Complete Chinese numerals:

* Use standard table override instead of pre-config hooks.
* Add few test strings.

scriptshifter/hooks/chinese/__init__.py

@@ -2,42 +2,20 @@
 
 
 from logging import getLogger
-from os import path
 from re import I, compile, search, sub
 
-from yaml import load
-try:
-    from yaml import CLoader as Loader
-except ImportError:
-    from yaml import Loader
-
-
-HOOK_DIR = path.dirname(path.realpath(__file__))
 
 logger = getLogger(__name__)
 
 
-def merge_numerals_pre_config(tdata):
-    """
-    Add numerals mapping to configuration.
-
-    This overrides the existing character mappings.
-    """
-    num_map_yml = path.join(HOOK_DIR, "numerals.yml")
-    with open(num_map_yml, "r") as fh:
-        num_map = load(fh, Loader=Loader)
-
-    tdata["script_to_roman"]["map"].update(num_map)
-
-
 def parse_numerals(ctx):
     """
     Parse Chinese numerals in the already romanized result.
 
     This is run at post-assembly.
     """
     # Only apply to specific MARC fields.
-    use_num_v = ctx.options.get("marc_field") in ("245", "830")
+    use_num_v = ctx.options.get("marc_field") in ("245n", "830n")
 
     # tokens = split(r"[\W^#]", ctx.dest)  # Original logic.
     tokens = [tk.strip() for tk in ctx.dest_ls]

scriptshifter/tables/__init__.py

@@ -11,7 +11,7 @@
 except ImportError:
     from yaml import Loader
 
-from scriptshifter.exceptions import ConfigError
+from scriptshifter.exceptions import BREAK, ConfigError
 
 
 __doc__ = """
@@ -28,6 +28,7 @@
 
 # Available hook names.
 HOOKS = (
+    "pre_config",
     "post_config",
     "post_normalize",
     "begin_input_token",
@@ -126,7 +127,10 @@ def __hash__(self):
 @cache
 def list_tables():
     """
-    List all the available tables.
+    List all the indexed tables.
+
+    Note that this may not correspond to all the table files in the data
+    folder, but only those exposed in the index.
     """
     with open(path.join(TABLE_DIR, "index.yml")) as fh:
         tdata = load(fh, Loader=Loader)
@@ -150,7 +154,18 @@ def load_table(tname):
     with open(fname) as fh:
         tdata = load(fh, Loader=Loader)
 
-    # NOTE Only one level of inheritance. No need for recursion for now.
+    # Pre-config hooks.
+    # If any of these hooks returns BREAK, interrupt the configuration
+    # parsing and return whatever is obtained so far.
+    if "hooks" in tdata:
+        tdata["hooks"] = load_hook_fn(tname, tdata)
+    pre_cfg_hooks = tdata.get("hooks", {}).get("pre_config", [])
+    for hook_def in pre_cfg_hooks:
+        kwargs = hook_def[1] if len(hook_def) > 1 else {}
+        ret = hook_def[0](tdata, **kwargs)
+        if ret == BREAK:
+            return tdata
+
     parents = tdata.get("general", {}).get("parents", [])
 
     if "script_to_roman" in tdata: