From 1ccba65c9aeda9e3dd95aad5762fc5aa304914f4 Mon Sep 17 00:00:00 2001
From: Torben <59419684+entorb@users.noreply.github.com>
Date: Sat, 23 Nov 2024 09:07:46 +0100
Subject: [PATCH] check_chapters

---
 scripts/check_chapters.py      | 24 ++++++++++-
 scripts/check_chapters_test.py | 75 ++++++++++++++++++++++------------
 2 files changed, 72 insertions(+), 27 deletions(-)

diff --git a/scripts/check_chapters.py b/scripts/check_chapters.py
index bf323664b..fc9142196 100755
--- a/scripts/check_chapters.py
+++ b/scripts/check_chapters.py
@@ -184,8 +184,17 @@ def fix_ellipsis(s: str) -> str:
     # remove all spaces around ellipsis
     s = re.sub(r" *… *", r"…", s)
 
-    # after punctuation: add space
-    s = re.sub(r"(?<=[\.\?!:,;])…", r" …", s)
+    if settings["lang"] != "DE":
+        # after punctuation: add space
+        s = re.sub(r"(?<=[\.\?!:,;])…", r" …", s)
+
+    # new rule for German
+    if settings["lang"] == "DE":
+        # before: add space if not at start of line or quote
+        s = re.sub(r"(?<![ „‚\n^])…", r" …", s)
+        # after: add space if not followed by punctuation
+        s = re.sub(r"…(?![ \.\?!:,;“‘\n$])", r"… ", s)
+
     return s
 
 
@@ -474,6 +483,17 @@ def fix_hyphens(s: str) -> str:
     s = re.sub(r"(\d)\-(?=\d)", r"\1–", s)
     # NOT: mid-length dash ->  em dash (caution: false positives!)
     # s = s.replace("–", "—")
+
+    # new rule for German
+    if settings["lang"] == "DE":
+        # remove all spaces around hyphens
+        s = re.sub(r" *— *", r"—", s)
+
+        # before: add space if not at start of line or quote
+        s = re.sub(r"(?<![ „‚\n^])—", r" —", s)
+        # after: add space if not followed by punctuation
+        s = re.sub(r"—(?![ \.\?!:,;“‘\n$])", r"— ", s)
+
     return s
 
 
diff --git a/scripts/check_chapters_test.py b/scripts/check_chapters_test.py
index 9dfcf4447..b366107b2 100644
--- a/scripts/check_chapters_test.py
+++ b/scripts/check_chapters_test.py
@@ -50,13 +50,33 @@ def test_fix_common_typos(lang: str) -> None:
 @pytest.mark.parametrize("lang", ["EN", "DE"])
 def test_fix_ellipsis(lang: str) -> None:
     settings["lang"] = lang
-    pairs = [
-        ("foo…bar", "foo…bar"),
-        ("foo … bar", "foo…bar"),
-        ("foo… bar", "foo…bar"),
-        ("foo …bar", "foo…bar"),
-        ("foo, …", "foo, …"),
-    ]
+    pairs = []
+    if lang != "DE":
+        pairs.extend(
+            [
+                ("foo...bar", "foo…bar"),
+                ("foo…bar", "foo…bar"),
+                ("foo … bar", "foo…bar"),
+                ("foo… bar", "foo…bar"),
+                ("foo …bar", "foo…bar"),
+                ("foo, …", "foo, …"),
+                ("foo …! bar", "foo…! bar"),
+            ]
+        )
+    if lang == "DE":
+        pairs.extend(
+            [
+                ("foo...bar", "foo … bar"),
+                ("foo…bar", "foo … bar"),
+                ("foo … bar", "foo … bar"),
+                ("foo… bar", "foo … bar"),
+                ("foo …bar", "foo … bar"),
+                ("foo, …“", "foo, …“"),
+                ("foo,…“", "foo, …“"),
+                ("foo …! bar", "foo …! bar"),
+            ]
+        )
+
     checkit(fix_ellipsis, pairs)
 
 
@@ -90,26 +110,31 @@ def test_fix_hyphens(lang: str) -> None:
     settings["lang"] = lang
     pairs = [
         ("2-3-4", "2–3–4"),
-        (" —,", "—,"),
-        (" —.", "—."),
-        (" —!", "—!"),
-        (" —?", "—?"),
-        ("— asdf", "—asdf"),
-        ("- asdf", "—asdf"),
-        ("-asdf", "—asdf"),
     ]
+    if lang != "DE":
+        pairs.extend(
+            (
+                (" —,", "—,"),
+                (" —.", "—."),
+                (" —!", "—!"),
+                (" —?", "—?"),
+                ("— asdf", "—asdf"),
+                ("- asdf", "—asdf"),
+                ("-asdf", "—asdf"),
+            )
+        )
     if lang == "DE":
         pairs.extend(
-            [
-                ("Text —", "Text—"),
-                ("Text—„", "Text— „"),
-                ("Text —„", "Text— „"),
-                ("Text „ —Quote", "Text „—Quote"),
-                ("Text „ — Quote", "Text „—Quote"),
-                ("Text—„— Quote", "Text— „—Quote"),
-                ("Text -“asdf", "Text—“ asdf"),
-                ("Text —“", "Text—“"),
-            ]
+            (
+                ("Text — Text", "Text — Text"),
+                ("Text—„", "Text — „"),
+                ("Text —„", "Text — „"),
+                ("Text „ —Quote", "Text „— Quote"),
+                ("Text „ — Quote", "Text „— Quote"),
+                ("Text—„— Quote", "Text — „— Quote"),
+                ("Text -“asdf", "Text —“ asdf"),
+                ("Text —“", "Text —“"),
+            )
         )
     checkit(fix_hyphens, pairs)
 
@@ -217,4 +242,4 @@ def checkit(fct: Callable, pairs: list[tuple[str, str]]) -> None:
         # test in complete fix_line context
         assert (
             fix_line(text) == expected_output
-        ), f"'{fix_line(text)}' != '{expected_output}'"
+        ), f"'{fix_line(text)}' != '{expected_output}' (fix_line)"