Add --chinese (CJK) figure-label rendering, plus related fixes

src/jcvi/apps/base.py

@@ -497,6 +497,7 @@ def set_image_options(
             GRAPHIC_FORMATS,
             ImageOptions,
             is_tex_available,
+            set_chinese_font,
             setup_theme,
         )
 
@@ -552,6 +553,12 @@ def set_image_options(
         group.add_argument(
             "--notex", default=False, action="store_true", help="Do not use tex"
         )
+        group.add_argument(
+            "--chinese",
+            default=False,
+            action="store_true",
+            help="Render Chinese labels: implies --notex and selects a CJK font",
+        )
         # https://github.com/tanghaibao/jcvi/issues/515#issuecomment-1327305211
         if (
             "--seed" not in self._option_string_actions
@@ -572,6 +579,11 @@ def set_image_options(
         assert opts.dpi > 0
         assert "x" in opts.figsize
 
+        # Chinese labels cannot be rendered through latex, so --chinese implies
+        # --notex and additionally selects a CJK-capable font below.
+        if opts.chinese:
+            opts.notex = True
+
         iopts = ImageOptions(opts)
 
         if opts.notex:
@@ -584,6 +596,9 @@ def set_image_options(
 
         setup_theme(style=opts.style, font=opts.font, usetex=iopts.usetex)
 
+        if opts.chinese:
+            set_chinese_font()
+
         return opts, args, iopts
 
     def set_dotplot_opts(self, theme: int = 2):

src/jcvi/compara/ks.py

@@ -15,7 +15,11 @@
 from typing import Optional
 
 from Bio import AlignIO, SeqIO
-from Bio.Align.Applications import ClustalwCommandline, MuscleCommandline
+
+try:
+    from Bio.Align.Applications import ClustalwCommandline, MuscleCommandline
+except ImportError:  # Biopython >= 1.78 removed the Applications command wrappers
+    ClustalwCommandline = MuscleCommandline = None
 import numpy as np
 
 from ..apps.base import (
@@ -49,7 +53,11 @@ def run(self):
 class YnCommandline(AbstractCommandline):
     """Little commandline for yn00."""
 
-    def __init__(self, ctl_file, command=PAML_BIN("yn00")):
+    def __init__(self, ctl_file, command=None):
+        # Resolve PAML path lazily so importing this module does not prompt for
+        # a yn00 binary that is only needed by the calc action.
+        if command is None:
+            command = PAML_BIN("yn00")
         self.ctl_file = ctl_file
         self.parameters = []
         self.command = command
@@ -67,8 +75,12 @@ def __init__(
         nuc_file,
         output_file,
         outfmt="paml",
-        command=PAL2NAL_BIN("pal2nal.pl"),
+        command=None,
     ):
+        # Resolve PAL2NAL lazily so importing this module does not prompt for a
+        # binary only needed by the calc action.
+        if command is None:
+            command = PAL2NAL_BIN("pal2nal.pl")
         self.prot_align_file = prot_align_file
         self.nuc_file = nuc_file
         self.output_file = output_file

src/jcvi/graphics/align.py

@@ -297,8 +297,8 @@ def delete(self, a, b):
 
     def duplicate(self, a, b):
         ai, bi = self.get_endpoints(a, b)
-        ai += self.wiggle / 2
-        bi += self.wiggle / 2
+        ai += self.wiggle // 2
+        bi += self.wiggle // 2
         ci, di = ai - self.wiggle, ai
         bb = self.sizes[ai:bi]
         bs = len(bb)
@@ -479,6 +479,12 @@ def main():
     (mode,) = args
     assert mode == "demo"
 
+    row_headers = ("Inversion", "Indel", "Duplication")
+    col_headers = ("Assembly alignment", "Read alignment", "Optical map alignment")
+    if opts.chinese:
+        row_headers = ("倒位", "插入缺失", "重复")
+        col_headers = ("拼接比对", "读段比对", "光学图谱比对")
+
     a, b = 30, 70
     pad = 0.08
     w = 0.31
@@ -494,14 +500,14 @@ def main():
     # Row headers
     xx = pad * 0.6
     yy = 1 - pad - 0.5 * w
-    for title in ("Inversion", "Indel", "Duplication"):
+    for title in row_headers:
         root.text(xx, yy, title, ha="center", va="center")
         yy -= w
 
