v1: executing-plans 模式生成，54 文件 1320 行 Python

comparator/normalizer.py

@@ -0,0 +1,89 @@
+from dataclasses import dataclass
+
+
+EBCDIC_TO_ASCII = {
+    0xC1: 'A', 0xC2: 'B', 0xC3: 'C', 0xC4: 'D', 0xC5: 'E',
+    0xC6: 'F', 0xC7: 'G', 0xC8: 'H', 0xC9: 'I', 0xD1: 'J',
+    0xD2: 'K', 0xD3: 'L', 0xD4: 'M', 0xD5: 'N', 0xD6: 'O',
+    0xD7: 'P', 0xD8: 'Q', 0xD9: 'R', 0xE2: 'S', 0xE3: 'T',
+    0xE4: 'U', 0xE5: 'V', 0xE6: 'W', 0xE7: 'X', 0xE8: 'Y',
+    0xE9: 'Z', 0xF0: '0', 0xF1: '1', 0xF2: '2', 0xF3: '3',
+    0xF4: '4', 0xF5: '5', 0xF6: '6', 0xF7: '7', 0xF8: '8',
+    0xF9: '9', 0x40: ' ', 0x4B: '.', 0x6B: ',', 0x5A: '!',
+}
+
+
+@dataclass
+class CobolIRField:
+    raw_hex: str
+    decoded_value: str
+    encoding: str
+    field_type: str
+    length: int
+    scale: int
+    signed: bool
+
+
+@dataclass
+class JavaIRField:
+    raw_value: str
+    decoded_value: str
+    field_type: str
+    nullable: bool
+
+
+@dataclass
+class IRRecord:
+    field_name: str
+    cobol: CobolIRField | None = None
+    java: JavaIRField | None = None
+
+
+class Normalizer:
+    def normalize_encoding(self, raw: bytes, encoding: str) -> str:
+        if encoding == "EBCDIC":
+            return self._ebcdic_to_ascii(raw)
+        return raw.decode("ascii", errors="replace")
+
+    def normalize_comp3(self, raw: bytes) -> str:
+        if not raw:
+            return "0"
+        nibbles = []
+        for b in raw:
+            nibbles.append((b >> 4) & 0x0F)
+            nibbles.append(b & 0x0F)
+        sign = nibbles.pop()
+        value = 0
+        for n in nibbles:
+            value = value * 10 + n
+        if sign in (0x0D, 0x0B):
+            value = -value
+        return str(value)
+
+    def normalize_date(self, date_str: str) -> str:
+        s = date_str.strip()
+        if len(s) == 8 and s.isdigit():
+            return f"{s[0:4]}-{s[4:6]}-{s[6:8]}"
+        return s
+
+    def to_ir_record(self, field_name, raw_hex, decoded_value,
+                     encoding, field_type, length=0, scale=0, signed=False) -> IRRecord:
+        return IRRecord(
+            field_name=field_name,
+            cobol=CobolIRField(
+                raw_hex=raw_hex, decoded_value=decoded_value,
+                encoding=encoding, field_type=field_type,
+                length=length, scale=scale, signed=signed))
+
+    def to_null_ir(self, field_name, side="java") -> IRRecord:
+        if side == "java":
+            return IRRecord(field_name=field_name,
+                            cobol=None, java=JavaIRField(raw_value="", decoded_value="", field_type="null", nullable=True))
+        return IRRecord(field_name=field_name,
+                        cobol=None, java=JavaIRField(raw_value="", decoded_value="", field_type="null", nullable=True))
+
+    def _ebcdic_to_ascii(self, raw: bytes) -> str:
+        result = []
+        for b in raw:
+            result.append(EBCDIC_TO_ASCII.get(b, chr(b) if 32 <= b < 127 else '?'))
+        return ''.join(result)