf-8 encoded xml document + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function delete($rawPost) + { + return $this->_sendRawPost($this->_updateUrl, $rawPost); + } + + /** + * Create a delete document based on document ID + * + * @param string $id Expected to be utf-8 encoded + * @param boolean $fromPending + * @param boolean $fromCommitted + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function deleteById($id, $fromPending = true, $fromCommitted = true) + { + $pendingValue = $fromPending ? 'true' : 'false'; + $committedValue = $fromCommitted ? 'true' : 'false'; + + //escape special xml characters + $id = htmlspecialchars($id, ENT_NOQUOTES, 'UTF-8'); + + $rawPost = '' . $id . ''; + + return $this->delete($rawPost); + } + + /** + * Create a delete document based on a query and submit it + * + * @param string $rawQuery Expected to be utf-8 encoded + * @param boolean $fromPending + * @param boolean $fromCommitted + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function deleteByQuery($rawQuery, $fromPending = true, $fromCommitted = true) + { + $pendingValue = $fromPending ? 'true' : 'false'; + $committedValue = $fromCommitted ? 'true' : 'false'; + + // escape special xml characters + $rawQuery = htmlspecialchars($rawQuery, ENT_NOQUOTES, 'UTF-8'); + + $rawPost = '' . $rawQuery . ''; + + return $this->delete($rawPost); + } + + /** + * Send an optimize command. Will be synchronous unless both wait parameters are set + * to false. + * + * @param boolean $waitFlush + * @param boolean $waitSearcher + * @param float $timeout Maximum expected duration of the commit operation on the server (otherwise, will throw a communication exception) + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function optimize($waitFlush = true, $waitSearcher = true, $timeout = 3600) + { + $flushValue = $waitFlush ? 'true' : 'false'; + $searcherValue = $waitSearcher ? 'true' : 'false'; + + $rawPost = ''; + + return $this->_sendRawPost($this->_updateUrl, $rawPost, $timeout); + } + + /** + * Simple Search interface + * + * @param string $query The raw query string + * @param int $offset The starting offset for result documents + * @param int $limit The maximum number of result documents to return + * @param array $params key / value pairs for other query parameters (see Solr documentation), use arrays for parameter keys used more than once (e.g. facet.field) + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function search($query, $offset = 0, $limit = 10, $params = array()) + { + if (!is_array($params)) + { + $params = array(); + } + + // construct our full parameters + // sending the version is important in case the format changes + $params['version'] = self::SOLR_VERSION; + + // common parameters in this interface + $params['wt'] = self::SOLR_WRITER; + $params['json.nl'] = $this->_namedListTreatment; + + $params['q'] = $query; + $params['start'] = $offset; + $params['rows'] = $limit; + + // use http_build_query to encode our arguments because its faster + // than urlencoding all the parts ourselves in a loop + $queryString = http_build_query($params, null, $this->_queryStringDelimiter); + + // because http_build_query treats arrays differently than we want to, correct the query + // string by changing foo[#]=bar (# being an actual number) parameter strings to just + // multiple foo=bar strings. This regex should always work since '=' will be urlencoded + // anywhere else the regex isn't expecting it + $queryString = preg_replace('/%5B(?:[0-9]|[1-9][0-9]+)%5D=/', '=', $queryString); - return $this->_sendRawGet($this->_searchUrl . $this->_queryDelimiter . $queryString); - } + return $this->_sendRawGet($this->_searchUrl . $this->_queryDelimiter . $queryString); + } } \ No newline at end of file Index: SolrPhpClient/Apache/Solr/Service/Balancer.php =================================================================== RCS file: /cvs/drupal-contrib/contributions/modules/apachesolr/SolrPhpClient/Apache/Solr/Service/Balancer.php,v retrieving revision 1.1.4.4 diff -u -r1.1.4.4 Balancer.php --- SolrPhpClient/Apache/Solr/Service/Balancer.php 25 Oct 2008 16:49:25 -0000 1.1.4.4 +++ SolrPhpClient/Apache/Solr/Service/Balancer.php 14 Dec 2008 11:16:09 -0000 @@ -1,5 +1,4 @@ addReadService($service); - } - - //setup writeable services - foreach ($writeableServices as $service) - { - $this->addWriteService($service); - } - } - - public function setReadPingTimeout($timeout) - { - $this->_readPingTimeout = $timeout; - } - - public function setWritePingTimeout($timeout) - { - $this->_writePingTimeout = $timeout; - } - - public