adapted the code to do a transfer on several genomes, by clearing the features target genome paths between each genome

adapted the multicopies code to the output of -var. corrected the positions in the output gff for the inversions

added gene multicopies handling. new function detect_copies, and now the target_genome_path of child features and the transfer depend on the parent feature

changed how the features are transfered. now after the transfer of a gene, all its child features are transfered with it. this automatically sorts the output gff, and makes easier the handling of the cov and id filters (if a gene didnt pass, dont transfer its chids, etc)

now always use the function get_seg_occ to get the right segment occurence. so now the option -ann outputs the target genome gff with all the copies of the genes if -cov and -id are at 0

added a function to get the right segment occurence, given the segment_id,walk_name,feature_id,copy_id

modified how the feature paths in the target genomes are found; now several can be found, on several walks (chr/ctg); adapted how the features are transfered (only for the genome annotation transfer, not for the graph_gaf/gff, variations and alignment)

the dict segments_on_target_genome now stores the segment twice if it is present twice in a given walk. adapted the rest of the code to use the last occurence (like before, will change that later)

added a function to order the dictionnary that stores the features. now the parent features are before their childs; it makes future manipulation of this dictionnary easier

modified the condition for the transfer. now the filters coverage and id only apply to genes. if the gene passes the filters, all its child features are transfered. if the gene doesnt pass the filters, none of its child features are transfered

inverted the start and stop positions in the output gff when the stop was before the start (in inverted regions)

changed the usage of id and cov; -ann is now required, and these options only work on -ann and -pav (the rest will come later)

the function that generates the target genomes gff now doesnt transfer if the coverage or id is too low

added a dict that stores the size of the segments, to get the coverage and sequence id of the transfers

changed the creation of Feature objects to handle intersect files where a feature can appear after its childs, to handle complete gff files

corrected the emptying of the dict segments_on_target_genome; this reduced the number of fragmented genes, that were due to segments from previous genome staying in the dict