I think a better way to patch the contigs and perform genome finishing is, one should go for some long-read sequencing (like pacBio OR nanopore) and build the assembly using a long read, and correct it using short-read sequencing (just performing a hybrid assembly). You can also use the SPAdes scaffolds.fa file instead of contig.fa, it may have patched contigs. There are few tools available though, which serve the purpose like CONTIGuator (A genome finishing tool for bacterial genome; I am not sure whether it would work for eukaryotic genomes or not; I hope you would find other for eukaryotic genome finishing).
For the gene prediction of the eukaryotic genome, there are a number of tools available, like GENSCAN, AUGUSTUS, GeneMark, etc. (If you just google it, you would get plenty number of hits).
I was using AUGUSTUS to annotate the Fungal genome, and it was quite accurate, but I am not sure about annotating the chloroplast genome though. You can give it a try and post your view for us :)