Build Flask WAF log converter app

app/services/parser.py

@@ -0,0 +1,47 @@
+import shlex
+from collections import OrderedDict
+from io import BufferedIOBase, TextIOBase
+
+
+class LogParseError(ValueError):
+    """Raised when the uploaded log file cannot be parsed."""
+
+
+def parse_log_file(stream: BufferedIOBase | TextIOBase) -> tuple[list[dict[str, str]], list[str]]:
+    """Parse a UTF-8 log file where each line contains shell-like key/value tokens."""
+    raw_bytes = stream.read()
+    if isinstance(raw_bytes, str):
+        content = raw_bytes
+    else:
+        content = raw_bytes.decode("utf-8")
+
+    records: list[dict[str, str]] = []
+    seen_keys: OrderedDict[str, None] = OrderedDict()
+
+    for line_number, raw_line in enumerate(content.splitlines(), start=1):
+        line = raw_line.strip()
+        if not line:
+            continue
+
+        try:
+            tokens = shlex.split(line, posix=True)
+        except ValueError as exc:
+            raise LogParseError(f"Line {line_number}: invalid shell-style quoting.") from exc
+
+        record: dict[str, str] = {}
+        for token in tokens:
+            if "=" not in token:
+                raise LogParseError(
+                    f"Line {line_number}: token '{token}' is missing '='."
+                )
+
+            key, value = token.split("=", 1)
+            if not key:
+                raise LogParseError(f"Line {line_number}: empty key is not allowed.")
+
+            record[key] = value
+            seen_keys.setdefault(key, None)
+
+        records.append(record)
+
+    return records, list(seen_keys.keys())