function setUseBackoff($enable) - { - $this->_useBackoff = $enable; - } - - /** - * Generates a service ID - * - * @param string $host - * @param integer $port - * @param string $path - * @return string - */ - protected function _getServiceId($host, $port, $path) - { - return $host . ':' . $port . $path; - } - - /** - * Adds a service instance or service descriptor (if it is already - * not added) - * - * @param mixed $service - * - * @throws Exception If service descriptor is not valid - */ - public function addReadService($service) - { - if ($service instanceof Apache_Solr_Service) - { - $id = $this->_getServiceId($service->getHost(), $service->getPort(), $service->getPath()); - - $this->_readableServices[$id] = $service; - } - else if (is_array($service)) - { - if (isset($service['host']) && isset($service['port']) && isset($service['path'])) - { - $id = $this->_getServiceId((string)$service['host'], (int)$service['port'], (string)$service['path']); - - $this->_readableServices[$id] = $service; - } - else - { - throw new Exception('A Readable Service description array does not have all required elements of host, port, and path'); - } - } - } - - /** - * Removes a service instance or descriptor from the available services - * - * @param mixed $service - * - * @throws Exception If service descriptor is not valid - */ - public function removeReadService($service) - { - $id = ''; - - if ($service instanceof Apache_Solr_Service) - { - $id = $this->_getServiceId($service->getHost(), $service->getPort(), $service->getPath()); - } - else if (is_array($service)) - { - if (isset($service['host']) && isset($service['port']) && isset($service['path'])) - { - $id = $this->_getServiceId((string)$service['host'], (int)$service['port'], (string)$service['path']); - } - else - { - throw new Exception('A Readable Service description array does not have all required elements of host, port, and path'); - } - } - else if (is_string($service)) - { - $id = $service; - } - - if ($id && isset($this->_readableServices[$id])) - { - unset($this->_readableServices[$id]); - } - } - - /** - * Adds a service instance or service descriptor (if it is already - * not added) - * - * @param mixed $service - * - * @throws Exception If service descriptor is not valid - */ - public function addWriteService($service) - { - if ($service instanceof Apache_Solr_Service) - { - $id = $this->_getServiceId($service->getHost(), $service->getPort(), $service->getPath()); - - $this->_writeableServices[$id] = $service; - } - else if (is_array($service)) - { - if (isset($service['host']) && isset($service['port']) && isset($service['path'])) - { - $id = $this->_getServiceId((string)$service['host'], (int)$service['port'], (string)$service['path']); - - $this->_writeableServices[$id] = $service; - } - else - { - throw new Exception('A Writeable Service description array does not have all required elements of host, port, and path'); - } - } - } - - /** - * Removes a service instance or descriptor from the available services - * - * @param mixed $service - * - * @throws Exception If service descriptor is not valid - */ - public function removeWriteService($service) - { - $id = ''; - - if ($service instanceof Apache_Solr_Service) - { - $id = $this->_getServiceId($service->getHost(), $service->getPort(), $service->getPath()); - } - else if (is_array($service)) - { - if (isset($service['host']) && isset($service['port']) && isset($service['path'])) - { - $id = $this->_getServiceId((string)$service['host'], (int)$service['port'], (string)$service['path']); - } - else - { - throw new Exception('A Readable Service description array does not have all required elements of host, port, and path'); - } - } - else if (is_string($service)) - { - $id = $service; - } - - if ($id && isset($this->_writeableServices[$id])) - { - unset($this->_writeableServices[$id]); - } - } - - /** - * Iterate through available read services and select the first with a ping - * that satisfies configured timeout restrictions (or the default) - * - * @return Apache_Solr_Service - * - * @throws Exception If there are no read services that meet requirements - */ - protected function _selectReadService($forceSelect = false) - { - if (!$this->_currentReadService || !isset($this->_readableServices[$this->_currentReadService]) || $forceSelect) - { - if ($this->_currentReadService && isset($this->_readableServices[$this->_currentReadService]) && $forceSelect) - { - // we probably had a communication error, ping the current read service, remove it if it times out - if ($this->_readableServices[$this->_currentReadService]->ping($this->_readPingTimeout) === false) - { - $this->removeReadService($this->_currentReadService); - } - } - - if (count($this->_readableServices)) - { - // select one of the read services at random - $ids = array_keys($this->_readableServices); - - $id = $ids[rand(0, count($ids) - 1)]; - $service = $this->_readableServices[$id]; - - if (is_array($service)) - { - //convert the array definition to a client object - $service = new Apache_Solr_Service($service['host'], $service['port'], $service['path']); - $this->_readableServices[$id] = $service; - } - - $service->setCreateDocuments($this->_createDocuments); - $this->_currentReadService = $id; - } - else - { - throw new Exception('No read services were available'); - } - } - - return $this->_readableServices[$this->_currentReadService]; - } - - /** - * Iterate through available write services and select the first with a ping - * that satisfies configured timeout restrictions (or the default) - * - * @return Apache_Solr_Service - * - * @throws Exception If there are no write services that meet requirements - */ - protected function _selectWriteService($forceSelect = false) - { - if($this->_useBackoff) - { - return $this->_selectWriteServiceSafe($forceSelect); - } - - if (!$this->_currentWriteService || !isset($this->_writeableServices[$this->_currentWriteService]) || $forceSelect) - { - if ($this->_currentWriteService && isset($this->_writeableServices[$this->_currentWriteService]) && $forceSelect) - { - // we probably had a communication error, ping the current read service, remove it if it times out - if ($this->_writeableServices[$this->_currentWriteService]->ping($this->_writePingTimeout) === false) - { - $this->removeWriteService($this->_currentWriteService); - } - } - - if (count($this->_writeableServices)) - { - // select one of the read services at random - $ids = array_keys($this->_writeableServices); - - $id = $ids[rand(0, count($ids) - 1)]; - $service = $this->_writeableServices[$id]; - - if (is_array($service)) - { - //convert the array definition to a client object - $service = new Apache_Solr_Service($service['host'], $service['port'], $service['path']); - $this->_writeableServices[$id] = $service; - } - - $this->_currentWriteService = $id; - } - else - { - throw new Exception('No write services were available'); - } - } - - return $this->_writeableServices[$this->_currentWriteService]; - } - - /** - * Iterate through available write services and select the first with a ping - * that satisfies configured timeout restrictions (or the default). The - * timeout period will increase until a connection is made or the limit is - * reached. This will allow for increased reliability with heavily loaded - * server(s). - * - * @return Apache_Solr_Service - * - * @throws Exception If there are no write services that meet requirements - */ - - protected function _selectWriteServiceSafe($forceSelect = false) - { - if (!$this->_currentWriteService || !isset($this->_writeableServices[$this->_currentWriteService]) || $forceSelect) - { - if (count($this->_writeableServices)) - { - $backoff = $this->_defaultBackoff; - - do { - // select one of the read services at random - $ids = array_keys($this->_writeableServices); - - $id = $ids[rand(0, count($ids) - 1)]; - $service = $this->_writeableServices[$id]; - - if (is_array($service)) - { - //convert the array definition to a client object - $service = new Apache_Solr_Service($service['host'], $service['port'], $service['path']); - $this->_writeableServices[$id] = $service; - } - - $this->_currentWriteService = $id; - - $backoff *= $this->_backoffEscalation; - - if($backoff > $this->_backoffLimit) - { - throw new Exception('No write services were available. All timeouts exceeded.'); - } - - } while($this->_writeableServices[$this->_currentWriteService]->ping($backoff) === false); - } - else - { - throw new Exception('No write services were available'); - } - } - - return $this->_writeableServices[$this->_currentWriteService]; - } - - public function setCreateDocuments($createDocuments) - { - $this->_createDocuments = (bool) $createDocuments; - - if ($this->_currentReadService) - { - $service = $this->_selectReadService(); - $service->setCreateDocuments($createDocuments); - } - } - - public function getCreateDocuments() - { - return $this->_createDocuments; - } - - /** - * Raw Add Method. Takes a raw post body and sends it to the update service. Post body - * should be a complete and well formed "add" xml document. - * - * @param string $rawPost - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function add($rawPost) - { - $service = $this->_selectWriteService(); - - do - { - try - { - return $service->add($rawPost); - } - catch (Exception $e) - { - if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR - { - throw $e; - } - } - - $service = $this->_selectWriteService(true); - } while ($service); - - return false; - } - - /** - * Add a Solr Document to the index - * - * @param Apache_Solr_Document $document - * @param boolean $allowDups - * @param boolean $overwritePending - * @param boolean $overwriteCommitted - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function addDocument(Apache_Solr_Document $document, $allowDups = false, $overwritePending = true, $overwriteCommitted = true) - { - $service = $this->_selectWriteService(); - - do - { - try - { - return $service->addDocument($document, $allowDups, $overwritePending, $overwriteCommitted); - } - catch (Exception $e) - { - if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR - { - throw $e; - } - } - - $service = $this->_selectWriteService(true); - } while ($service); - - return false; - } - - /** - * Add an array of Solr Documents to the index all at once - * - * @param array $documents Should be an array of Apache_Solr_Document instances - * @param boolean $allowDups - * @param boolean $overwritePending - * @param boolean $overwriteCommitted - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function addDocuments($documents, $allowDups = false, $overwritePending = true, $overwriteCommitted = true) - { - $service = $this->_selectWriteService(); - - do - { - try - { - return $service->addDocuments($documents, $allowDups, $overwritePending, $overwriteCommitted); - } - catch (Exception $e) - { - if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR - { - throw $e; - } - } - - $service = $this->_selectWriteService(true); - } while ($service); - - return false; - } - - /** - * Send a commit command. Will be synchronous unless both wait parameters are set - * to false. - * - * @param boolean $waitFlush - * @param boolean $waitSearcher - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function commit($optimize = true, $waitFlush = true, $waitSearcher = true, $timeout = 3600) - { - $service = $this->_selectWriteService(); - - do - { - try - { - return $service->commit($optimize, $waitFlush, $waitSearcher, $timeout); - } - catch (Exception $e) - { - if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR - { - throw $e; - } - } - - $service = $this->_selectWriteService(true); - } while ($service); - - return false; - } - - /** - * Raw Delete Method. Takes a raw post body and sends it to the update service. Body should be - * a complete and well formed "delete" xml document - * - * @param string $rawPost - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function delete($rawPost) - { - $service = $this->_selectWriteService(); - - do - { - try - { - return $service->delete($rawPost); - } - catch (Exception $e) - { - if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR - { - throw $e; - } - } - - $service = $this->_selectWriteService(true); - } while ($service); - - return false; - } - - /** - * Create a delete document based on document ID - * - * @param string $id - * @param boolean $fromPending - * @param boolean $fromCommitted - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function deleteById($id, $fromPending = true, $fromCommitted = true) - { - $service = $this->_selectWriteService(); - - do - { - try - { - return $service->deleteById($id, $fromPending, $fromCommitted); - } - catch (Exception $e) - { - if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR - { - throw $e; - } - } - - $service = $this->_selectWriteService(true); - } while ($service); - - return false; - } - - /** - * Create a delete document based on a query and submit it - * - * @param string $rawQuery - * @param boolean $fromPending - * @param boolean $fromCommitted - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function deleteByQuery($rawQuery, $fromPending = true, $fromCommitted = true) - { - $service = $this->_selectWriteService(); - - do - { - try - { - return $service->deleteByQuery($rawQuery, $fromPending, $fromCommitted); - } - catch (Exception $e) - { - if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR - { - throw $e; - } - } - - $service = $this->_selectWriteService(true); - } while ($service); - - return false; - } - - /** - * Send an optimize command. Will be synchronous unless both wait parameters are set - * to false. - * - * @param boolean $waitFlush - * @param