feat: normalize option

2026-04-08 07:32:25 +02:00
parent 1e0f8e2868
commit 587d5dbe46
5 changed files with 245 additions and 22 deletions
@@ -4,7 +4,7 @@ build-backend = "hatchling.build"

 [project]
 name = "lrx-cli"
-version = "0.6.2"
+version = "0.6.3"
 description = "Fetch line-synced lyrics for your music player."
 readme = "README.md"
 requires-python = ">=3.13"
@@ -99,7 +99,17 @@ def fetch(
    plain: Annotated[
        bool,
        cyclopts.Parameter(
-            name="--plain", negative="", help="Output only the raw lyrics without tags."
+            name="--plain",
+            negative="",
+            help="Output only plain lyrics without tags (highest priority over --normalize).",
+        ),
+    ] = False,
+    normalize: Annotated[
+        bool,
+        cyclopts.Parameter(
+            name="--normalize",
+            negative="",
+            help="Output normalized LRC (ignored when --plain is also set).",
        ),
    ] = False,
 ):
@@ -123,7 +133,12 @@ def fetch(
        logger.error("No lyrics found.")
        sys.exit(1)

-    print(result.lyrics.to_text(plain=plain))
+    if plain:
+        print(result.lyrics.to_plain())
+    elif normalize:
+        print(result.lyrics.to_normalized_text())
+    else:
+        print(result.lyrics.to_text())


 # search
@@ -179,7 +194,17 @@ def search(
    plain: Annotated[
        bool,
        cyclopts.Parameter(
-            name="--plain", negative="", help="Output only the raw lyrics without tags."
+            name="--plain",
+            negative="",
+            help="Output only plain lyrics without tags (highest priority over --normalize).",
+        ),
+    ] = False,
+    normalize: Annotated[
+        bool,
+        cyclopts.Parameter(
+            name="--normalize",
+            negative="",
+            help="Output normalized LRC (ignored when --plain is also set).",
        ),
    ] = False,
 ):
@@ -214,7 +239,12 @@ def search(
        logger.error("No lyrics found.")
        sys.exit(1)

-    print(result.lyrics.to_text(plain=plain))
+    if plain:
+        print(result.lyrics.to_plain())
+    elif normalize:
+        print(result.lyrics.to_normalized_text())
+    else:
+        print(result.lyrics.to_text())


 # export
@@ -253,7 +283,17 @@ def export(
    plain: Annotated[
        bool,
        cyclopts.Parameter(
-            name="--plain", negative="", help="Export only the raw lyrics without tags."
+            name="--plain",
+            negative="",
+            help="Export only plain lyrics (.txt, highest priority over --normalize).",
+        ),
+    ] = False,
+    normalize: Annotated[
+        bool,
+        cyclopts.Parameter(
+            name="--normalize",
+            negative="",
+            help="Export normalized LRC output (ignored when --plain is also set).",
        ),
    ] = False,
 ):
@@ -307,8 +347,10 @@ def export(
        with open(output, "w", encoding="utf-8") as f:
            if plain:
                f.write(result.lyrics.to_plain())
+            elif normalize:
+                f.write(result.lyrics.to_normalized_text())
            else:
-                f.write(str(result.lyrics))
+                f.write(result.lyrics.to_text())
        logger.info(f"Exported lyrics to {output}")
    except Exception as e:
        logger.error(f"Failed to write file: {e}")
@@ -233,6 +233,14 @@ def _is_single_doc_tag_line(line: str) -> Optional[tuple[str, str]]:
    return key, value


+def _parse_offset_value(value: str) -> Optional[int]:
+    """Parse doc offset value in milliseconds, returning None for invalid values."""
+    try:
+        return int(value.strip())
+    except ValueError:
+        return None
+
+
 class LRCData:
    _lines: list[BaseLine]
    _doc_tags: dict[str, str]
@@ -265,7 +273,7 @@ class LRCData:
        self._lines = parsed

    def __str__(self) -> str:
-        return self.to_text(plain=False, include_word_sync=False)
+        return self._serialize_lines(self._lines, include_word_sync=True)

    def __repr__(self) -> str:
        return f"LRCData(doc_tags={self._doc_tags!r}, lines={self._lines!r})"
@@ -293,7 +301,7 @@ class LRCData:
            else CacheStatus.SUCCESS_UNSYNCED
        )

-    def normalize_unsynced(self):
+    def normalize_unsynced(self) -> "LRCData":
        """Convert lyrics into unsynced LRC form with [00:00.00] tags.

        - Leading blank lyric lines are skipped.
@@ -329,6 +337,59 @@ class LRCData:
        ret._doc_tags = dict(self._doc_tags)
        return ret

+    def normalize(self) -> "LRCData":
+        """Normalize LRC for decode/export oriented output.
+
+        Rules:
+        - Move all doc tags to the beginning, preserving line order and duplicates.
+        - Keep doc tags unchanged except removing all offset tags.
+        - Remove word-sync tags.
+        - Convert untagged non-empty lyric lines to [00:00.00] lyrics.
+        - Drop empty lyric lines.
+        - Expand lyric lines with multiple time tags into one line per tag.
+        - Apply offset (ms) to lyric timestamps and sort by timestamp.
+        """
+        out_doc_tags: list[DocTagLine] = []
+        lyric_entries: list[tuple[int, str]] = []
+        offset_ms = 0
+
+        # Resolve offset first so it applies to all lyric lines, independent of tag position.
+        for line in self._lines:
+            if isinstance(line, DocTagLine) and line.key.strip().lower() == "offset":
+                parsed_offset = _parse_offset_value(line.value)
+                if parsed_offset is not None:
+                    offset_ms = parsed_offset
+
+        for line in self._lines:
+            if isinstance(line, DocTagLine):
+                if line.key.strip().lower() == "offset":
+                    continue
+                out_doc_tags.append(DocTagLine(key=line.key, value=line.value))
+                continue
+
+            assert isinstance(line, LyricLine)
+
+            lyric_text = line.text
+            if not lyric_text.strip():
+                continue
+
+            line_times = line.line_times_ms if line.line_times_ms else [0]
+            for time_ms in line_times:
+                shifted = max(0, time_ms + offset_ms)
+                lyric_entries.append((shifted, lyric_text))
+
+        lyric_entries.sort(key=lambda item: item[0])
+
+        out_lyrics: list[LyricLine] = [
+            LyricLine(line_times_ms=[time_ms], words=[LrcWordSegment(text=text)])
+            for time_ms, text in lyric_entries
+        ]
+
+        ret = LRCData()
+        ret._lines = [*out_doc_tags, *out_lyrics]
+        ret._doc_tags = {line.key: line.value for line in out_doc_tags}
+        return ret
+
    def to_plain(
        self,
        deduplicate: bool = False,
@@ -366,23 +427,32 @@ class LRCData:

        return "\n".join(sorted_lines).strip()

+    @staticmethod
+    def _serialize_lines(lines: list[BaseLine], include_word_sync: bool) -> str:
+        return "\n".join(
+            line.to_text(include_word_sync=include_word_sync) for line in lines
+        )
+
    def to_text(
        self,
-        plain: bool = False,
        include_word_sync: bool = False,
    ) -> str:
-        """Serialize to LRC text or plain text.
+        """Serialize to non-normalized LRC text.

-        - plain=True returns to_plain().
-        - include_word_sync controls rendering of per-word tags for word-sync lines.
+        - Unsynced lyrics are converted to [00:00.00]-tagged form.
+        - include_word_sync only controls rendering of per-word tags.
+        - This method does not apply normalize() rules.
        """
-        if plain:
-            return self.to_plain(deduplicate=False)
+        res = self if self.is_synced() else self.normalize_unsynced()
+        return self._serialize_lines(res._lines, include_word_sync=include_word_sync)

-        lines: list[str] = [
-            line.to_text(include_word_sync=include_word_sync) for line in self._lines
-        ]
-        return "\n".join(lines)
+    def to_normalized_text(self) -> str:
+        """Serialize using normalize() rules.
+
+        Normalized output always strips word-sync tags.
+        """
+        normalized = self.normalize()
+        return self._serialize_lines(normalized._lines, include_word_sync=False)


 def get_audio_path(audio_url: str, ensure_exists: bool = False) -> Optional[Path]:
@@ -149,6 +149,83 @@ def test_normalize_unsynced_result_is_always_unsynced() -> None:
    assert normalized.detect_sync_status() is CacheStatus.SUCCESS_UNSYNCED


+def test_normalize_moves_doc_tags_to_top_and_removes_offset_tag() -> None:
+    text = "\n".join(
+        [
+            "[00:02.00]b",
+            "[ar:Artist]",
+            "[offset:500]",
+            "[00:01.00]a",
+            "[ti:Song]",
+        ]
+    )
+
+    normalized = LRCData(text).to_normalized_text()
+
+    assert normalized == "\n".join(
+        [
+            "[ar:Artist]",
+            "[ti:Song]",
+            "[00:01.50]a",
+            "[00:02.50]b",
+        ]
+    )
+
+
+def test_normalize_expands_multi_time_tags_and_sorts_lyrics() -> None:
+    text = "\n".join(
+        [
+            "[00:03.00]c",
+            "[00:02.00][00:01.00]x",
+        ]
+    )
+
+    normalized = LRCData(text).to_normalized_text()
+
+    assert normalized == "\n".join(["[00:01.00]x", "[00:02.00]x", "[00:03.00]c"])
+
+
+def test_normalize_converts_unsynced_lines_and_removes_word_sync_tags() -> None:
+    text = "\n".join(
+        [
+            "plain",
+            "<00:01.00>he <00:01.50>llo",
+            "[00:02.00]<00:02.20>world",
+            "",
+        ]
+    )
+
+    normalized = LRCData(text).to_normalized_text()
+
+    assert normalized == "\n".join(
+        [
+            "[00:00.00]plain",
+            "[00:00.00]he llo",
+            "[00:02.00]world",
+        ]
+    )
+
+
+def test_to_normalized_text_is_separate_from_plain() -> None:
+    data = LRCData("[offset:500]\n[00:02.00]b\n[00:01.00]a")
+
+    assert data.to_plain() == "a\nb"
+    assert data.to_normalized_text() == "[00:01.50]a\n[00:02.50]b"
+
+
+def test_to_text_default_forces_unsynced_tagging() -> None:
+    data = LRCData("line\nother")
+
+    assert data.to_text() == "[00:00.00]line\n[00:00.00]other"
+
+
+def test_str_keeps_raw_unsynced_while_to_text_adds_unsynced_tags() -> None:
+    data = LRCData("line\nother")
+
+    assert str(data) == "line\nother"
+    assert data.to_text() == "[00:00.00]line\n[00:00.00]other"
+
+
 def test_to_plain_duplicates_lines_for_multi_line_times() -> None:
    text = "\n".join(
        [
@@ -288,6 +365,40 @@ def test_word_sync_markup_only_changes_output_when_enabled() -> None:
    assert b.to_text(include_word_sync=True) == "[00:01.00]lyric"


+def test_str_preserves_word_sync_markup() -> None:
+    data = LRCData("[00:01.00]<00:00.50>lyric")
+
+    assert str(data) == "[00:01.00]<00:00.50>lyric"
+    assert data.to_text(include_word_sync=False) == "[00:01.00]lyric"
+
+
+def test_str_preserves_offset_tag_and_does_not_apply_it() -> None:
+    data = LRCData("[offset:500]\n[00:01.00]a")
+
+    assert str(data) == "[offset:500]\n[00:01.00]a"
+    assert data.to_normalized_text() == "[00:01.50]a"
+
+
+def test_str_preserves_doc_tag_order_and_duplicates_exactly() -> None:
+    data = LRCData("[ar:First]\n[ti:Song]\n[ar:Second]\n[00:01.00]line")
+
+    assert str(data) == "[ar:First]\n[ti:Song]\n[ar:Second]\n[00:01.00]line"
+
+
+def test_str_does_not_expand_or_sort_multi_time_lines() -> None:
+    data = LRCData("[00:03.00]c\n[00:02.00][00:01.00]x")
+
+    assert str(data) == "[00:03.00]c\n[00:02.00][00:01.00]x"
+    assert data.to_normalized_text() == "[00:01.00]x\n[00:02.00]x\n[00:03.00]c"
+
+
+def test_str_preserves_plain_text_lines_without_injecting_time_tags() -> None:
+    data = LRCData("plain line\n[ar:Artist]\nother line")
+
+    assert str(data) == "plain line\n[ar:Artist]\nother line"
+    assert data.to_text() == "[00:00.00]plain line\n[ar:Artist]\n[00:00.00]other line"
+
+
 def test_word_sync_line_with_empty_tail_keeps_word_tag_only_when_enabled() -> None:
    data = LRCData("[00:01.00]<00:02.00>")

@@ -295,10 +406,10 @@ def test_word_sync_line_with_empty_tail_keeps_word_tag_only_when_enabled() -> No
    assert data.to_text(include_word_sync=True) == "[00:01.00]<00:02.00>"


-def test_to_text_plain_true_matches_to_plain_output() -> None:
+def test_to_plain_outputs_sorted_plain_text() -> None:
    data = LRCData("[00:02.00]b\n[00:01.00]a")

-    assert data.to_text(plain=True) == data.to_plain()
+    assert data.to_plain() == "a\nb"


 def test_duplicate_doc_tag_key_last_value_wins_but_lines_are_kept() -> None:
@@ -153,7 +153,7 @@ wheels = [

 [[package]]
 name = "lrx-cli"
-version = "0.6.2"
+version = "0.6.3"
 source = { editable = "." }
 dependencies = [
    { name = "cyclopts" },