implement 'format-separator' option (#2737)

a global option, that servers as a workaround for shortcomings due to
lack of a proper format string parser

docs/configuration.rst

@@ -3710,6 +3710,20 @@ Description
     this cache.
 
 
+format-separator
+----------------
+Type
+    ``string``
+Default
+    ``"/"``
+Description
+    Character(s) used as argument separator in format string
+    `format specifiers <formatting.md#format-specifiers>`__.
+
+    For example, setting this option to ``"#"`` would allow a replacement
+    operation to be ``Rold#new#`` instead of the default ``Rold/new/``
+
+
 signals-ignore
 --------------
 Type

gallery_dl/__init__.py

@@ -177,6 +177,12 @@ def main():
             extractor.modules = modules
             extractor._module_iter = iter(modules)
 
+        # format string separator
+        separator = config.get((), "format-separator")
+        if separator:
+            from . import formatter
+            formatter._SEPARATOR = separator
+
         # loglevels
         output.configure_logging(args.loglevel)
         if args.loglevel >= logging.ERROR:

gallery_dl/formatter.py

@@ -264,7 +264,7 @@ def _build_format_func(format_spec, default):
 
 
 def _parse_optional(format_spec, default):
-    before, after, format_spec = format_spec.split("/", 2)
+    before, after, format_spec = format_spec.split(_SEPARATOR, 2)
     before = before[1:]
     fmt = _build_format_func(format_spec, default)
 
@@ -284,7 +284,7 @@ def _parse_slice(format_spec, default):
 
 
 def _parse_maxlen(format_spec, default):
-    maxlen, replacement, format_spec = format_spec.split("/", 2)
+    maxlen, replacement, format_spec = format_spec.split(_SEPARATOR, 2)
     maxlen = text.parse_int(maxlen[1:])
     fmt = _build_format_func(format_spec, default)
 
@@ -295,7 +295,7 @@ def _parse_maxlen(format_spec, default):
 
 
 def _parse_join(format_spec, default):
-    separator, _, format_spec = format_spec.partition("/")
+    separator, _, format_spec = format_spec.partition(_SEPARATOR)
     separator = separator[1:]
     fmt = _build_format_func(format_spec, default)
 
@@ -305,7 +305,7 @@ def _parse_join(format_spec, default):
 
 
 def _parse_replace(format_spec, default):
-    old, new, format_spec = format_spec.split("/", 2)
+    old, new, format_spec = format_spec.split(_SEPARATOR, 2)
     old = old[1:]
     fmt = _build_format_func(format_spec, default)
 
@@ -315,7 +315,7 @@ def _parse_replace(format_spec, default):
 
 
 def _parse_datetime(format_spec, default):
-    dt_format, _, format_spec = format_spec.partition("/")
+    dt_format, _, format_spec = format_spec.partition(_SEPARATOR)
     dt_format = dt_format[1:]
     fmt = _build_format_func(format_spec, default)
 
@@ -325,7 +325,7 @@ def _parse_datetime(format_spec, default):
 
 
 def _parse_offset(format_spec, default):
-    offset, _, format_spec = format_spec.partition("/")
+    offset, _, format_spec = format_spec.partition(_SEPARATOR)
     offset = offset[1:]
     fmt = _build_format_func(format_spec, default)
 
@@ -363,6 +363,7 @@ class Literal():
 _literal = Literal()
 
 _CACHE = {}
+_SEPARATOR = "/"
 _GLOBALS = {
     "_env": lambda: os.environ,
     "_lit": lambda: _literal,

test/test_formatter.py

@@ -236,6 +236,26 @@ class TestFormatter(unittest.TestCase):
         # parse and format datetime
         self._run_test("{ds:D%Y-%m-%dT%H:%M:%S%z/%Y%m%d}", "20100101")
 
+    def test_separator(self):
+        orig_separator = formatter._SEPARATOR
+        try:
+            formatter._SEPARATOR = "|"
+            self._run_test("{a:Rh|C|RE|e|RL|l|}", "Cello wOrld")
+            self._run_test("{d[b]!s:R1|Q|R2|A|R0|Y|}", "Y")
+
+            formatter._SEPARATOR = "##"
+            self._run_test("{l:J-##Rb##E##}", "a-E-c")
+            self._run_test("{l:J-##[1:-1]}", "-b-")
+
+            formatter._SEPARATOR = "\0"
+            self._run_test("{d[a]:?<\0>\0L1\0too long\0}", "<too long>")
+            self._run_test("{d[c]:?<\0>\0L5\0too long\0}", "")
+
+            formatter._SEPARATOR = "?"
+            self._run_test("{ds:D%Y-%m-%dT%H:%M:%S%z?%Y%m%d}", "20100101")
+        finally:
+            formatter._SEPARATOR = orig_separator
+
     def test_globals_env(self):
         os.environ["FORMATTER_TEST"] = value = self.kwdict["a"]