def validate(self): for key, value in self.items(): if key in self.BOOLEAN_OPTIONS: try: value = asbool(value) except ValueError: raise SearchQueryError('Value for search option %r must be True or False (1 or 0) but received %r' % (key, value)) elif key in self.INTEGER_OPTIONS: try: value = int(value) except ValueError: raise SearchQueryError('Value for search option %r must be an integer but received %r' % (key, value)) elif key in self.UNSUPPORTED_OPTIONS: raise SearchQueryError('Search option %r is not supported' % key) self[key] = value
def convert_legacy_parameters_to_solr( legacy_params: dict[str, Any]) -> dict[str, Any]: '''API v1 and v2 allowed search params that the SOLR syntax does not support, so use this function to convert those to SOLR syntax. See tests for examples. raises SearchQueryError on invalid params. ''' options = QueryOptions(**legacy_params) options.validate() solr_params = legacy_params.copy() solr_q_list: list[str] = [] if solr_params.get('q'): solr_q_list.append(solr_params['q'].replace('+', ' ')) non_solr_params = set(legacy_params.keys()) - VALID_SOLR_PARAMETERS for search_key in non_solr_params: value_obj = legacy_params[search_key] value = value_obj.replace('+', ' ') if isinstance(value_obj, str) else value_obj if search_key == 'all_fields': if value: solr_params['fl'] = '*' elif search_key == 'offset': solr_params['start'] = value elif search_key == 'limit': solr_params['rows'] = value elif search_key == 'order_by': solr_params['sort'] = '%s asc' % value elif search_key == 'tags': if isinstance(value_obj, list): tag_list = value_obj elif isinstance(value_obj, str): tag_list = [value_obj] else: raise SearchQueryError( 'Was expecting either a string or JSON list for the tags parameter: %r' % value) solr_q_list.extend([ 'tags:"%s"' % escape_legacy_argument(tag) for tag in tag_list ]) else: if len(value.strip()): value = escape_legacy_argument(value) if ' ' in value: value = '"%s"' % value solr_q_list.append('%s:%s' % (search_key, value)) del solr_params[search_key] solr_params['q'] = ' '.join(solr_q_list) if non_solr_params: log.debug('Converted legacy search params from %r to %r', legacy_params, solr_params) return solr_params
def run(self, query): ''' Performs a dataset search using the given query. @param query - dictionary with keys like: q, fq, sort, rows, facet @return - dictionary with keys results and count May raise SearchQueryError or SearchError. ''' assert isinstance(query, (dict, MultiDict)) # check that query keys are valid if not set(query.keys()) <= VALID_SOLR_PARAMETERS: invalid_params = [ s for s in set(query.keys()) - VALID_SOLR_PARAMETERS ] raise SearchQueryError("Invalid search parameters: %s" % invalid_params) # default query is to return all documents q = query.get('q') if not q or q == '""' or q == "''": query['q'] = "*:*" # number of results rows_to_return = min(1000, int(query.get('rows', 10))) if rows_to_return > 0: # #1683 Work around problem of last result being out of order # in SOLR 1.4 rows_to_query = rows_to_return + 1 else: rows_to_query = rows_to_return query['rows'] = rows_to_query # order by score if no 'sort' term given order_by = query.get('sort') if order_by == 'rank' or order_by is None: query['sort'] = 'score desc, name asc' # show only results from this CKAN instance fq = query.get('fq', '') if not '+site_id:' in fq: fq += ' +site_id:"%s"' % config.get('ckan.site_id') # filter for package status if not '+state:' in fq: fq += " +state:active" query['fq'] = fq # faceting query['facet'] = query.get('facet', 'true') query['facet.limit'] = query.get( 'facet.limit', config.get('search.facets.limit', '50')) query['facet.mincount'] = query.get('facet.mincount', 1) # return the package ID and search scores query['fl'] = query.get('fl', 'name') # return results as json encoded string query['wt'] = query.get('wt', 'json') # If the query has a colon in it then consider it a fielded search and do use dismax. if ':' not in query['q']: query['defType'] = 'dismax' query['tie'] = '0.1' # this minimum match is explained # http://wiki.apache.org/solr/DisMaxQParserPlugin#mm_.28Minimum_.27Should.27_Match.29 query['mm'] = '2<-1 5<80%' query['qf'] = query.get('qf', QUERY_FIELDS) conn = make_connection() log.debug('Package query: %r' % query) try: solr_response = conn.raw_query(**query) except SolrException, e: raise SearchError( 'SOLR returned an error running query: %r Error: %r' % (query, e.reason))
