Manage annotation with joined coordinates

[JeanMainguy]

PPanGGOLiN encountered issues when handling joined coordinates present in input annotation files (GFF or GBFF). Such annotations were disregarded when encountered in GBFF files and improperly managed in GFF files.

This PR solves this issue by managing properly these annotations.

Implemented solution

• Joined annotation cases have been handled when parsing the input files to retrieve different start and stop coordinates.
• The information of the different coordinates is stored in an attribute coordinates of gene objects. This attribute is a list of start and stop tuples.
• To keep the information of the joined coordinates, a new table has been added in the hdf5 call joinedCoordinates and stores the multiple start and stop of joined genes.
• Some pytest functions have been added to check the correctness of the implementation.
• Ensure proper handling of circular RGPs, addressing issues observed in the spot plot (refer to issue RGP drawn in spot figure is incorrect  #124) and ProkSee output.

[jpjarnoux]

Great work on code. I just would like to do some improvement in docstring and variable name

[jpjarnoux]

Test on few pangenomes:

species	D.pigrum		P.aeruginosa		E.Coli
	Before	After	Before	After	Before	After
Genes	59999	59999	4 933 371	4 939 509	14 932 814	15 017 229
Genomes	32	32	802	802	3190	3190
Families	4123	4123	32 196	32 279	45 002	45 077
Edges	5742	5742	59 301	60 524	131 142	139 257
Persistent	1436	1436	5 181	5 028	3 137	3 135
Shell	477	477	5 982	6 286	7 250	7 234
Cloud	2210	2210	21 033	20 965	34 676	34 708
RGP	806	806	33 894	34 148	236 867	236 824
Spots	97	97	915	829	2 037	1 721
Modules	122	122	1 420	1422	2 219	2 219

The difference in RGP count comes from a difference in clustering, which affects partitioning.

[jpjarnoux]

The following image shows the difference in RGP on the GCF_016904235.1_ASM1690423v1_CDS_2283 genome with proksee.

There is a deletion of RGP66 due to the change in partition of the ECK0501 gene family from shell to persistent.

RGPs 12 and 19 merge into RGP9 due to the change in partition of the ECK1494 gene family from persistent to shell.