"""A Python Unicode Shaping Engine.""" from fontFeatures import FontFeatures from .BaseShaper import BaseShaper from .ArabicShaper import ArabicShaper from .IndicShaper import IndicShaper from .MyanmarShaper import MyanmarShaper from .HangulShaper import HangulShaper from .KhmerShaper import KhmerShaper from .USEShaper import USEShaper import logging import re class Shaper: """Initialize a shaping engine. Args: ff: A :py:class:`fontFeatures.FontFeatures` object. font: A ``Babelfont`` font object. message_function: A function called with a message and buffer object. """ def __init__(self, ff, font, message_function=None): assert isinstance(ff, FontFeatures) self.fontfeatures = ff self.babelfont = font if message_function: self.msg = message_function else: self.msg = self.default_message_function def execute(self, buf, features=[]): """Run the shaper on a given buffer. Args: buf: :py:class:`fontFeatures.shaperLib.buffer.Buffer` object. features: either a list of feature tags or a feature string ("+foox,-barx") """ # Choose complex shaper self.complexshaper = self.categorize(buf)(self, self.babelfont, buf, features) self.msg("Using %s" % type(self.complexshaper).__name__) self.stages = [[]] if isinstance(features, str): self.user_features = self._parse_user_feature_string(features) else: self.user_features = features self.collect_features(buf) self.complexshaper.shape() buf.clear_mask() return buf def default_message_function(self, msg, buffer=None, serialize_options=None): """A logger function to be used if one is not provided in the constructor. By default, debugging and informational log messages are reported to the ``fontFeatures.shaperLib`` facility.""" ser = "" if buffer: ser = buffer.serialize(additional=serialize_options) msg = msg + " : " + ser logging.getLogger("fontFeatures.shaperLib").info(msg) def _parse_user_feature_string(self, features): features = features.split(",") outfeat = [] for f in features: f = f.rstrip() m = re.match(r"^([+\-]?)(\w+)", f) if m: outfeat.append({"tag": m[2], "value": m[1] != "-"}) continue m = re.match(r"^(\w+)=([10])", f) if m: outfeat.append({"tag": m[1], "value": m[2] == "1"}) else: outfeat.append({"tag": f, "value": True}) return outfeat def add_pause(self, thing=None): """Add a pause in the shaping stage. Used internally.""" if thing: self.stages.append(thing) self.stages.append([]) def add_features(self, *tags): """Adds feature tags to the current shaping stage.""" for t in tags: if any([isinstance(x, list) and t in x for x in self.stages]): continue self.stages[-1].append(t) def disable_feature(self, tag): """Removes a feature from processing.""" for s in self.stages: if isinstance(s, list) and tag in s: s.remove(tag) def collect_features(self, buf): """Determine the features, and their order, to process the buffer.""" self.add_features("rvrn") self.add_pause() if buf.direction == "LTR": self.add_features("ltra", "ltrm") elif buf.direction == "RTL": self.add_features("rtla", "rtlm") self.add_features("frac", "numr", "dnom", "rand") # trak? self.complexshaper.collect_features(self) # common features self.add_features("abvm", "blwm", "ccmp", "locl", "mark", "mkmk", "rlig") if buf.direction == "LTR" or buf.direction == "RTL": self.add_features("calt", "clig", "curs", "dist", "kern", "liga", "rclt") else: self.add_features("vert") for uf in self.user_features: if not uf["value"]: # Turn it off if it's already on self.disable_feature(uf["tag"]) else: self.add_features(uf["tag"]) if hasattr(self.complexshaper, "override_features"): self.complexshaper.override_features(self) def categorize(self, buf): """Returns the appropriate complex shaper class to shape this buffer.""" if buf.script == "Arabic": return ArabicShaper connected_scripts = { "Mongolian": "mong", "Syriac": "syrc", "Nko": "nko ", "Phags_Pa": "phag", "Mandaic": "mand", "Manichaean": "mand", "Psalter_Pahlavi": "phlp", "Adlam": "adlm", "Hanifi_Rohingya": "rohg", "Sogdian": "sogd", } if buf.script in connected_scripts: if self.fontfeatures.hasScriptSupport(connected_scripts[buf.script]): return ArabicShaper else: return BaseShaper # if buf.script in ["Thai", "Lao"]: # return ThaiShaper if buf.script == "Hangul": return HangulShaper # if buf.script == "Hebrew": # return HebrewShaper if buf.script in [ "Bengali", "Devanagari", "Gujarati", "Gurmukhi", "Kannada", "Malayalam", "Oriya", "Tamil", "Telugu", "Sinhala", ]: indic23map = { "Bengali": "bng", "Devanagari": "dev", "Gujarati": "gjr", "Gurmukhi": "gur", "Kannada": "knd", "Malayalam": "mlm", "Oriya": "ory", "Tamil": "tml", "Telugu": "tel", } # Sinhala is different if buf.script in indic23map and self.fontfeatures.hasScriptSupport( indic23map[buf.script] + "3" ): return USEShaper else: return IndicShaper if buf.script == "Khmer": return KhmerShaper if buf.script == "Myanmar": if self.fontfeatures.hasScriptSupport("mymr"): return BaseShaper else: return MyanmarShaper # if buf.script = "Qaag": return MyanmarZawgyiShaper if buf.script in [ "Tibetan", "Buhid", "Hanunoo", "Tagalog", "Tagbanwa", "Limbu", "Tai_Le", "Buginese", "Kharoshthi", "Syloti_Nagri", "Tifinagh", "Balinese", "Cham", "Kayah_Li", "Lepcha", "Rejang", "Saurashtra", "Sundanese", "Egyptian_Hieroglyphs", "Javanese", "Kaithi", "Meetei_Mayek", "Tai_Tham", "Tai_Viet", "Batak", "Brahmi", "Chakma", "Sharada", "Takri", "Duployan", "Grantha", "Khojki", "Khudawadi", "Mahajani", "Modi", "Pahawh_Hmong", "Siddham", "Tirhuta", "Ahom", "Bhaiksuki", "Marchen", "Newa", "Masaram_Gondi", "Soyombo", "Zanabazar_Square", "Dogra", "Gunjala_Gondi", "Makasar", "Nandinagari", ]: return USEShaper return BaseShaper def _script_direction(script): if script in [ "Arabic", "Hebrew", "Syriac", "Thaana", "Cypriot", "Kharoshthi", "Phoenician", "Nko", "Lydian", "Avestan", "Imperial_Aramaic", "Inscriptional_Parthian", "Inscriptional_Pahlavi", "Old_South_Arabian", "Old_Turkic", "Samaritan", "Mandaic", "Meroitic_Cursive", "Meroitic_Hieroglyphs", "Manichaean", "Mende_Kikakui", "Nabataean", "Old_North_Arabian", "Palmyrene", "Psalter_Pahlavi", "Hatran", "Adlam", "Hanifi_Rohingya", "Old_Sogdian", "Sogdian", "Elymaic", "Chorasmian", "Yezidi", ]: return "RTL" if script in ["Old_Hungarian", "Old_Italic", "Runic"]: return "invalid" return "LTR"