runtime opti, changed feature_seg_list in dictionnary to check if seg present in list, to compute child features paths and to compute cov/id; much faster on human data, not so much on rice and ecoli data

fix bug when there are several haplotypes and option -ht is used; didnt fix bug when there are several haplotypes but option -ht is absent, todo add a warning

improved how paths and walks are detected and how bed files are attributed to genomes; fixes bug when a genome name was included in another genome name

bug fix when a segment is present in both orientation in the source features; modified search_segment to find the orientation that corresponds to the current feature

continued code refactoring. removed global variable segments_on_target_genome, sorted functions in several files

added option to not output duplicated features in annotation, only the best occurence is kept. corrected the alignment option, it now doesnt print features that didnt pass the filters.

load intersect is now much quicker. however there might be unexpected behaviour in detect_gene_copies (or maybe elsewhere) if a segment is present several times in a feature (duplication within the feature). this case was not encountered in rice and human data, but it is technically possible.

changed how Segments are stored while loading the intersect. now can handle duplicate segments in the source genome (it used to cause negative positions in var output)