def run(self, query, permission_labels=None, **kwargs): ''' Performs a dataset search using the given query. :param query: dictionary with keys like: q, fq, sort, rows, facet :type query: dict :param permission_labels: filter results to those that include at least one of these labels. None to not filter (return everything) :type permission_labels: list of unicode strings; or None :returns: dictionary with keys results and count May raise SearchQueryError or SearchError. ''' assert isinstance(query, (dict, MultiDict)) # check that query keys are valid if not set(query.keys()) <= VALID_SOLR_PARAMETERS: invalid_params = [ s for s in set(query.keys()) - VALID_SOLR_PARAMETERS ] raise SearchQueryError("Invalid search parameters: %s" % invalid_params) # default query is to return all documents q = query.get('q') if not q or q == '""' or q == "''": query['q'] = "*:*" # number of results rows_to_return = min(1000, int(query.get('rows', 10))) if rows_to_return > 0: # #1683 Work around problem of last result being out of order # in SOLR 1.4 rows_to_query = rows_to_return + 1 else: rows_to_query = rows_to_return query['rows'] = rows_to_query fq = [] if 'fq' in query: fq.append(query['fq']) fq.extend(query.get('fq_list', [])) # show only results from this CKAN instance fq.append('+site_id:%s' % solr_literal(config.get('ckan.site_id'))) # filter for package status if not '+state:' in query.get('fq', ''): fq.append('+state:active') # only return things we should be able to see if permission_labels is not None: fq.append('+permission_labels:(%s)' % ' OR '.join(solr_literal(p) for p in permission_labels)) query['fq'] = fq # faceting query['facet'] = query.get('facet', 'true') query['facet.limit'] = query.get( 'facet.limit', config.get('search.facets.limit', '50')) query['facet.mincount'] = query.get('facet.mincount', 1) # return the package ID and search scores query['fl'] = query.get('fl', 'name') # return results as json encoded string query['wt'] = query.get('wt', 'json') # If the query has a colon in it then consider it a fielded search and do use dismax. defType = query.get('defType', 'dismax') if ':' not in query['q'] or defType == 'edismax': query['defType'] = defType query['tie'] = query.get('tie', '0.1') # this minimum match is explained # http://wiki.apache.org/solr/DisMaxQParserPlugin#mm_.28Minimum_.27Should.27_Match.29 query['mm'] = query.get('mm', '2<-1 5<80%') query['qf'] = query.get('qf', QUERY_FIELDS) conn = make_connection(decode_dates=False) log.debug('Package query: %r' % query) try: solr_response = conn.search(**query) except pysolr.SolrError as e: # Error with the sort parameter. You see slightly different # error messages depending on whether the SOLR JSON comes back # or Jetty gets in the way converting it to HTML - not sure why # if e.args and isinstance(e.args[0], str): if "Can't determine a Sort Order" in e.args[0] or \ "Can't determine Sort Order" in e.args[0] or \ 'Unknown sort order' in e.args[0]: raise SearchQueryError('Invalid "sort" parameter') raise SearchError( 'SOLR returned an error running query: %r Error: %r' % (query, e)) self.count = solr_response.hits self.results = solr_response.docs # #1683 Filter out the last row that is sometimes out of order self.results = self.results[:rows_to_return] # get any extras and add to 'extras' dict for result in self.results: extra_keys = filter(lambda x: x.startswith('extras_'), result.keys()) extras = {} for extra_key in extra_keys: value = result.pop(extra_key) extras[extra_key[len('extras_'):]] = value if extra_keys: result['extras'] = extras # if just fetching the id or name, return a list instead of a dict if query.get('fl') in ['id', 'name']: self.results = [r.get(query.get('fl')) for r in self.results] # get facets and convert facets list to a dict self.facets = solr_response.facets.get('facet_fields', {}) for field, values in six.iteritems(self.facets): self.facets[field] = dict(zip(values[0::2], values[1::2])) return {'results': self.results, 'count': self.count}
def run(self, query, permission_labels=None, **kwargs): ''' Performs a dataset search using the given query. :param query: dictionary with keys like: q, fq, sort, rows, facet :type query: dict :param permission_labels: filter results to those that include at least one of these labels. None to not filter (return everything) :type permission_labels: list of unicode strings; or None :returns: dictionary with keys results and count May raise SearchQueryError or SearchError. ''' assert isinstance(query, (dict, MultiDict)) # check that query keys are valid valid_solr_parameters = VALID_SOLR_PARAMETERS for item in plugins.PluginImplementations(plugins.IPackageController): if 'update_valid_solr_parameters' in dir(item): valid_solr_parameters = item.update_valid_solr_parameters( valid_solr_parameters) if not set(query.keys()) <= valid_solr_parameters: invalid_params = [ s for s in set(query.keys()) - valid_solr_parameters ] raise SearchQueryError("Invalid search parameters: %s" % invalid_params) # default query is to return all documents q = query.get('q') if not q or q == '""' or q == "''": query['q'] = "*:*" # number of results rows_to_return = min(1000, int(query.get('rows', 10))) if