Attempts to update #150 after my other merges

charcoal/compare_taxonomy.py

@@ -120,7 +120,6 @@ def choose_genome_lineage(guessed_genome_lineage, provided_lineage, match_rank,
 
 def get_genome_taxonomy(genome_name, genome_gather_json_filename, provided_lineage,
                         tax_assign, match_rank, min_f_ident, min_f_major):
-
     guessed_genome_lineage, f_major, f_ident = "", 0.0, 0.0
     # did we get gather results?
     genome_info = utils.load_contigs_gather_json(genome_gather_json_filename)

charcoal/gather_taxonomy.py

@@ -44,7 +44,7 @@ def main(args):
     new_siglist = []
     for ss in siglist:
         if genome_sig.similarity(ss) == 1.0:
-            print(f'removing an identical match: {ss.name()}')
+            print(f'removing an identical match: {ss.name}')
         else:
             new_siglist.append(ss)
     siglist = new_siglist
@@ -110,7 +110,7 @@ def main(args):
     # genome search
     genome_len=0
     entire_mh = genome_sig.minhash
-    genome_name = os.path.basename(genome_sig.name())
+    genome_name = os.path.basename(genome_sig.name)
     num_hashes = len(entire_mh.hashes)
     if not genome_len:
         for record in screed_iter:

charcoal/just_taxonomy.py

@@ -386,10 +386,10 @@ def report(*args):
             new_siglist.append(ss)
         else:
             if args.lineage and args.lineage != 'NA':
-                report(f'found exact match: {ss.name()}. removing.')
+                report(f'found exact match: {ss.name}. removing.')
                 identical_match_removed = True
             else:
-                report(f'found exact match: {ss.name()}. but no provided lineage! exiting.')
+                report(f'found exact match: {ss.name}. but no provided lineage! exiting.')
                 comment = "Exact match in matches, but no provided lineage."
                 create_empty_output(genomebase, comment, args.summary,
                                     None, args.contig_report,

charcoal/utils.py

@@ -180,7 +180,7 @@ def summarize_at_rank(lincounts, rank):
 
 def get_ident(sig):
     "Hack and slash identifiers."
-    ident = sig.name()
+    ident = sig.name
     ident = ident.split()[0]
     ident = ident.split('.')[0]
     return ident

tests/test-data/loomba/contam.json

@@ -1 +1 @@
-[[[["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"]], [["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes"]], 14], [[["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Oscillospirales"]], [["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Lachnospirales"]], 12], [[["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Oscillospirales"], ["family", "f__Acutalibacteraceae"]], [["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Oscillospirales"], ["family", "f__Ruminococcaceae"]], 24], [[["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Oscillospirales"], ["family", "f__Acutalibacteraceae"]], [["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Oscillospirales"], ["family", "f__Oscillospiraceae"]], 17], [[["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Oscillospirales"], ["family", "f__Acutalibacteraceae"], ["genus", "g__Anaeromassilibacillus"]], [["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Oscillospirales"], ["family", "f__Acutalibacteraceae"], ["genus", "g__Anaeromassilibacillus"]], 1503]]
+{"loomba": [[[["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Oscillospirales"], ["family", "f__Acutalibacteraceae"], ["genus", "g__Anaeromassilibacillus"]], [["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Lachnospirales"], ["family", "f__Anaerotignaceae"], ["genus", "g__Anaerotignum"]], 7], [[["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Oscillospirales"], ["family", "f__Acutalibacteraceae"], ["genus", "g__Anaeromassilibacillus"]], [["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Oscillospirales"], ["family", "f__Acutalibacteraceae"], ["genus", "g__An200"]], 3], [[["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Oscillospirales"], ["family", "f__Acutalibacteraceae"], ["genus", "g__Anaeromassilibacillus"]], [["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Oscillospirales"], ["family", "f__Oscillospiraceae"], ["genus", "g__Flavonifractor"]], 1], [[["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Oscillospirales"], ["family", "f__Acutalibacteraceae"], ["genus", "g__Anaeromassilibacillus"]], [["superkingdom", "d__Bacteria"], ["phylum", "p__Firmicutes_A"], ["class", "c__Clostridia"], ["order", "o__Oscillospirales"], ["family", "f__Oscillospiraceae"], ["genus", "g__Flavonifractor"]], 3]]}

tests/test_decontam.py

@@ -70,7 +70,7 @@ def make_lca_and_lineages(match_files, lineages_csv, scaled, ksize,
 
     # build database of matches & lineages!
     for ss in siglist:
-        print(ss.name(), ss.minhash.scaled)
+        print(ss.name, ss.minhash.scaled)
         ident = just_taxonomy.get_ident(ss)
         lineage = tax_assign[ident]