WIP capitalization changes.

Author: scossu

scriptshifter/hooks/general/__init__.py

@@ -27,6 +27,25 @@
 logger = getLogger(__name__)
 
 
+def capitalize_pre_assembly(ctx):
+    """
+    Capitalize a not-yet-assembled result list according to user options.
+    """
+    ctx.dest_ls = _capitalize(ctx.dest_ls, ctx.options.get("capitalize"))
+
+
+def capitalize_post_assembly(ctx):
+    """
+    Capitalize an already assembled result string according to user options.
+    """
+    dest_ls = ctx.dest.split(" ")
+
+    dest_ls = _capitalize(dest_ls, ctx.options.get("capitalize"))
+
+    return " ".join(dest_ls)
+
+
+
 def normalize_spacing_post_assembly(ctx):
     """
     Remove duplicate and unwanted whitespace around punctuation.
@@ -53,3 +72,18 @@ def normalize_spacing_post_assembly(ctx):
     # norm = NORM8_RE.sub(r"\1\2", norm)
 
     return norm
+
+
+def _capitalize(src, which):
+    """
+    Only capitalize first word and words preceded by space.
+
+    NOTE: this function is only used for capitalizing hook-generated
+    transliterations, which are not normally processed. Double cap rules are
+    not applicable here.
+    """
+    if which == "first":
+        ctx.dest_ls[0] = ctx.dest_ls[0].upper()
+
+    elif which == "all":
+        ctx.dest_ls = [tk[0].upper() + tk[1:] for tk in ctx.dest_ls]

scriptshifter/hooks/yiddish_/__init__.py

@@ -16,24 +16,16 @@
 from yiddish import detransliterate, transliterate
 
 from scriptshifter.exceptions import BREAK
-from scriptshifter.tools import capitalize
 
 
 def s2r_post_config(ctx):
     """
     Script to Roman.
     """
-    rom = transliterate(
+    ctx.dest = transliterate(
             ctx.src, loc=True,
             loshn_koydesh=ctx.options.get("loshn_koydesh"))
 
-    if ctx.options["capitalize"] == "all":
-        rom = capitalize(rom)
-    elif ctx.options["capitalize"] == "first":
-        rom = rom[0].upper() + rom[1:]
-
-    ctx.dest = rom
-
     return BREAK
 
 

scriptshifter/tools.py

@@ -112,7 +112,7 @@ def transliterate(src, lang, t_dir="s2r", capitalize=False, options={}):
             )
 
         # Normalize case before post_config and rule-based normalization.
-        if not ctx.general["case_sensitive"]:
+        if t_dir == FEAT_R2S and not ctx.general["case_sensitive"]:
             ctx._src = ctx.src.lower()
 
         # This hook may take over the whole transliteration process or delegate
@@ -270,7 +270,10 @@ def transliterate(src, lang, t_dir="s2r", capitalize=False, options={}):
                     # A match is found. Stop scanning tokens, append result,
                     # and proceed scanning the source.
 
-                    # Capitalization.
+                    # Capitalization. This applies double capitalization
+                    # rules. The external function in
+                    # scriptshifter.tools.capitalize used for non-table
+                    # languages does not.
                     if (
                         (ctx.options["capitalize"] == "first" and ctx.cur == 0)
                         or