🐛 Fix some complex emojis not being converted and enhance test coverage (#19)

src/dismoji/__init__.py

@@ -12,10 +12,13 @@ EMOJIS_PATH = Path(__file__).parent / "raw" / "build" / "emojis.json"
 with EMOJIS_PATH.open("r", encoding="utf-8") as f:
     EMOJIS = json.load(f)
 
-EMOJI_MAPPING: dict[str, str] = {k: EMOJIS["emojis"][v]["surrogates"] for k, v in EMOJIS["nameToEmoji"].items()}
+_VARIATION_SELECTOR = "\ufe0f"  # We remove this as it is not needed by discord and causes issues with tests
+EMOJI_MAPPING: dict[str, str] = {
+    k: EMOJIS["emojis"][v]["surrogates"].replace(_VARIATION_SELECTOR, "") for k, v in EMOJIS["nameToEmoji"].items()
+}
 
-# Create a reverse mapping for demojizing (emoji to name)
 REVERSE_EMOJI_MAPPING: dict[str, str] = {}
+
 for emoji_index_str, emoji_index in sorted(EMOJIS["surrogateToEmoji"].items(), key=lambda x: len(x[0]), reverse=True):
     # Get the first name in the list as the preferred name
     e = EMOJIS["emojis"][emoji_index]
@@ -23,9 +26,9 @@ for emoji_index_str, emoji_index in sorted(EMOJIS["surrogateToEmoji"].items(), k
     # e.g. :handshake_light_skin_tone_dark_skin_tone: vs :handshake_tone1_tone5:
     REVERSE_EMOJI_MAPPING[emoji_index_str] = e["names"][-1 if e.get("hasMultiDiversityParent") else 0]
 
-del EMOJIS  # Clean up to save memory
+del EMOJIS, _VARIATION_SELECTOR  # Clean up to save memory
 
-EMOJI_PATTERN = re.compile(r":([a-zA-Z0-9_-]+):")
+EMOJI_PATTERN = re.compile(r":([\w+-]+):")
 
 EMOJI_CHARS_PATTERN = re.compile("|".join(map(re.escape, REVERSE_EMOJI_MAPPING.keys())))
 

tests/emoji_test.py

@@ -1,7 +1,26 @@
 # Copyright (c) Paillat-dev
 # SPDX-License-Identifier: MIT
 
-from dismoji import REVERSE_EMOJI_MAPPING, demojize, emojize
+from dismoji import EMOJI_MAPPING, REVERSE_EMOJI_MAPPING, demojize, emojize
+
+
+def are_equal(a: str, b: str) -> bool:
+    """Check if two emojis are equal.
+
+    Allows for comparing emojis with modifiers even when they are in different orders.
+
+    Args:
+        a (str): First emoji string.
+        b (str): Second emoji string.
+
+    Returns:
+        bool: True if the emojis are equal, False otherwise.
+    """
+    if len(a) != len(b):
+        return False
+    if len(a) == 1:
+        return a == b
+    return a[0] == b[0] and set(a[1:]) == set(b[1:])
 
 
 def test_basic() -> None:
@@ -70,6 +89,11 @@ def test_emoji_with_special_characters() -> None:
         assert emojize(input_str) == expected_output
 
 
+def test_emojize_all() -> None:
+    for name, emoji in EMOJI_MAPPING.items():
+        assert are_equal(emojize(f":{name}:"), emoji)
+
+
 def test_demojize_basic() -> None:
     """Test basic functionality of demojize function."""
     assert demojize("Hello 😄") == "Hello :smile:"