Make _py_grapheme standalone by generating property tables

Add makegraphemedata() to Tools/unicode/makeunicodedata.py that
generates Lib/_py_grapheme_db.py from the Unicode data files
(GraphemeBreakProperty.txt, emoji-data.txt, DerivedCoreProperties.txt).

_py_grapheme.py now imports property tables from _py_grapheme_db
and uses bisect for lookups instead of calling unicodedata functions
added in 3.15. This makes the module usable on Python 3.13 and 3.14
by regenerating the tables for the appropriate Unicode version.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: ambv

Lib/_py_grapheme.py

@@ -1,10 +1,23 @@
 """Pure Python implementation of unicodedata.iter_graphemes().
 
 Uses the extended grapheme cluster rules from Unicode TR29.
+
+Property tables are in _py_grapheme_db.py, generated by
+Tools/unicode/makeunicodedata.py from the Unicode Character Database.
 """
 
 import sys
-import unicodedata
+from bisect import bisect_right
+
+from _py_grapheme_db import (
+    GCB_RANGES,
+    GCB_Other, GCB_Prepend, GCB_CR, GCB_LF, GCB_Control,
+    GCB_Extend, GCB_Regional_Indicator, GCB_SpacingMark,
+    GCB_L, GCB_V, GCB_T, GCB_LV, GCB_LVT, GCB_ZWJ,
+    EXT_PICT_RANGES,
+    INCB_RANGES,
+    InCB_None, InCB_Linker, InCB_Consonant, InCB_Extend,
+)
 
 
 class Segment:
@@ -24,28 +37,6 @@ def __repr__(self):
         return f"<Segment {self.start}:{self.end}>"
 
 
-# Grapheme_Cluster_Break property values (matching C #defines)
-_GCB_Other = "Other"
-_GCB_Prepend = "Prepend"
-_GCB_CR = "CR"
-_GCB_LF = "LF"
-_GCB_Control = "Control"
-_GCB_Extend = "Extend"
-_GCB_Regional_Indicator = "Regional_Indicator"
-_GCB_SpacingMark = "SpacingMark"
-_GCB_L = "L"
-_GCB_V = "V"
-_GCB_T = "T"
-_GCB_LV = "LV"
-_GCB_LVT = "LVT"
-_GCB_ZWJ = "ZWJ"
-
-# Indic_Conjunct_Break property values
-_InCB_None = "None"
-_InCB_Linker = "Linker"
-_InCB_Consonant = "Consonant"
-_InCB_Extend = "Extend"
-
 # Extended Pictographic FSM states (for GB11)
 _EP_INIT = 0
 _EP_STARTED = 1
@@ -58,65 +49,97 @@ def __repr__(self):
 _INCB_LINKER = 2
 _INCB_MATCHED = 3
 
+# Hangul syllable constants
+_HANGUL_S_BASE = 0xAC00
+_HANGUL_S_COUNT = 11172
+_HANGUL_T_COUNT = 28
+
+# Precomputed start arrays for bisect lookup
+_GCB_STARTS = tuple(r[0] for r in GCB_RANGES)
+_EXT_PICT_STARTS = tuple(r[0] for r in EXT_PICT_RANGES)
+_INCB_STARTS = tuple(r[0] for r in INCB_RANGES)
+
 
-def _update_ext_pict_state(state, gcb, ext_pict):
-    if ext_pict:
-        return _EP_MATCHED if state == _EP_ZWJ else _EP_STARTED
-    if state == _EP_STARTED or state == _EP_MATCHED:
-        if gcb == _GCB_Extend:
-            return _EP_STARTED
-        if gcb == _GCB_ZWJ:
-            return _EP_ZWJ
-    return _EP_INIT
+# ---------------------------------------------------------------------------
+# Property lookup functions
+# ---------------------------------------------------------------------------
 
+def _get_gcb(cp):
+    """Return the Grapheme_Cluster_Break value for a codepoint."""
+    idx = bisect_right(_GCB_STARTS, cp) - 1
+    if idx >= 0:
+        entry = GCB_RANGES[idx]
+        if cp <= entry[1]:
+            return entry[2]
+    # Hangul syllables: LV if (cp - S_BASE) % T_COUNT == 0, else LVT
+    if _HANGUL_S_BASE <= cp < _HANGUL_S_BASE + _HANGUL_S_COUNT:
+        if (cp - _HANGUL_S_BASE) % _HANGUL_T_COUNT == 0:
+            return GCB_LV
+        return GCB_LVT
+    return GCB_Other
 
-def _update_incb_state(state, incb):
-    if incb == _InCB_Consonant:
-        return _INCB_MATCHED if state == _INCB_LINKER else _INCB_STARTED
-    if state != _INCB_INIT:
-        if incb == _InCB_Extend:
-            return _INCB_LINKER if state == _INCB_LINKER else _INCB_STARTED
-        if incb == _InCB_Linker:
-            return _INCB_LINKER
-    return _INCB_INIT
 
+def _get_ext_pict(cp):
+    """Return True if the codepoint has the Extended_Pictographic property."""
+    idx = bisect_right(_EXT_PICT_STARTS, cp) - 1
+    return idx >= 0 and cp <= EXT_PICT_RANGES[idx][1]
+
+
+def _get_incb(cp):
+    """Return the Indic_Conjunct_Break value for a codepoint."""
+    idx = bisect_right(_INCB_STARTS, cp) - 1
+    if idx >= 0:
+        entry = INCB_RANGES[idx]
+        if cp <= entry[1]:
+            return entry[2]
+    return InCB_None
+
+
+# ---------------------------------------------------------------------------
+# Grapheme break algorithm (TR29)
+# ---------------------------------------------------------------------------
 
 def _grapheme_break(prev_gcb, curr_gcb, ep_state, ri_flag, incb_state):
     """Return True if a grapheme cluster break occurs between two characters."""
     # GB3: Do not break between a CR and LF.
-    if prev_gcb == _GCB_CR and curr_gcb == _GCB_LF:
+    if prev_gcb == GCB_CR and curr_gcb == GCB_LF:
         return False
 
     # GB4: Break after controls.
-    if prev_gcb in (_GCB_CR, _GCB_LF, _GCB_Control):
+    if prev_gcb == GCB_CR or prev_gcb == GCB_LF or prev_gcb == GCB_Control:
         return True
 
     # GB5: Break before controls.
-    if curr_gcb in (_GCB_CR, _GCB_LF, _GCB_Control):
+    if curr_gcb == GCB_CR or curr_gcb == GCB_LF or curr_gcb == GCB_Control:
         return True
 
     # GB6: Do not break Hangul syllable sequences (L).
-    if prev_gcb == _GCB_L and curr_gcb in (_GCB_L, _GCB_V, _GCB_LV, _GCB_LVT):
+    if prev_gcb == GCB_L and (
+        curr_gcb == GCB_L or curr_gcb == GCB_V
+        or curr_gcb == GCB_LV or curr_gcb == GCB_LVT
+    ):
         return False
 
     # GB7: Do not break Hangul syllable sequences (LV, V).
-    if prev_gcb in (_GCB_LV, _GCB_V) and curr_gcb in (_GCB_V, _GCB_T):
+    if (prev_gcb == GCB_LV or prev_gcb == GCB_V) and (
+        curr_gcb == GCB_V or curr_gcb == GCB_T
+    ):
         return False
 
     # GB8: Do not break Hangul syllable sequences (LVT, T).
-    if prev_gcb in (_GCB_LVT, _GCB_T) and curr_gcb == _GCB_T:
+    if (prev_gcb == GCB_LVT or prev_gcb == GCB_T) and curr_gcb == GCB_T:
         return False
 
     # GB9: Do not break before extending characters or ZWJ.
-    if curr_gcb in (_GCB_Extend, _GCB_ZWJ):
+    if curr_gcb == GCB_Extend or curr_gcb == GCB_ZWJ:
         return False
 
     # GB9a: Do not break before SpacingMarks.
-    if curr_gcb == _GCB_SpacingMark:
+    if curr_gcb == GCB_SpacingMark:
         return False
 
     # GB9b: Do not break after Prepend characters.
-    if prev_gcb == _GCB_Prepend:
+    if prev_gcb == GCB_Prepend:
         return False
 
     # GB9c: Do not break within Indic conjunct clusters.
@@ -128,13 +151,17 @@ def _grapheme_break(prev_gcb, curr_gcb, ep_state, ri_flag, incb_state):
         return False
 
     # GB12/GB13: Do not break within emoji flag sequences.
-    if prev_gcb == _GCB_Regional_Indicator and curr_gcb == _GCB_Regional_Indicator:
+    if prev_gcb == GCB_Regional_Indicator and curr_gcb == GCB_Regional_Indicator:
         return ri_flag
 
     # GB999: Otherwise, break everywhere.
     return True
 
 
+# ---------------------------------------------------------------------------
+# Public API
+# ---------------------------------------------------------------------------
+
 def iter_graphemes(string, /, start=0, end=sys.maxsize):
     """Iterate over grapheme clusters in a string.
 
@@ -150,7 +177,6 @@ def iter_graphemes(string, /, start=0, end=sys.maxsize):
         )
 
     length = len(string)
-    # Adjust indices (matching CPython's ADJUST_INDICES macro)
     if end > length:
         end = length
     if end < 0:
@@ -166,22 +192,51 @@ def iter_graphemes(string, /, start=0, end=sys.maxsize):
 
 
 def _iter_grapheme_clusters(string, start, end):
-    gcb = _GCB_Other
+    gcb = GCB_Other
     ep_state = _EP_INIT
     incb_state = _INCB_INIT
     ri_flag = False
 
     cluster_start = start
     pos = start
     while pos < end:
-        ch = string[pos]
-        curr_gcb = unicodedata.grapheme_cluster_break(ch)
-        ext_pict = unicodedata.extended_pictographic(ch)
-        incb = unicodedata.indic_conjunct_break(ch)
-
-        ep_state = _update_ext_pict_state(ep_state, curr_gcb, ext_pict)
-        ri_flag = (not ri_flag) if curr_gcb == _GCB_Regional_Indicator else False
-        incb_state = _update_incb_state(incb_state, incb)
+        cp = ord(string[pos])
+        curr_gcb = _get_gcb(cp)
+
+        # Update Extended Pictographic FSM (GB11)
+        ext_pict = _get_ext_pict(cp)
+        if ext_pict:
+            ep_state = _EP_MATCHED if ep_state == _EP_ZWJ else _EP_STARTED
+        elif ep_state == _EP_STARTED or ep_state == _EP_MATCHED:
+            if curr_gcb == GCB_Extend:
+                ep_state = _EP_STARTED
+            elif curr_gcb == GCB_ZWJ:
+                ep_state = _EP_ZWJ
+            else:
+                ep_state = _EP_INIT
+        else:
+            ep_state = _EP_INIT
+
+        # Update Regional Indicator flag (GB12/GB13)
+        ri_flag = not ri_flag if curr_gcb == GCB_Regional_Indicator else False
+
+        # Update Indic Conjunct Break FSM (GB9c)
+        curr_incb = _get_incb(cp)
+        if curr_incb == InCB_Consonant:
+            incb_state = (
+                _INCB_MATCHED if incb_state == _INCB_LINKER else _INCB_STARTED
+            )
+        elif incb_state != _INCB_INIT:
+            if curr_incb == InCB_Extend:
+                incb_state = (
+                    _INCB_LINKER if incb_state == _INCB_LINKER else _INCB_STARTED
+                )
+            elif curr_incb == InCB_Linker:
+                incb_state = _INCB_LINKER
+            else:
+                incb_state = _INCB_INIT
+        else:
+            incb_state = _INCB_INIT
 
         prev_gcb = gcb
         gcb = curr_gcb