warn on unicode decoding errors in PDF annotations

in some cases the the annotations may contain some junk that hinders annotations processing altogether. This allows to ignore the error and warn instead, which is configurable via warn_unicode_error arguments in the PDF initializer and/or open() method.

pdfplumber/page.py

@@ -13,6 +13,7 @@ from typing import (
     Union,
 )
 from unicodedata import normalize as normalize_unicode
+from warnings import warn
 
 from pdfminer.converter import PDFPageAggregator
 from pdfminer.layout import (
@@ -306,7 +307,12 @@ class Page(Container):
                     try:
                         extras[k] = v.decode("utf-8")
                     except UnicodeDecodeError:
-                        extras[k] = v.decode("utf-16")
+                        try:
+                            extras[k] = v.decode("utf-16")
+                        except UnicodeDecodeError:
+                            if not self.pdf.warn_unicode_error:
+                                raise
+                            warn(f"Could not decode {k} for annotation. {k} will be missing.")
 
             parsed = {
                 "page_number": self.page_number,

pdfplumber/pdf.py

@@ -35,6 +35,7 @@ class PDF(Container):
         password: Optional[str] = None,
         strict_metadata: bool = False,
         unicode_norm: Optional[Literal["NFC", "NFKC", "NFD", "NFKD"]] = None,
+        warn_unicode_error: bool = False,
     ):
         self.stream = stream
         self.stream_is_external = stream_is_external
@@ -43,6 +44,7 @@ class PDF(Container):
         self.laparams = None if laparams is None else LAParams(**laparams)
         self.password = password
         self.unicode_norm = unicode_norm
+        self.warn_unicode_error = warn_unicode_error
 
         self.doc = PDFDocument(PDFParser(stream), password=password or "")
         self.rsrcmgr = PDFResourceManager()
@@ -76,6 +78,7 @@ class PDF(Container):
         repair: bool = False,
         gs_path: Optional[Union[str, pathlib.Path]] = None,
         repair_setting: T_repair_setting = "default",
+        warn_unicode_error: bool = False,
     ) -> "PDF":
 
         stream: Union[BufferedReader, BytesIO]
@@ -107,6 +110,7 @@ class PDF(Container):
                 strict_metadata=strict_metadata,
                 unicode_norm=unicode_norm,
                 stream_is_external=stream_is_external,
+                warn_unicode_error=warn_unicode_error,
             )
 
         except PSException: