Esempio n. 1
0
def rankinfo_main(args):
    """
    rankinfo!
    """
    p = argparse.ArgumentParser(prog="sourmash lca rankinfo")
    p.add_argument('db', nargs='+')
    p.add_argument('--scaled', type=float)
    p.add_argument('-d', '--debug', action='store_true')
    args = p.parse_args(args)

    if not args.db:
        error('Error! must specify at least one LCA database with --db')
        sys.exit(-1)

    if args.debug:
        set_debug(args.debug)

    if args.scaled:
        args.scaled = int(args.scaled)

    # load all the databases
    dblist, ksize, scaled = lca_utils.load_databases(args.db, args.scaled)

    # count all the LCAs across these databases
    counts = make_lca_counts(dblist)

    # collect counts across all ranks
    counts_by_rank = defaultdict(int)
    for lineage, count in counts.items():
        if lineage:
            lineage_tup = lineage[-1]
            counts_by_rank[lineage_tup.rank] += count

    # output!
    total = float(sum(counts_by_rank.values()))
    for rank in lca_utils.taxlist():
        count = counts_by_rank.get(rank, 0)
        print('{}: {} ({:.1f}%)'.format(rank, count, count / total * 100.))
def gather_main(args):
    """
    """
    p = argparse.ArgumentParser()
    p.add_argument('--debug', action='store_true')
    p.add_argument('spreadsheet')
    p.add_argument('species')
    p.add_argument('--sbt')
    p.add_argument('-o', '--output', type=argparse.FileType('wt'))
    args = p.parse_args(args)

    if args.debug:
        set_debug(args.debug)

    assignments, num_rows = load_taxonomy_assignments(args.spreadsheet)

    found = False
    for ident, lineage in assignments.items():
        for vv in lineage:
            if vv.rank == 'species' and vv.name == args.species:
                found = True
                found_lineage = lineage
                break

    if not found:
        print('nothing found for {}; quitting'.format(args.species))
        sys.exit(-1)

    print('found:', ", ".join(lca_utils.zip_lineage(found_lineage)))

    lineage_search = dict(found_lineage)

    rank_found = defaultdict(list)
    rank_idents = defaultdict(list)
    taxlist = list(reversed(list(lca_utils.taxlist())))

    for ident, lineage in assignments.items():
        dd = dict(lineage)
        for k in taxlist:
            if dd.get(k) and dd.get(k) == lineage_search.get(k):
                rank_found[k].append(lineage)
                rank_idents[k].append(ident)
                break

    retrieve_idents = defaultdict(set)
    gimme_idents = {}
    for k in rank_found:
        print('at', k, 'found', len(rank_found.get(k)))

        num_to_extract = min(len(rank_idents[k]), 10)
        gimme = random.sample(rank_idents[k], num_to_extract)
        for g in gimme:
            gimme_idents[g] = k

    if not args.output or not args.sbt:
        print('no output arg or SBT arg given; quitting without extracting')
        sys.exit(-1)

    print('looking for:', len(gimme_idents))

    tree = sourmash_lib.load_sbt_index(args.sbt)

    w = csv.writer(args.output)
    for n, leaf in enumerate(tree.leaves()):
        if n % 1000 == 0:
            print('...', n)
        name = leaf.data.name()
        # hack for NCBI-style names, etc.
        name = name.split(' ')[0].split('.')[0]

        if name in gimme_idents:
            level = gimme_idents[name]
            level_n = taxlist.index(level)
            filename = leaf.data.d['filename']

            w.writerow([level, level_n, filename, leaf.data.name()])
            print('FOUND!', leaf.data.name(), level)
Esempio n. 3
0
def gather_main(args):
    """
    Do a greedy search for the hash components of a query against an LCA db.

    Here we don't actually do a least-common-ancestor search of any kind; we
    do essentially the same kind of search as we do in `sourmash gather`, with
    the main difference that we are implicitly combining different genomes of
    identical lineages.

    This takes advantage of the structure of the LCA db, where we store the
    full lineage information for each known hash, as opposed to storing only
    the least-common-ancestor information for it.
    """
    p = argparse.ArgumentParser(prog="sourmash lca gather")
    p.add_argument('query')
    p.add_argument('db', nargs='+')
    p.add_argument('-d', '--debug', action='store_true')
    p.add_argument('-o', '--output', type=argparse.FileType('wt'),
                   help='output CSV containing matches to this file')
    p.add_argument('--output-unassigned', type=argparse.FileType('wt'),
                   help='output unassigned portions of the query as a signature to this file')
    p.add_argument('--ignore-abundance',  action='store_true',
                   help='do NOT use k-mer abundances if present')
    args = p.parse_args(args)

    if args.debug:
        set_debug(args.debug)

    # load all the databases
    dblist, ksize, scaled = lca_utils.load_databases(args.db, None)

    # for each query, gather all the matches across databases
    query_sig = sourmash_args.load_query_signature(args.query, ksize, 'DNA')
    debug('classifying', query_sig.name())

    # make sure we're looking at the same scaled value as database
    query_sig.minhash = query_sig.minhash.downsample_scaled(scaled)

    # do the classification, output results
    found = []
    for result, f_unassigned, est_bp, remaining_mins in gather_signature(query_sig, dblist, args.ignore_abundance):
        # is this our first time through the loop? print headers, if so.
        if not len(found):
            print_results("")
            print_results("overlap     p_query p_match ")
            print_results("---------   ------- --------")

        # output!
        pct_query = '{:.1f}%'.format(result.f_unique_to_query*100)
        pct_match = '{:.1f}%'.format(result.f_match*100)
        str_bp = format_bp(result.intersect_bp)
        name = format_lineage(result.lineage)

        equal_match_str = ""
        if result.n_equal_matches:
            equal_match_str = " (** {} equal matches)".format(result.n_equal_matches)

        print_results('{:9}   {:>6}  {:>6}      {}{}', str_bp, pct_query,
                      pct_match, name, equal_match_str)

        found.append(result)

    if found:
        print_results('')
        if f_unassigned:
            print_results('{:.1f}% ({}) of hashes have no assignment.', f_unassigned*100,
                          format_bp(est_bp))
        else:
            print_results('Query is completely assigned.')
            print_results('')
    # nothing found.
    else:
        est_bp = len(query_sig.minhash.get_mins()) * query_sig.minhash.scaled
        print_results('')
        print_results('No assignment for est {} of sequence.',
                      format_bp(est_bp))
        print_results('')

    if not found:
        sys.exit(0)

    if args.output:
        fieldnames = ['intersect_bp', 'f_match', 'f_unique_to_query', 'f_unique_weighted',
                      'average_abund', 'name', 'n_equal_matches'] + list(lca_utils.taxlist())

        w = csv.DictWriter(args.output, fieldnames=fieldnames)
        w.writeheader()
        for result in found:
            lineage = result.lineage
            d = dict(result._asdict())
            del d['lineage']

            for (rank, value) in lineage:
                d[rank] = value

            w.writerow(d)

    if args.output_unassigned:
        if not found:
            notify('nothing found - entire query signature unassigned.')
        elif not remaining_mins:
            notify('no unassigned hashes! not saving.')
        else:
            outname = args.output_unassigned.name
            notify('saving unassigned hashes to "{}"', outname)

            e = query_sig.minhash.copy_and_clear()
            e.add_many(remaining_mins)

            sourmash_lib.save_signatures([ sourmash_lib.SourmashSignature(e) ],
                                         args.output_unassigned)
Esempio n. 4
0
def classify(args):
    """
    main single-genome classification function.
    """
    p = argparse.ArgumentParser()
    p.add_argument('--db', nargs='+', action='append')
    p.add_argument('--query', nargs='+', action='append')
    p.add_argument('--threshold', type=int, default=DEFAULT_THRESHOLD)
    p.add_argument('-o',
                   '--output',
                   type=argparse.FileType('wt'),
                   help='output CSV to this file instead of stdout')
    p.add_argument('--scaled', type=float)
    p.add_argument('--traverse-directory',
                   action='store_true',
                   help='load all signatures underneath directories.')
    p.add_argument('-d', '--debug', action='store_true')
    args = p.parse_args(args)

    if not args.db:
        error('Error! must specify at least one LCA database with --db')
        sys.exit(-1)

    if not args.query:
        error('Error! must specify at least one query signature with --query')
        sys.exit(-1)

    if args.debug:
        set_debug(args.debug)

    # flatten --db and --query
    args.db = [item for sublist in args.db for item in sublist]
    args.query = [item for sublist in args.query for item in sublist]

    # load all the databases
    dblist, ksize, scaled = lca_utils.load_databases(args.db, args.scaled)
    notify('ksize={} scaled={}', ksize, scaled)

    # find all the queries
    notify('finding query signatures...')
    if args.traverse_directory:
        inp_files = list(sourmash_args.traverse_find_sigs(args.query))
    else:
        inp_files = list(args.query)

    # set up output
    csvfp = csv.writer(sys.stdout)
    if args.output:
        notify("outputting classifications to '{}'", args.output.name)
        csvfp = csv.writer(args.output)
    else:
        notify("outputting classifications to stdout")
    csvfp.writerow(['ID', 'status'] + list(lca_utils.taxlist()))

    # for each query, gather all the matches across databases
    total_count = 0
    n = 0
    total_n = len(inp_files)
    for query_filename in inp_files:
        n += 1
        for query_sig in sourmash_lib.load_signatures(query_filename,
                                                      ksize=ksize):
            notify(u'\r\033[K', end=u'')
            notify('... classifying {} (file {} of {})',
                   query_sig.name(),
                   n,
                   total_n,
                   end='\r')
            debug('classifying', query_sig.name())
            total_count += 1

            # make sure we're looking at the same scaled value as database
            query_sig.minhash = query_sig.minhash.downsample_scaled(scaled)

            # do the classification
            lineage, status = classify_signature(query_sig, dblist,
                                                 args.threshold)
            debug(lineage)

            # output each classification to the spreadsheet
            row = [query_sig.name(), status]
            row += lca_utils.zip_lineage(lineage)

            # when outputting to stdout, make output intelligible
            if not args.output:
                notify(u'\r\033[K', end=u'')
            csvfp.writerow(row)

    notify(u'\r\033[K', end=u'')
    notify('classified {} signatures total', total_count)
Esempio n. 5
0
def index(args):
    """
    main function for building an LCA database.
    """
    p = argparse.ArgumentParser()
    p.add_argument('csv', help='taxonomy spreadsheet')
    p.add_argument('lca_db_out', help='name to save database to')
    p.add_argument('signatures', nargs='+',
                   help='one or more sourmash signatures')
    p.add_argument('--scaled', default=10000, type=float)
    p.add_argument('-k', '--ksize', default=31, type=int)
    p.add_argument('-d', '--debug', action='store_true')
    p.add_argument('-C', '--start-column', default=2, type=int,
                   help='column at which taxonomic assignments start')
    p.add_argument('--tabs', action='store_true',
                   help='input spreadsheet is tab-delimited (default: commas)')
    p.add_argument('--no-headers', action='store_true',
                   help='no headers present in taxonomy spreadsheet')
    p.add_argument('--split-identifiers', action='store_true',
                   help='split names in signatures on whitspace and period')
    p.add_argument('-f', '--force', action='store_true')
    p.add_argument('--traverse-directory', action='store_true',
                   help='load all signatures underneath directories.')
    p.add_argument('--report', help='output a report on anomalies, if any.')
    args = p.parse_args(args)

    if args.start_column < 2:
        error('error, --start-column cannot be less than 2')
        sys.exit(-1)

    if args.debug:
        set_debug(args.debug)

    args.scaled = int(args.scaled)

