Added escape() and unescape()

tikuma-lsuhsc · tikuma-lsuhsc · commit 32340be42fa8 · 2022-03-20T17:08:23.000-05:00
diff --git a/src/ffmpegio/utils/__init__.py b/src/ffmpegio/utils/__init__.py
@@ -4,6 +4,81 @@
 from .._utils import *
 
 
+def escape(txt):
+    """apply FFmpeg single quote escaping
+
+    :param txt: Unescaped string
+    :type txt: any stringifiable object
+    :return: Escaped string
+    :rtype: str
+
+    See https://ffmpeg.org/ffmpeg-utils.html#Quoting-and-escaping
+    """
+
+    txt = str(txt)
+
+    if re.search(r"\s", txt, re.MULTILINE):
+        # quote if txt has any white space
+        txt = txt.replace("'", r"'\''")
+        return f"'{txt}'"
+    else:
+        # if not quoted, escape quotes and backslashes
+        return re.sub(r"(['\\])", r"\\\1", txt)
+
+
+def unescape(txt):
+    """undo FFmpeg single quote escaping
+
+    :param txt: Escaped string
+    :type txt: str
+    :return: Original string
+    :rtype: str
+
+    See https://ffmpeg.org/ffmpeg-utils.html#Quoting-and-escaping
+    """
+
+    n = len(txt)
+    if not n:
+        return txt
+
+    re_start = re.compile(r"[^\\](?:\\\\)*'")
+    re_sub = re.compile(r"\\([\\'])")
+
+    blks = []
+
+    # look for a first quoted text block
+    m = re.search(r"(?:^|[^\\])(?:\\\\)*'", txt)
+    if m:
+        i0 = m.end()
+        if i0 > 1:
+            # unescape the initial unquoted block
+            blks.append(re_sub.sub(r"\1", txt[0 : i0 - 1]))
+    else:
+        # no quoted text block, unescape the whole string
+        return re_sub.sub(r"\1", txt)
+
+    # always starts with quoted block
+    in_quote = True
+
+    while i0 < n:
+
+        if in_quote:
+            # find the end quote
+            i1 = txt.find("'", i0)
+            if i1 < 0:
+                raise ValueError("incorrectly escaped text: missing a closing quote.")
+            blks.append(txt[i0:i1])
+        else:
+            # find the next starting quote
+            m = re_start.search(txt, i0 - 1)
+            i1 = m.end() - 1 if m else n
+            blks.append(re_sub.sub(r"\1", txt[i0:i1]))
+        i0 = i1 + 1
+        in_quote = not in_quote
+
+    return "".join(blks)
+
+
 def parse_spec_stream(spec, file_index=False):
     if isinstance(spec, str):
         out = {}
diff --git a/tests/test_utils.py b/tests/test_utils.py
@@ -2,6 +2,37 @@
 from ffmpegio import utils
 import pytest
 
+def test_string_escaping():
+    raw = "Crime d'Amour"
+    esc = utils.escape(raw)
+    assert esc == r"'Crime d'\''Amour'"
+    assert utils.unescape(esc) == raw
+
+    raw = "  this string starts and ends with whitespaces  "
+    esc = utils.escape(raw)
+    assert esc == "'  this string starts and ends with whitespaces  '"
+    assert utils.unescape(esc) == raw
+
+    esc = r"' The string '\'string\'' is a string '"
+    raw = r" The string 'string' is a string "
+    assert raw == utils.unescape(utils.escape(raw))
+    assert raw == utils.unescape(esc)
+
+    esc = r"'c:\foo' can be written as c:\\foo"
+    raw = r"c:\foo can be written as c:\foo"
+    assert raw == utils.unescape(esc)
+    assert raw == utils.unescape(utils.escape(raw))
+
+    raw = "d'Amour"
+    esc = utils.escape(raw)
+    assert esc == r"d\'Amour"
+    assert utils.unescape(esc) == raw
+
+    raw = r"c:\foo"
+    esc = utils.escape(raw)
+    assert esc == r"c:\\foo"
+    assert utils.unescape(esc) == raw
+
 
 def test_parse_spec_stream():
     assert utils.parse_spec_stream(1) == {"index": 1}
@@ -81,5 +112,3 @@ def test_get_audio_format():
 if __name__ == "__main__":
     import re
 
-    spec = "p:4"
-    print(re.split(r"(?<![pi]\:|m\:.+?\:)\:", spec))