     # Column headers
     xx = pad + 0.5 * w
     yy = 1 - pad / 2
-    for title in ("Assembly alignment", "Read alignment", "Optical map alignment"):
+    for title in col_headers:
         root.text(xx, yy, title, ha="center", va="center")
         xx += w
 

src/jcvi/graphics/base.py

@@ -417,6 +417,22 @@ def markup(s: str):
     Change the string to latex format, and italicize the text between *.
     """
     if not rcParams["text.usetex"]:
+        # Without LaTeX, strip LaTeX-only markup so labels stay readable (and so
+        # CJK labels are not hidden behind \textit{}); keep $...$ mathtext intact.
+        s = s.replace(r"\noindent", "")
+
+        def _italic(m):
+            # Render \textit{...} as mathtext italic so Latin names stay italic
+            # without latex; fall back to plain for non-ASCII (e.g. CJK).
+            x = m.group(1)
+            if x.isascii():
+                return r"$\mathit{" + x.replace(" ", r"\ ") + r"}$"
+            return x
+
+        s = re.sub(r"\\textit\{([^{}]*)\}", _italic, s)
+        s = re.sub(r"\\text(?:bf|rm|sf)\{([^{}]*)\}", r"\1", s)
+        s = re.sub(r"\*([^*]+)\*", r"\1", s)  # *italic* markup -> plain text
+        s = s.replace(r"\\", "\n")  # LaTeX line break -> real newline
         return s
     if "$" in s:
         return s
@@ -433,6 +449,31 @@ def append_percentage(s):
         return s + "%"
 
 
+# Cross-platform CJK-capable sans-serif fonts, tried in order (macOS / Windows /
+# Linux). Used as a per-glyph fallback so Chinese labels render with --chinese.
+CJK_FONTS = [
+    "Arial Unicode MS",
+    "PingFang SC",
+    "Hiragino Sans GB",
+    "Heiti SC",
+    "Songti SC",
+    "Microsoft YaHei",
+    "SimHei",
+    "SimSun",
+    "Noto Sans CJK SC",
+    "Noto Sans SC",
+    "Source Han Sans SC",
+    "WenQuanYi Zen Hei",
+]
+
+
+def set_chinese_font():
+    """Select a CJK-capable font and disable latex so Chinese labels render."""
+    rc("text", usetex=False)
+    rc("font", **{"family": "sans-serif", "sans-serif": CJK_FONTS + ["DejaVu Sans"]})
+    rcParams["axes.unicode_minus"] = False
+
+
 def setup_theme(
     context="notebook",
     style="darkgrid",
@@ -457,7 +498,9 @@ def setup_theme(
         rc("text", usetex=False)
 
     if font == "Helvetica":
-        rc("font", **{"family": "sans-serif", "sans-serif": ["Helvetica"]})
+        # CJK fonts appended as a per-glyph fallback so Chinese labels render
+        # (only used when text.usetex is False; Latin still comes from Helvetica)
+        rc("font", **{"family": "sans-serif", "sans-serif": ["Helvetica"] + CJK_FONTS})
     elif font == "Palatino":
         rc("font", **{"family": "serif", "serif": ["Palatino"]})
     elif font == "Schoolbook":

src/jcvi/graphics/coverage.py

@@ -19,6 +19,7 @@
 from .base import (
     Rectangle,
     adjust_spines,
+    markup,
     mb_float_formatter,
     mb_formatter,
     plt,
@@ -195,7 +196,7 @@ def __init__(
                 xy.import_hlfile(hlfile, chr, diverge=diverge)
             if plot_label:
                 label = labels_dict.get(label, label.capitalize())
-                label = r"\textit{{{0}}}".format(label)
+                label = markup(r"\textit{{{0}}}".format(label))
                 root.text(x - 0.015, yy + yinterval / 2, label, ha="right", va="center")
             xy.draw()
             ax.set_xlim(*xlim)

src/jcvi/graphics/dotplot.py

@@ -261,6 +261,7 @@ def dotplot(
     stdpf: bool = True,
     chpf: bool = True,
     usetex: bool = True,
+    chinese: bool = False,
 ):
     """
     Draw a dotplot from an anchor file.
@@ -381,11 +382,18 @@ def dotplot(
             xstart += 0.1
 
     if title is None:
-        title = f"Inter-genomic comparison: {gx} vs {gy}"
-        if is_self:
-            title = f"Intra-genomic comparison within {gx}"
-            npairs //= 2
-        title += f" ({thousands(npairs)} gene pairs)"
+        if chinese:
+            title = f"基因组间比较：{gx} vs {gy}"
+            if is_self:
+                title = f"{gx} 基因组内比较"
+                npairs //= 2
+            title += f"（{thousands(npairs)} 个基因对）"
+        else:
+            title = f"Inter-genomic comparison: {gx} vs {gy}"
+            if is_self:
+                title = f"Intra-genomic comparison within {gx}"
+                npairs //= 2
+            title += f" ({thousands(npairs)} gene pairs)"
     root.set_title(title, x=0.5, y=0.96, color="k")
     if title:
         logger.debug("Dot plot title: %s", title)