    # first, load taxonomy spreadsheet
    delimiter = ','
    if args.tabs:
        delimiter = '\t'
    assignments, num_rows = load_taxonomy_assignments(args.csv,
                                               delimiter=delimiter,
                                               start_column=args.start_column,
                                               use_headers=not args.no_headers,
                                               force=args.force)

    # convert lineages to numbers.
    next_lineage_index = 0
    lineage_dict = {}

    assignments_idx = {}
    lineage_to_idx = {}
    for (ident, lineage) in assignments.items():
        idx = lineage_to_idx.get(lineage)
        if idx is None:
            idx = next_lineage_index
            next_lineage_index += 1

            lineage_dict[idx] = lineage
            lineage_to_idx[lineage] = idx

        assignments_idx[ident] = idx

    notify('{} distinct lineages in spreadsheet out of {} rows',
           len(lineage_dict), num_rows)

    # load signatures, construct index of hashvals to lineages
    hashval_to_lineage = defaultdict(set)
    md5_to_lineage = {}

    notify('finding signatures...')
    if args.traverse_directory:
        yield_all_files = False           # only pick up *.sig files?
        if args.force:
            yield_all_files = True
        inp_files = list(sourmash_args.traverse_find_sigs(args.signatures,
                                                          yield_all_files=yield_all_files))
    else:
        inp_files = list(args.signatures)

    n = 0
    total_n = len(inp_files)
    record_duplicates = set()
    record_no_lineage = set()
    record_remnants = set(assignments_idx.keys())
    for filename in inp_files:
        n += 1
        for sig in sourmash_lib.load_signatures(filename, ksize=args.ksize):
            notify(u'\r\033[K', end=u'')
            notify('... loading signature {} (file {} of {})', sig.name()[:30], n, total_n, end='\r')
            debug(filename, sig.name())

            if sig.md5sum() in md5_to_lineage:
                notify('\nWARNING: in file {}, duplicate md5sum: {}; skipping', filename, sig.md5sum())
                record_duplicates.add(filename)
                continue

            name = sig.name()
            if args.split_identifiers: # hack for NCBI-style names, etc.
                name = name.split(' ')[0].split('.')[0]

            # is this one for which we have a lineage assigned?
            lineage_idx = assignments_idx.get(name)
            if lineage_idx is None:
               notify('\nWARNING: no lineage assignment for {}.', name)
               record_no_lineage.add(name)
            else:
                # remove from our list of remnant lineages
                record_remnants.remove(name)

                # downsample to specified scaled; this has the side effect of
                # making sure they're all at the same scaled value!
                minhash = sig.minhash.downsample_scaled(args.scaled)

                # connect hashvals to lineage
                for hashval in minhash.get_mins():
                    hashval_to_lineage[hashval].add(lineage_idx)

                # store md5 -> lineage too
                md5_to_lineage[sig.md5sum()] = lineage_idx

    notify(u'\r\033[K', end=u'')
    notify('...found {} genomes with lineage assignments!!',
           len(md5_to_lineage))

    # remove those lineages with no genomes associated
    assigned_lineages = set(md5_to_lineage.values())
    lineage_dict_2 = {}
    for idx in assigned_lineages:
        lineage_dict_2[idx] = lineage_dict[idx]

    unused_lineages = set(lineage_dict.values()) - set(lineage_dict_2.values())

    notify('{} assigned lineages out of {} distinct lineages in spreadsheet',
           len(lineage_dict_2), len(lineage_dict))
    lineage_dict = lineage_dict_2

    # now, save!
    db_outfile = args.lca_db_out
    if not (db_outfile.endswith('.lca.json') or db_outfile.endswith('.lca.json.gz')):
        db_outfile += '.lca.json'
    notify('saving to LCA DB: {}'.format(db_outfile))

    db = lca_utils.LCA_Database()
    db.lineage_dict = lineage_dict
    db.hashval_to_lineage_id = hashval_to_lineage
    db.ksize = int(args.ksize)
    db.scaled = int(args.scaled)
    db.signatures_to_lineage = md5_to_lineage

    db.save(db_outfile)

    if record_duplicates or record_no_lineage or record_remnants or unused_lineages:
        if record_duplicates:
            notify('WARNING: {} duplicate signatures.', len(record_duplicates))
        if record_no_lineage:
            notify('WARNING: no lineage provided for {} signatures.',
                   len(record_no_lineage))
        if record_remnants:
            notify('WARNING: no signatures for {} lineage assignments.',
                   len(record_remnants))
        if unused_lineages:
            notify('WARNING: {} unused lineages.', len(unused_lineages))

        if args.report:
            notify("generating a report and saving in '{}'", args.report)
            generate_report(record_duplicates, record_no_lineage,
                            record_remnants, unused_lineages, args.report)
        else:
            notify('(You can use --report to generate a detailed report.)')
Esempio n. 6
0
def main():
    p = argparse.ArgumentParser()
    p.add_argument('sigs', nargs='+')
    p.add_argument('--traverse-directory',
                   action='store_true',
                   help='load all signatures underneath directories.')
    p.add_argument('-k', '--ksize', default=31, type=int)
    p.add_argument('-d', '--debug', action='store_true')
    p.add_argument('-f', '--force', action='store_true')
    p.add_argument('--scaled', type=float, default=10000)
    p.add_argument('--plot', default=None)
    p.add_argument('-o',
                   '--output',
                   type=argparse.FileType('wt'),
                   help='CSV output')
    p.add_argument('--step', type=int, default=1000)
    p.add_argument('--repeat', type=int, default=5)
    p.add_argument('--db', nargs='+', action='append')
    args = p.parse_args()

    if args.debug:
        set_debug(args.debug)

    args.scaled = int(args.scaled)

    dblist = []
    known_hashes = set()
    if args.db:
        args.db = [item for sublist in args.db for item in sublist]
        dblist, ksize, scaled = lca_utils.load_databases(args.db, args.scaled)
        assert ksize == args.ksize
        notify('loaded {} LCA databases', len(dblist))

        for db in dblist:
            known_hashes.update(db.hashval_to_lineage_id.keys())
        notify('got {} known hashes!', len(known_hashes))

    notify('finding signatures...')
    if args.traverse_directory:
        yield_all_files = False  # only pick up *.sig files?
        if args.force:
            yield_all_files = True
        inp_files = list(
            sourmash_args.traverse_find_sigs(args.sigs,
                                             yield_all_files=yield_all_files))
    else:
        inp_files = list(args.sigs)

    n = 0
    total_n = len(inp_files)
    sigs = []
    total_hashvals = list()
    for filename in inp_files:
        n += 1
        for sig in sourmash_lib.load_signatures(filename, ksize=args.ksize):
            notify(u'\r\033[K', end=u'')
            notify('... loading signature {} (file {} of {})',
                   sig.name()[:30],
                   n,
                   total_n,
                   end='\r')
            debug(filename, sig.name())

            sig.minhash = sig.minhash.downsample_scaled(args.scaled)

            total_hashvals.extend(sig.minhash.get_mins())
            sigs.append(sig)

    notify(u'\r\033[K', end=u'')
    notify('...found {} signatures total in {} files.', len(sigs), total_n)

    distinct_hashvals = set(total_hashvals)
    notify('{} distinct out of {} total hashvals.', (len(distinct_hashvals)),
           len(total_hashvals))
    if known_hashes:
        n_known = len(known_hashes.intersection(distinct_hashvals))
        notify('{} of them known, or {:.1f}%', n_known,
               n_known / float(len(distinct_hashvals)) * 100)

    x = []
    y = []
    z = []
    notify('subsampling...')
    for n in range(0, len(total_hashvals), args.step):
        notify(u'\r\033[K', end=u'')
        notify('... {} of {}', n, len(total_hashvals), end='\r')
        avg = 0
        known = 0
        for j in range(0, args.repeat):
            subsample = random.sample(total_hashvals, n)
            distinct = len(set(subsample))
            if known_hashes:
                known += len(set(subsample).intersection(known_hashes))
            avg += distinct

        x.append(n)
        y.append(avg / args.repeat)
        z.append(known / args.repeat)

    notify('\n...done!')

    if args.output:
        w = csv.writer(args.output)
        w.writerow(['n', 'k', 'known'])
        for a, b, c in zip(x, y, z):
            w.writerow([a, b, c])

    if args.plot:
        from matplotlib import pyplot
        pyplot.plot(x, y)
        pyplot.savefig(args.plot)
Esempio n. 7
0
def summarize_main(args):
    """
    main summarization function.
    """
    p = argparse.ArgumentParser()
    p.add_argument('--db', nargs='+', action='append')
    p.add_argument('--query', nargs='+', action='append')
    p.add_argument('--threshold', type=int, default=DEFAULT_THRESHOLD)
    p.add_argument('--traverse-directory',
                   action='store_true',
                   help='load all signatures underneath directories.')
    p.add_argument('-o',
                   '--output',
                   type=argparse.FileType('wt'),
                   help='CSV output')
    p.add_argument('--scaled', type=float)
    p.add_argument('-d', '--debug', action='store_true')
    args = p.parse_args(args)

    if not args.db:
        error('Error! must specify at least one LCA database with --db')
        sys.exit(-1)

    if not args.query:
        error('Error! must specify at least one query signature with --query')
        sys.exit(-1)

    if args.debug:
        set_debug(args.debug)

    if args.scaled:
        args.scaled = int(args.scaled)

    # flatten --db and --query
    args.db = [item for sublist in args.db for item in sublist]
    args.query = [item for sublist in args.query for item in sublist]

    # load all the databases
    dblist, ksize, scaled = lca_utils.load_databases(args.db, args.scaled)
    notify('ksize={} scaled={}', ksize, scaled)

    # find all the queries
    notify('finding query signatures...')
    if args.traverse_directory:
        inp_files = list(sourmash_args.traverse_find_sigs(args.query))
    else:
        inp_files = list(args.query)

    # for each query, gather all the hashvals across databases
    total_count = 0
    n = 0
    total_n = len(inp_files)
    hashvals = defaultdict(int)
    for query_filename in inp_files:
        n += 1
        for query_sig in sourmash_lib.load_signatures(query_filename,
                                                      ksize=ksize):
            notify(u'\r\033[K', end=u'')
            notify('... loading {} (file {} of {})',
                   query_sig.name(),
                   n,
                   total_n,
                   end='\r')
            total_count += 1

            mh = query_sig.minhash.downsample_scaled(scaled)
            for hashval in mh.get_mins():
                hashvals[hashval] += 1

    notify(u'\r\033[K', end=u'')
    notify('loaded {} signatures from {} files total.', total_count, n)

    # get the full counted list of lineage counts in this signature
    lineage_counts = summarize(hashvals, dblist, args.threshold)

    # output!
    total = float(len(hashvals))
    for (lineage, count) in lineage_counts.items():
        if lineage:
            lineage = lca_utils.zip_lineage(lineage, truncate_empty=True)
            lineage = ';'.join(lineage)
        else:
            lineage = '(root)'

        p = count / total * 100.
        p = '{:.1f}%'.format(p)

        print_results('{:5} {:>5}   {}'.format(p, count, lineage))