boolean $waitSearcher - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function optimize($waitFlush = true, $waitSearcher = true) - { - $service = $this->_selectWriteService(); - - do - { - try - { - return $service->optimize($waitFlush, $waitSearcher); - } - catch (Exception $e) - { - if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR - { - throw $e; - } - } - - $service = $this->_selectWriteService(true); - } while ($service); - - return false; - } - - /** - * Simple Search interface - * - * @param string $query The raw query string - * @param int $offset The starting offset for result documents - * @param int $limit The maximum number of result documents to return - * @param array $params key / value pairs for query parameters, use arrays for multivalued parameters - * @return Apache_Solr_Response - * - * @throws Exception If an error occurs during the service call - */ - public function search($query, $offset = 0, $limit = 10, $params = array()) - { - $service = $this->_selectReadService(); - - do - { - try - { - return $service->search($query, $offset, $limit, $params); - } - catch (Exception $e) - { - if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR - { - throw $e; - } - } - - $service = $this->_selectReadService(true); - } while ($service); - - return false; - } + protected $_readPingTimeout = 2; + protected $_writePingTimeout = 4; + + // Configuration for server selection backoff intervals + protected $_useBackoff = false; // Set to true to use more resillient write server selection + protected $_backoffLimit = 600; // 10 minute default maximum + protected $_backoffEscalation = 2.0; // Rate at which to increase backoff period + protected $_defaultBackoff = 2.0; // Default backoff interval + + /** + * Escape a value for special query characters such as ':', '(', ')', '*', '?', etc. + * + * NOTE: inside a phrase fewer characters need escaped, use {@link Apache_Solr_Service::escapePhrase()} instead + * + * @param string $value + * @return string + */ + static public function escape($value) + { + return Apache_Solr_Service::escape($value); + } + + /** + * Escape a value meant to be contained in a phrase for special query characters + * + * @param string $value + * @return string + */ + static public function escapePhrase($value) + { + return Apache_Solr_Service::escapePhrase($value); + } + + /** + * Convenience function for creating phrase syntax from a value + * + * @param string $value + * @return string + */ + static public function phrase($value) + { + return Apache_Solr_Service::phrase($value); + } + + /** + * Constructor. Takes arrays of read and write service instances or descriptions + * + * @param array $readableServices + * @param array $writeableServices + */ + public function __construct($readableServices = array(), $writeableServices = array()) + { + //setup readable services + foreach ($readableServices as $service) + { + $this->addReadService($service); + } + + //setup writeable services + foreach ($writeableServices as $service) + { + $this->addWriteService($service); + } + } + + public function setReadPingTimeout($timeout) + { + $this->_readPingTimeout = $timeout; + } + + public function setWritePingTimeout($timeout) + { + $this->_writePingTimeout = $timeout; + } + + public function setUseBackoff($enable) + { + $this->_useBackoff = $enable; + } + + /** + * Generates a service ID + * + * @param string $host + * @param integer $port + * @param string $path + * @return string + */ + protected function _getServiceId($host, $port, $path) + { + return $host . ':' . $port . $path; + } + + /** + * Adds a service instance or service descriptor (if it is already + * not added) + * + * @param mixed $service + * + * @throws Exception If service descriptor is not valid + */ + public function addReadService($service) + { + if ($service instanceof Apache_Solr_Service) + { + $id = $this->_getServiceId($service->getHost(), $service->getPort(), $service->getPath()); + + $this->_readableServices[$id] = $service; + } + else if (is_array($service)) + { + if (isset($service['host']) && isset($service['port']) && isset($service['path'])) + { + $id = $this->_getServiceId((string)$service['host'], (int)$service['port'], (string)$service['path']); + + $this->_readableServices[$id] = $service; + } + else + { + throw new Exception('A Readable Service description array does not have all required elements of host, port, and path'); + } + } + } + + /** + * Removes a service instance or descriptor from the available services + * + * @param mixed $service + * + * @throws Exception If service descriptor is not valid + */ + public function removeReadService($service) + { + $id = ''; + + if ($service instanceof Apache_Solr_Service) + { + $id = $this->_getServiceId($service->getHost(), $service->getPort(), $service->getPath()); + } + else if (is_array($service)) + { + if (isset($service['host']) && isset($service['port']) && isset($service['path'])) + { + $id = $this->_getServiceId((string)$service['host'], (int)$service['port'], (string)$service['path']); + } + else + { + throw new Exception('A Readable Service description array does not have all required elements of host, port, and path'); + } + } + else if (is_string($service)) + { + $id = $service; + } + + if ($id && isset($this->_readableServices[$id])) + { + unset($this->_readableServices[$id]); + } + } + + /** + * Adds a service instance or service descriptor (if it is already + * not added) + * + * @param mixed $service + * + * @throws Exception If service descriptor is not valid + */ + public function addWriteService($service) + { + if ($service instanceof Apache_Solr_Service) + { + $id = $this->_getServiceId($service->getHost(), $service->getPort(), $service->getPath()); + + $this->_writeableServices[$id] = $service; + } + else if (is_array($service)) + { + if (isset($service['host']) && isset($service['port']) && isset($service['path'])) + { + $id = $this->_getServiceId((string)$service['host'], (int)$service['port'], (string)$service['path']); + + $this->_writeableServices[$id] = $service; + } + else + { + throw new Exception('A Writeable Service description array does not have all required elements of host, port, and path'); + } + } + } + + /** + * Removes a service instance or descriptor from the available services + * + * @param mixed $service + * + * @throws Exception If service descriptor is not valid + */ + public function removeWriteService($service) + { + $id = ''; + + if ($service instanceof Apache_Solr_Service) + { + $id = $this->_getServiceId($service->getHost(), $service->getPort(), $service->getPath()); + } + else if (is_array($service)) + { + if (isset($service['host']) && isset($service['port']) && isset($service['path'])) + { + $id = $this->_getServiceId((string)$service['host'], (int)$service['port'], (string)$service['path']); + } + else + { + throw new Exception('A Readable Service description array does not have all required elements of host, port, and path'); + } + } + else if (is_string($service)) + { + $id = $service; + } + + if ($id && isset($this->_writeableServices[$id])) + { + unset($this->_writeableServices[$id]); + } + } + + /** + * Iterate through available read services and select the first with a ping + * that satisfies configured timeout restrictions (or the default) + * + * @return Apache_Solr_Service + * + * @throws Exception If there are no read services that meet requirements + */ + protected function _selectReadService($forceSelect = false) + { + if (!$this->_currentReadService || !isset($this->_readableServices[$this->_currentReadService]) || $forceSelect) + { + if ($this->_currentReadService && isset($this->_readableServices[$this->_currentReadService]) && $forceSelect) + { + // we probably had a communication error, ping the current read service, remove it if it times out + if ($this->_readableServices[$this->_currentReadService]->ping($this->_readPingTimeout) === false) + { + $this->removeReadService($this->_currentReadService); + } + } + + if (count($this->_readableServices)) + { + // select one of the read services at random + $ids = array_keys($this->_readableServices); + + $id = $ids[rand(0, count($ids) - 1)]; + $service = $this->_readableServices[$id]; + + if (is_array($service)) + { + //convert the array definition to a client object + $service = new Apache_Solr_Service($service['host'], $service['port'], $service['path']); + $this->_readableServices[$id] = $service; + } + + $service->setCreateDocuments($this->_createDocuments); + $this->_currentReadService = $id; + } + else + { + throw new Exception('No read services were available'); + } + } + + return $this->_readableServices[$this->_currentReadService]; + } + + /** + * Iterate through available write services and select the first with a ping + * that satisfies configured timeout restrictions (or the default) + * + * @return Apache_Solr_Service + * + * @throws Exception If there are no write services that meet requirements + */ + protected function _selectWriteService($forceSelect = false) + { + if($this->_useBackoff) + { + return $this->_selectWriteServiceSafe($forceSelect); + } + + if (!$this->_currentWriteService || !isset($this->_writeableServices[$this->_currentWriteService]) || $forceSelect) + { + if ($this->_currentWriteService && isset($this->_writeableServices[$this->_currentWriteService]) && $forceSelect) + { + // we probably had a communication error, ping the current read service, remove it if it times out + if ($this->_writeableServices[$this->_currentWriteService]->ping($this->_writePingTimeout) === false) + { + $this->removeWriteService($this->_currentWriteService); + } + } + + if (count($this->_writeableServices)) + { + // select one of the read services at random + $ids = array_keys($this->_writeableServices); + + $id = $ids[rand(0, count($ids) - 1)]; + $service = $this->_writeableServices[$id]; + + if (is_array($service)) + { + //convert the array definition to a client object + $service = new Apache_Solr_Service($service['host'], $service['port'], $service['path']); + $this->_writeableServices[$id] = $service; + } + + $this->_currentWriteService = $id; + } + else + { + throw new Exception('No write services were available'); + } + } + + return $this->_writeableServices[$this->_currentWriteService]; + } + + /** + * Iterate through available write services and select the first with a ping + * that satisfies configured timeout restrictions (or the default). The + * timeout period will increase until a connection is made or the limit is + * reached. This will allow for increased reliability with heavily loaded + * server(s). + * + * @return Apache_Solr_Service + * + * @throws Exception If there are no write services that meet requirements + */ + + protected function _selectWriteServiceSafe($forceSelect = false) + { + if (!$this->_currentWriteService || !isset($this->_writeableServices[$this->_currentWriteService]) || $forceSelect) + { + if (count($this->_writeableServices)) + { + $backoff = $this->_defaultBackoff; + + do { + // select one of the read services at random + $ids = array_keys($this->_writeableServices); + + $id = $ids[rand(0, count($ids) - 1)]; + $service = $this->_writeableServices[$id]; + + if (is_array($service)) + { + //convert the array definition to a client object + $service = new Apache_Solr_Service($service['host'], $service['port'], $service['path']); + $this->_writeableServices[$id] = $service; + } + + $this->_currentWriteService = $id; + + $backoff *= $this->_backoffEscalation; + + if($backoff > $this->_backoffLimit) + { + throw new Exception('No write services were available. All timeouts exceeded.'); + } + + } while($this->_writeableServices[$this->_currentWriteService]->ping($backoff) === false); + } + else + { + throw new Exception('No write services were available'); + } + } + + return $this->_writeableServices[$this->_currentWriteService]; + } + + public function setCreateDocuments($createDocuments) + { + $this->_createDocuments = (bool) $createDocuments; + + if ($this->_currentReadService) + { + $service = $this->_selectReadService(); + $service->setCreateDocuments($createDocuments); + } + } + + public function getCreateDocuments() + { + return $this->_createDocuments; + } + + /** + * Raw Add Method. Takes a raw post body and sends it to the update service. Post body + * should be a complete and well formed "add" xml document. + * + * @param string $rawPost + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function add($rawPost) + { + $service = $this->_selectWriteService(); + + do + { + try + { + return $service->add($rawPost); + } + catch (Exception $e) + { + if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR + { + throw $e; + } + } + + $service = $this->_selectWriteService(true); + } while ($service); + + return false; + } + + /** + * Add a Solr Document to the index + * + * @param Apache_Solr_Document $document + * @param boolean $allowDups + * @param boolean $overwritePending + * @param boolean $overwriteCommitted + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function addDocument(Apache_Solr_Document $document, $allowDups = false, $overwritePending = true, $overwriteCommitted = true) + { + $service = $this->_selectWriteService(); + + do + { + try + { + return $service->addDocument($document, $allowDups, $overwritePending, $overwriteCommitted); + } + catch (Exception $e) + { + if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR + { + throw $e; + } + } + + $service = $this->_selectWriteService(true); + } while ($service); + + return false; + } + + /** + * Add an array of Solr Documents to the index all at once + * + * @param array $documents Should be an array of Apache_Solr_Document instances + * @param boolean $allowDups + * @param boolean $overwritePending + * @param boolean $overwriteCommitted + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function addDocuments($documents, $allowDups = false, $overwritePending = true, $overwriteCommitted = true) + { + $service = $this->_selectWriteService(); + + do + { + try + { + return $service->addDocuments($documents, $allowDups, $overwritePending, $overwriteCommitted); + } + catch (Exception $e) + { + if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR + { + throw $e; + } + } + + $service = $this->_selectWriteService(true); + } while ($service); + + return false; + } + + /** + * Send a commit command. Will be synchronous unless both wait parameters are set + * to false. + * + * @param boolean $waitFlush + * @param boolean $waitSearcher + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function commit($optimize = true, $waitFlush = true, $waitSearcher = true, $timeout = 3600) + { + $service = $this->_selectWriteService(); + + do + { + try + { + return $service->commit($optimize, $waitFlush, $waitSearcher, $timeout); + } + catch (Exception $e) + { + if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR + { + throw $e; + } + } + + $service = $this->_selectWriteService(true); + } while ($service); + + return false; + } + + /** + * Raw Delete Method. Takes a raw post body and sends it to the update service. Body should be + * a complete and well formed "delete" xml document + * + * @param string $rawPost + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function delete($rawPost) + { + $service = $this->_selectWriteService(); + + do + { + try + { + return $service->delete($rawPost); + } + catch (Exception $e) + { + if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR + { + throw $e; + } + } + + $service = $this->_selectWriteService(true); + } while ($service); + + return false; + } + + /** + * Create a delete document based on document ID + * + * @param string $id + * @param boolean $fromPending + * @param boolean $fromCommitted + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function deleteById($id, $fromPending = true, $fromCommitted = true) + { + $service = $this->_selectWriteService(); + + do + { + try + { + return $service->deleteById($id, $fromPending, $fromCommitted); + } + catch (Exception $e) + { + if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR + { + throw $e; + } + } + + $service = $this->_selectWriteService(true); + } while ($service); + + return false; + } + + /** + * Create a delete document based on a query and submit it + * + * @param string $rawQuery + * @param boolean $fromPending + * @param boolean $fromCommitted + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function deleteByQuery($rawQuery, $fromPending = true, $fromCommitted = true) + { + $service = $this->_selectWriteService(); + + do + { + try + { + return $service->deleteByQuery($rawQuery, $fromPending, $fromCommitted); + } + catch (Exception $e) + { + if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR + { + throw $e; + } + } + + $service = $this->_selectWriteService(true); + } while ($service); + + return false; + } + + /** + * Send an optimize command. Will be synchronous unless both wait parameters are set + * to false. + * + * @param boolean $waitFlush + * @param boolean $waitSearcher + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function optimize($waitFlush = true, $waitSearcher = true) + { + $service = $this->_selectWriteService(); + + do + { + try + { + return $service->optimize($waitFlush, $waitSearcher); + } + catch (Exception $e) + { + if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR + { + throw $e; + } + } + + $service = $this->_selectWriteService(true); + } while ($service); + + return false; + } + + /** + * Simple Search interface + * + * @param string $query The raw query string + * @param int $offset The starting offset for result documents + * @param int $limit The maximum number of result documents to return + * @param array $params key / value pairs for query parameters, use arrays for multivalued parameters + * @return Apache_Solr_Response + * + * @throws Exception If an error occurs during the service call + */ + public function search($query, $offset = 0, $limit = 10, $params = array()) + { + $service = $this->_selectReadService(); + + do + { + try + { + return $service->search($query, $offset, $limit, $params); + } + catch (Exception $e) + { + if ($e->getCode() != 0) //IF NOT COMMUNICATION ERROR + { + throw $e; + } + } + + $service = $this->_selectReadService(true); + } while ($service); + + return false; + } }