src/jcvi/graphics/karyotype.py

@@ -383,10 +383,16 @@ def __init__(
             if generank:
                 sz = dict((x, len(list(bed.sub_bed(x)))) for x in seqids)
             else:
-                sz = sizes or dict(
-                    (x, max(z.end for z in list(bed.sub_bed(x)))) for x in seqids
-                )
-                assert sz is not None, "sizes not available and cannot be inferred"
+                if sizes:
+                    # Restrict the provided sizes to this track's seqids; otherwise
+                    # every track inherits the genome-wide total and the per-chr
+                    # gauge/coverage axes collapse.
+                    sz = dict((x, sizes[x]) for x in seqids if x in sizes)
+                else:
+                    sz = dict(
+                        (x, max(z.end for z in list(bed.sub_bed(x)))) for x in seqids
+                    )
+                assert sz, "sizes not available and cannot be inferred"
             t.seqids = seqids
             # validate if all seqids are non-empty
             for k, v in sz.items():

src/jcvi/graphics/tree.py

@@ -170,6 +170,7 @@ def draw_tree(
     wgdinfo=None,
     geoscale=False,
     groups=[],
+    chinese=False,
 ):
     """
     main function for drawing phylogenetic tree
@@ -240,6 +241,11 @@ def rescale_divergence(divergence):
         xx = rescale(dist)
 
         if n.is_leaf:
+            # On a time-calibrated (geoscale) tree all extant tips sit at the
+            # present, so align leaf tips at max_dist (keeps the outgroup leaf,
+            # e.g. the rerooted outgroup, flush with the others).
+            if geoscale:
+                xx = rescale(max_dist)
             yy = ystart + i * yinterval
             i += 1
 
@@ -337,7 +343,13 @@ def rescale_divergence(divergence):
     # scale bar
     if geoscale:
         draw_geoscale(
-            ax, ytop, margin=margin, rmargin=rmargin, yy=ymargin, max_dist=max_dist
+            ax,
+            ytop,
+            margin=margin,
+            rmargin=rmargin,
+            yy=ymargin,
+            max_dist=max_dist,
+            chinese=chinese,
         )
     else:
         br = 0.1
@@ -437,7 +449,14 @@ def read_trees(tree):
 
 
 def draw_geoscale(
-    ax, ytop, margin=0.1, rmargin=0.2, yy=0.1, max_dist=3.0, contrast_epochs=True
+    ax,
+    ytop,
+    margin=0.1,
+    rmargin=0.2,
+    yy=0.1,
+    max_dist=3.0,
+    contrast_epochs=True,
+    chinese=False,
 ):
     """
     Draw geological epoch on million year ago (mya) scale.
@@ -462,7 +481,7 @@ def cv(x):
     ax.text(
         (a + b) / 2,
         yy - 5 * tick,
-        "Time before present (million years)",
+        "距今时间（百万年）" if chinese else "Time before present (million years)",
         ha="center",
         va="center",
     )
@@ -478,6 +497,16 @@ def cv(x):
         ("Permian", 252.17, 298.9),
         ("Carboniferous", 298.9, 358.9),
     )
+    # Chinese display names; English keys are kept for the contrast logic below
+    CN_ERA = {
+        "Neogene": "新近纪",
+        "Paleogene": "古近纪",
+        "Cretaceous": "白垩纪",
+        "Jurassic": "侏罗纪",
+        "Triassic": "三叠纪",
+        "Permian": "二叠纪",
+        "Carboniferous": "石炭纪",
+    }
     h = 0.05
     for (era, start, end), color in zip(Geo, set3_n(len(Geo))):
         if maxx - start < 10:  # not visible enough
@@ -488,7 +517,7 @@ def cv(x):
         ax.text(
             (start + end) / 2,
             yy + (tick + h) / 2,
-            era,
+            CN_ERA.get(era, era) if chinese else era,
             ha="center",
             va="center",
             size=8,

src/jcvi/projects/jcvi.py

@@ -23,6 +23,7 @@
     plt,
     savefig,
     set1,
+    set_chinese_font,
     setup_theme,
 )
 from ..graphics.chromosome import draw_chromosomes
@@ -43,7 +44,11 @@ def synteny(args):
     p = OptionParser(synteny.__doc__)
     p.set_beds()
     opts, args, iopts = p.set_image_options(args, figsize="14x7")
-    setup_theme(style="dark")
+    setup_theme(style="dark", usetex=iopts.usetex)
+
+    if opts.chinese:
+        # The extra setup_theme() above resets the font, so re-apply the CJK font
+        set_chinese_font()
 
     if len(args) != 6:
         sys.exit(not p.print_help())
@@ -71,6 +76,8 @@ def synteny(args):
         is_self=is_self,
         chrlw=0.5,
         sepcolor=set1[3],
+        genomenames="葡萄_桃" if opts.chinese else None,
+        chinese=opts.chinese,
     )
 
     # Panel B
@@ -86,7 +93,7 @@ def synteny(args):
     panel_labels(root, labels)
     normalize_axes(root, ax1_root, ax2_root, ax3_root)
 
-    image_name = "synteny.pdf"
+    image_name = "synteny." + iopts.format
     savefig(image_name, dpi=iopts.dpi, iopts=iopts)