add feature: include line number in toc parse error message

pdftocgen/app.py

@@ -90,7 +90,7 @@ def main():
         print(usage_s, file=sys.stderr)
         sys.exit(2)
 
-    recipe_file: TextIO = io.TextIOWrapper(sys.stdin.buffer, encoding='utf-8', errors='ignore')
+    recipe_file: TextIO = None
     readable: bool = False
     vpos: bool = False
     out: TextIO = io.TextIOWrapper(sys.stdout.buffer, encoding='utf-8', errors='ignore')
@@ -129,6 +129,8 @@ def main():
         print(usage_s, file=sys.stderr)
         sys.exit(1)
 
+    if not recipe_file:
+        recipe_file = io.TextIOWrapper(sys.stdin.buffer, encoding='utf-8', errors='ignore')
     path_in: str = args[0]
     # done parsing arguments
 

pdftocio/app.py

@@ -94,7 +94,7 @@ def main():
         print(usage_s, file=sys.stderr)
         sys.exit(2)
 
-    toc_file: TextIO = io.TextIOWrapper(sys.stdin.buffer, encoding='utf-8', errors='ignore')
+    toc_file: TextIO = None
     print_toc: bool = False
     readable: bool = False
     out: Optional[str] = None
@@ -131,6 +131,8 @@ def main():
         print(usage_s, file=sys.stderr)
         sys.exit(1)
 
+    if not toc_file:
+        toc_file = io.TextIOWrapper(sys.stdin.buffer, encoding='utf-8', errors='ignore')
     path_in: str = args[0]
     # done parsing arguments
 

pdftocio/tocparser.py

@@ -5,10 +5,10 @@
 
 from typing import IO, List
 from fitzutils import ToCEntry
-from itertools import takewhile
+from itertools import count, takewhile
 
 
-def parse_entry(entry: List) -> ToCEntry:
+def parse_entry(entry: List, nLine: int) -> ToCEntry:
     """parse a row in csv to a toc entry"""
 
     # a somewhat weird hack, csv reader would read spaces as an empty '', so we
@@ -23,8 +23,13 @@ def parse_entry(entry: List) -> ToCEntry:
             *entry[indent + 2:]      # vpos
         )
         return toc_entry
+    except ValueError as e:
+        print(f"Unable to parse toc entry {entry} from line {nLine};",
+              f"Couldn't convert '{entry[indent + 1]}' to a page number.",
+              file=sys.stderr)
+        raise e
     except IndexError as e:
-        print(f"Unable to parse toc entry {entry};",
+        print(f"Unable to parse toc entry {entry} from line {nLine};",
               f"Need at least {indent + 2} parts but only have {len(entry)}.",
               "Make sure the page number is present.",
               file=sys.stderr)
@@ -35,4 +40,4 @@ def parse_toc(file: IO) -> List[ToCEntry]:
     """Parse a toc file to a list of toc entries"""
     reader = csv.reader(file, lineterminator='\n',
                         delimiter=' ', quoting=csv.QUOTE_NONNUMERIC)
-    return list(map(parse_entry, reader))
+    return list(map(parse_entry, reader, count(1)))