From e4d4ad574732aa63d99912f29f3fc7ff0401cf9f Mon Sep 17 00:00:00 2001
From: thefool0 <74161809+thefool0000@users.noreply.github.com>
Date: Tue, 31 Dec 2024 00:06:42 +0800
Subject: [PATCH] feat: support unified diff and git headers

Signed-off-by: thefool0 <74161809+thefool0000@users.noreply.github.com>
---
 src/Patche/utils/header.py | 54 ++++++++++++++++++++++++++
 src/Patche/utils/parse.py  | 78 ++++++++++++++++++++++++++++++++++++++
 2 files changed, 132 insertions(+)
 create mode 100644 src/Patche/utils/header.py

diff --git a/src/Patche/utils/header.py b/src/Patche/utils/header.py
new file mode 100644
index 0000000..2153a7e
--- /dev/null
+++ b/src/Patche/utils/header.py
@@ -0,0 +1,54 @@
+import re
+from typing import Iterator, Optional
+
+from Patche.model import Header
+
+HEADER_OLD = re.compile(r"^--- ([^\t\n]+)(?:\t([^\n]*)|)$")
+HEADER_NEW = re.compile(r"^\+\+\+ ([^\t\n]+)(?:\t([^\n]*)|)$")
+HUNK_START = re.compile(r"^@@ -(\d+)(?:,(\d+))? \+(\d+)(?:,(\d+))? @@")
+CHANGE_LINE = re.compile(r"^([- +\\])(.*)$")
+GIT_HEADER = re.compile(r"^diff --git a/(.*) b/(.*)$")
+
+
+def parse_header(lines: Iterator[str]) -> Optional[Header]:
+    """解析 diff 头部信息"""
+    first_line = next(lines, "")
+
+    # 尝试解析 git header
+    git_match = GIT_HEADER.match(first_line)
+    if git_match:
+        # 跳过 index 行
+        next(lines, "")
+        old_line = next(lines, "")
+        new_line = next(lines, "")
+
+        old_match = HEADER_OLD.match(old_line)
+        new_match = HEADER_NEW.match(new_line)
+
+        if old_match and new_match:
+            return Header(
+                index_path=git_match.group(1),
+                old_path=old_match.group(1),
+                old_version=old_match.group(2) if old_match.group(2) else None,
+                new_path=new_match.group(1),
+                new_version=new_match.group(2) if new_match.group(2) else None,
+            )
+        return None
+
+    # 回退迭代器以处理统一 diff 格式
+    lines = iter([first_line] + list(lines))
+    old_line = next(lines, "")
+    new_line = next(lines, "")
+
+    old_match = HEADER_OLD.match(old_line)
+    new_match = HEADER_NEW.match(new_line)
+
+    if old_match and new_match:
+        return Header(
+            index_path=None,
+            old_path=old_match.group(1),
+            old_version=old_match.group(2) if old_match.group(2) else None,
+            new_path=new_match.group(1),
+            new_version=new_match.group(2) if new_match.group(2) else None,
+        )
+    return None
diff --git a/src/Patche/utils/parse.py b/src/Patche/utils/parse.py
index 8a42b43..20db8b9 100644
--- a/src/Patche/utils/parse.py
+++ b/src/Patche/utils/parse.py
@@ -1,10 +1,12 @@
 import re
+from typing import List, Optional
 
 from whatthepatch_pydantic import parse_patch as wtp_parse_patch
 from whatthepatch_pydantic.model import Diff as WTPDiff
 
 from Patche.config import settings
 from Patche.model import Change, Diff, Hunk, Patch
+from Patche.utils.header import CHANGE_LINE, HEADER_OLD, HUNK_START, parse_header
 
 git_diffcmd_header = re.compile("^diff --git a/(.+) b/(.+)$")
 unified_diff_header = re.compile("^---\s{1}")
@@ -94,6 +96,82 @@ def wtp_diff_to_diff(wtp_diff: WTPDiff) -> Diff:
     )
 
 
+def parse_unified_diff(text: str) -> Optional[List[Diff]]:
+    """解析 unified diff 格式的补丁"""
+    lines = iter(text.splitlines())
+    diffs: List[Diff] = []
+
+    while True:
+        try:
+            header = parse_header(lines)
+            if not header:
+                break
+
+            changes: List[Change] = []
+            hunk_index = 0
+
+            for line in lines:
+                # 检查是否是新的 diff 块开始
+                if HEADER_OLD.match(line):
+                    lines = iter([line] + list(lines))
+                    break
+
+                # 解析 hunk 头
+                hunk_match = HUNK_START.match(line)
+                if hunk_match:
+                    old_start = int(hunk_match.group(1))
+                    old_count = int(hunk_match.group(2) or "1")
+                    new_start = int(hunk_match.group(3))
+                    new_count = int(hunk_match.group(4) or "1")
+
+                    old_current = old_start
+                    new_current = new_start
+                    continue
+
+                # 解析变更行
+                change_match = CHANGE_LINE.match(line)
+                if change_match:
+                    change_type = change_match.group(1)
+                    content = change_match.group(2)
+
+                    if change_type == " ":
+                        # 上下文行
+                        changes.append(
+                            Change(
+                                old=old_current,
+                                new=new_current,
+                                line=content,
+                                hunk=hunk_index,
+                            )
+                        )
+                        old_current += 1
+                        new_current += 1
+                    elif change_type == "-":
+                        # 删除行
+                        changes.append(
+                            Change(
+                                old=old_current, new=None, line=content, hunk=hunk_index
+                            )
+                        )
+                        old_current += 1
+                    elif change_type == "+":
+                        # 新增行
+                        changes.append(
+                            Change(
+                                old=None, new=new_current, line=content, hunk=hunk_index
+                            )
+                        )
+                        new_current += 1
+
+            if header and changes:
+                diffs.append(Diff(header=header, changes=changes, text=text))
+
+        except StopIteration:
+            break
+
+    return diffs if diffs else None
+
+
 def parse_patch(text: str) -> Patch:
     """
     Parse a patch file