Force `GoldClusterizer` to cluster every sample in 1 unique single cluster

**Context & Problem** In the output of a `GoldClusterizer` (from `cluster_in_dataset` or `cluster_in_table`), every vectors are assigned to 1 cluster. Depending on the situation, multiple vectors might come from the same sample and asisgned to different clusters. For some situation, we might want to assign all the vectors of the same sample in the same cluster. **Proposed Solution** In `GoldClusterizer`, we add an new boolean attribute `force_same_cluster`. When it is True, we update the cluster of all the row of the same sample (using `idx`) instead of just the unique. `idx_vector` row. To do that we need to extract sample idx associated to the vector idx.