first commit

app/dat_parser.py

@@ -0,0 +1,128 @@
+"""Parser for clrmamepro (.dat) files.
+
+Handles the No-Intro / Redump text DAT format::
+
+    clrmamepro (
+        name "Sony - PlayStation Portable"
+        version "2021.10.24"
+    )
+    game (
+        name "WipEout Pure (Europe)"
+        region "Europe"
+        serial "ULES-00011"
+        rom ( name "WipEout Pure (Europe).iso" size 123 crc ABCD1234 md5 ... sha1 ... )
+    )
+
+``parse_dat_text`` returns ``(header, entries)``.  Each entry is a dict with
+scalar keys (``name``, ``description``, ``region``, ``serial``, ...) and
+repeated nested blocks exposed as lists under their block key (``rom``).
+"""
+import re
+from pathlib import Path
+
+_TOKEN_RE = re.compile(
+    r"""
+      (?P<ws>\s+)
+    | (?P<lparen>\()
+    | (?P<rparen>\))
+    | (?P<string>"(?:\\.|[^"\\])*")
+    | (?P<ident>[^"\s()]+)
+    """,
+    re.VERBOSE,
+)
+
+
+class _Token:
+    __slots__ = ("kind", "value")
+
+    def __init__(self, kind, value):
+        self.kind = kind
+        self.value = value
+
+
+def _tokenize(text):
+    tokens = []
+    for match in _TOKEN_RE.finditer(text):
+        kind = match.lastgroup
+        value = match.group()
+        if kind == "ws":
+            continue
+        if kind == "string":
+            value = value[1:-1].replace('\\"', '"').replace("\\\\", "\\")
+            kind = "string"
+        elif kind == "lparen":
+            kind = "lparen"
+            value = "("
+        elif kind == "rparen":
+            kind = "rparen"
+            value = ")"
+        else:
+            kind = "ident"
+        tokens.append(_Token(kind, value))
+    return tokens
+
+
+class _Parser:
+    def __init__(self, tokens):
+        self._tokens = tokens
+        self._pos = 0
+
+    def _peek(self):
+        return self._tokens[self._pos] if self._pos < len(self._tokens) else None
+
+    def _next(self):
+        token = self._peek()
+        self._pos += 1
+        return token
+
+    def parse_block(self, name=None):
+        if name is None:
+            name_token = self._next()
+            if name_token is None or name_token.kind != "ident":
+                raise ValueError(f"expected block name at token {self._pos}")
+            name = name_token.value
+        lparen = self._next()
+        if lparen is None or lparen.kind != "lparen":
+            raise ValueError(f"expected '(' after '{name}'")
+
+        block = {"_type": name}
+        while True:
+            token = self._peek()
+            if token is None:
+                raise ValueError(f"unexpected EOF inside block '{name}'")
+            if token.kind == "rparen":
+                self._next()
+                break
+            if token.kind != "ident":
+                raise ValueError(f"unexpected token '{token.value}' in block '{name}'")
+            key = self._next().value
+            nxt = self._peek()
+            if nxt is not None and nxt.kind == "lparen":
+                child = self.parse_block(name=key)
+                block.setdefault(key, []).append(child)
+            else:
+                value = self._next()
+                if value is None:
+                    raise ValueError(f"expected value for '{key}' in block '{name}'")
+                block[key] = value.value
+        return block
+
+
+def parse_dat_text(text):
+    tokens = _tokenize(text)
+    parser = _Parser(tokens)
+    header = None
+    entries = []
+    while parser._peek() is not None:
+        block = parser.parse_block()
+        kind = block.get("_type")
+        if kind == "clrmamepro":
+            header = block
+        elif kind in ("game", "machine"):
+            entries.append(block)
+        # ignore any other top-level blocks
+    return header, entries
+
+
+def parse_dat_file(path):
+    return parse_dat_text(Path(path).read_text(encoding="utf-8-sig"))