    # CSV:
    if args.output:
        w = csv.writer(args.output)
        headers = ['count'] + list(lca_utils.taxlist())
        w.writerow(headers)

        for (lineage, count) in lineage_counts.items():
            debug('lineage:', lineage)
            row = [count] + lca_utils.zip_lineage(lineage)
            w.writerow(row)
Esempio n. 8
0
def compare_csv(args):
    p = argparse.ArgumentParser()
    p.add_argument('csv1', help='taxonomy spreadsheet output by classify')
    p.add_argument('csv2', help='custom taxonomy spreadsheet')
    p.add_argument('-d', '--debug', action='store_true')
    p.add_argument('-C',
                   '--start-column',
                   default=2,
                   type=int,
                   help='column at which taxonomic assignments start')
    p.add_argument('--tabs',
                   action='store_true',
                   help='input spreadsheet is tab-delimited (default: commas)')
    p.add_argument('--no-headers',
                   action='store_true',
                   help='no headers present in taxonomy spreadsheet')
    p.add_argument('-f', '--force', action='store_true')
    args = p.parse_args(args)

    if args.start_column < 2:
        error('error, --start-column cannot be less than 2')
        sys.exit(-1)

    if args.debug:
        set_debug(args.debug)

    # first, load classify-style spreadsheet
    notify('loading classify output from: {}', args.csv1)
    assignments0, num_rows0 = load_taxonomy_assignments(args.csv1,
                                                        start_column=3)

    notify('loaded {} distinct lineages, {} rows',
           len(set(assignments0.values())), num_rows0)
    notify('----')

    # next, load custom taxonomy spreadsheet
    delimiter = ','
    if args.tabs:
        delimiter = '\t'

    notify('loading custom spreadsheet from: {}', args.csv2)
    assignments, num_rows = load_taxonomy_assignments(
        args.csv2,
        delimiter=delimiter,
        start_column=args.start_column,
        use_headers=not args.no_headers,
        force=args.force)
    notify('loaded {} distinct lineages, {} rows',
           len(set(assignments.values())), num_rows)

    # now, compute basic differences:
    missing_1 = set(assignments0.keys()) - set(assignments.keys())
    missing_2 = set(assignments.keys()) - set(assignments0.keys())
    if missing_2:
        notify('missing {} assignments in classify spreadsheet.',
               len(missing_2))
    if missing_1:
        notify('missing {} assignments in custom spreadsheet.', len(missing_1))
    if missing_1 or missing_2:
        notify('(these will not be evaluated any further)')
    else:
        notify('note: all IDs are in both spreadsheets!')

    # next, look at differences in lineages
    common = set(assignments0.keys())
    common.intersection_update(assignments.keys())

    n_total = 0
    n_different = 0
    n_compat = 0
    n_incompat = 0
    incompat_rank = defaultdict(int)
    for k in common:
        n_total += 1
        v0 = assignments0[k]
        v1 = assignments[k]
        if v0 != v1:
            n_different += 1
            tree = lca_utils.build_tree([v0])
            lca_utils.build_tree([v1], tree)

            lca, reason = lca_utils.find_lca(tree)
            if reason == 0:  # compatible lineages
                n_compat += 1
                print_results("{},compatible,{}", k,
                              ";".join(zip_lineage(lca)))
            else:
                n_incompat += 1
                print_results("{},incompatible,{}", k,
                              ";".join(zip_lineage(lca)))
                rank = next(iter(lca_utils.taxlist()))
                if lca:
                    rank = lca[-1].rank
                incompat_rank[rank] += 1

    notify("{} total assignments, {} differ between spreadsheets.", n_total,
           n_different)
    notify("{} are compatible (one lineage is ancestor of another.", n_compat)
    notify("{} are incompatible (there is a disagreement in the trees).",
           n_incompat)

    if n_incompat:
        for rank in lca_utils.taxlist():
            notify('{} incompatible at rank {}', incompat_rank[rank], rank)