rows_to_return > 0: # #1683 Work around problem of last result being out of order # in SOLR 1.4 rows_to_query = rows_to_return + 1 else: rows_to_query = rows_to_return query['rows'] = rows_to_query fq = [] if 'fq' in query: fq.append(query['fq']) fq.extend(query.get('fq_list', [])) # show only results from this CKAN instance fq.append('+site_id:%s' % solr_literal(config.get('ckan.site_id'))) # filter for package status if not '+state:' in query.get('fq', ''): fq.append('+state:active') # only return things we should be able to see if permission_labels is not None: fq.append('+permission_labels:(%s)' % ' OR '.join(solr_literal(p) for p in permission_labels)) query['fq'] = fq # faceting query['facet'] = query.get('facet', 'true') query['facet.limit'] = query.get( 'facet.limit', config.get('search.facets.limit', '50')) query['facet.mincount'] = query.get('facet.mincount', 1) # return the package ID and search scores query['fl'] = query.get('fl', 'name') # return results as json encoded string query['wt'] = query.get('wt', 'json') # If the query has a colon in it then consider it a fielded search and do use dismax. defType = query.get('defType', 'dismax') if ':' not in query['q'] or defType == 'edismax': query['defType'] = defType query['tie'] = query.get('tie', '0.1') # this minimum match is explained # http://wiki.apache.org/solr/DisMaxQParserPlugin#mm_.28Minimum_.27Should.27_Match.29 query['mm'] = query.get('mm', '2<-1 5<80%') query['qf'] = query.get('qf', QUERY_FIELDS) conn = make_connection(decode_dates=False) log.debug('Package query: %r' % query) try: solr_response = conn.search(**query) except pysolr.SolrError, e: # Error with the sort parameter. You see slightly different # error messages depending on whether the SOLR JSON comes back # or Jetty gets in the way converting it to HTML - not sure why # if e.args and isinstance(e.args[0], str): if "Can't determine a Sort Order" in e.args[0] or \ "Can't determine Sort Order" in e.args[0] or \ 'Unknown sort order' in e.args[0]: raise SearchQueryError('Invalid "sort" parameter') raise SearchError( 'SOLR returned an error running query: %r Error: %r' % (query, e))
def run(self, query): ''' Performs a dataset search using the given query. @param query - dictionary with keys like: q, fq, sort, rows, facet @return - dictionary with keys results and count May raise SearchQueryError or SearchError. ''' assert isinstance(query, (dict, MultiDict)) # check that query keys are valid if not set(query.keys()) <= VALID_SOLR_PARAMETERS: invalid_params = [ s for s in set(query.keys()) - VALID_SOLR_PARAMETERS ] raise SearchQueryError("Invalid search parameters: %s" % invalid_params) # default query is to return all documents q = query.get('q') if not q or q == '""' or q == "''": query['q'] = "*:*" # number of results rows_to_return = min(1000, int(query.get('rows', 10))) if rows_to_return > 0: # #1683 Work around problem of last result being out of order # in SOLR 1.4 rows_to_query = rows_to_return + 1 else: rows_to_query = rows_to_return query['rows'] = rows_to_query # show only results from this CKAN instance fq = query.get('fq', '') if not '+site_id:' in fq: fq += ' +site_id:"%s"' % config.get('ckan.site_id') # filter for package status if not '+state:' in fq: fq += " +state:active" query['fq'] = [fq] fq_list = query.get('fq_list', []) query['fq'].extend(fq_list) # faceting query['facet'] = query.get('facet', 'true') query['facet.limit'] = query.get( 'facet.limit', config.get('search.facets.limit', '50')) query['facet.mincount'] = query.get('facet.mincount', 1) # return the package ID and search scores query['fl'] = query.get('fl', 'name') # return results as json encoded string query['wt'] = query.get('wt', 'json') # If the query has a colon in it then consider it a fielded search and do use dismax. defType = query.get('defType', 'dismax') if ':' not in query['q'] or defType == 'edismax': query['defType'] = defType query['tie'] = query.get('tie', '0.1') # this minimum match is explained # http://wiki.apache.org/solr/DisMaxQParserPlugin#mm_.28Minimum_.27Should.27_Match.29 query['mm'] = query.get('mm', '2<-1 5<80%') query['qf'] = query.get('qf', QUERY_FIELDS) conn = make_connection(decode_dates=False) log.debug('Package query: %r' % query) try: solr_response = conn.search(**query) except pysolr.SolrError, e: # Error with the sort parameter. You see slightly different # error messages depending on whether the SOLR JSON comes back # or Jetty gets in the way converting it to HTML - not sure why # if e.args and isinstance(e.args[0], str): if "Can't determine a Sort Order" in e.args[0] or \ "Can't determine Sort Order" in e.args[0] or \ 'Unknown sort order' in e.args[0]: raise SearchQueryError('Invalid "sort" parameter') raise SearchError( 'SOLR returned an error running query: %r Error: %r' % (query, e))