B'Andrea has come up with what he believes to be a novel sorting algorithm for arrays of length n . The algorithm works as follows. Initially there is an array of n integers a_1, , a_2, , ... , , a_n . Then, k steps are executed. For each 1 <= i <= k , during the i -th step the subsequence of the array a with indexes j_{i,1}< j_{i,2}< ... < j_{i, q_i} is sorted, without changing the values with the remaining indexes. So, the subsequence a_{j_{i,1}}, , a_{j_{i,2}}, , ... , , a_{j_{i,q_i}} is sorted and all other elements of a are left untouched. Andrea, being eager to share his discovery with the academic community, sent a short paper describing his algorithm to the journal "Annals of Sorting Algorithms" and you are the referee of the paper (that is, the person who must judge the correctness of the paper). You must decide whether Andrea 's algorithm is correct, that is, if it sorts any array a of n integers. The first line contains two integers n and k ( 1 <= n <= 40 , 0 <= k <= 10 ) -- the length of the arrays handled by Andrea 's algorithm and the number of steps of Andrea 's algorithm. Then k lines follow, each describing the subsequence considered in a step of Andrea 's algorithm. The i -th of these lines contains the integer q_i ( 1 <= q_i <= n ) followed by q_i integers j_{i,1}, ,j_{i,2}, , ... , , j_{i,q_i} ( 1 <= j_{i,1}<j_{i,2}< cdots<j_{i,q_i} <= n ) -- the length of the subsequence considered in the i -th step and the indexes of the subsequence. If Andrea 's algorithm is correct print ACCEPTED, otherwise print REJECTED. Explanation of the first sample: The algorithm consists of 3 steps. The first one sorts the subsequence [a_1, a_2, a_3] , the second one sorts the subsequence [a_2, a_3, a_4] , the third one sorts the subsequence [a_1,a_2] . For example, if initially a=[6, 5, 6, 3] , the algorit'...