fix(#1433): allow leading underscore in attribute names

The attribute_name parser in declare.py was pp.Word over [a-z] init chars
and [a-z0-9_] body chars, rejecting any name starting with `_`. But the
framework already treats names starting with `_` as hidden attributes
(Heading.attributes filters by is_hidden = name.startswith("_")), and
internal hidden columns like _job_start_time, _job_duration, _job_version,
and _singleton are injected programmatically, bypassing the parser.

User-defined hidden attributes — documented at
docs.datajoint.com/reference/specs/table-declaration/#34-hidden-attributes —
hit the parser and failed with a cryptic pyparsing ParseException.

Allow `_` in the init-chars set so user code like

    _params_hash: varchar(32)
    unique index (tool, _params_hash)

declares cleanly. Names starting with a digit are still rejected.

Fixes #1433

Author: dimitri-yatsenko

src/datajoint/declare.py

@@ -147,7 +147,7 @@ def build_attribute_parser() -> pp.ParserElement:
     """
     quoted = pp.QuotedString('"') ^ pp.QuotedString("'")
     colon = pp.Literal(":").suppress()
-    attribute_name = pp.Word(pp.srange("[a-z]"), pp.srange("[a-z0-9_]")).set_results_name("name")
+    attribute_name = pp.Word(pp.srange("[_a-z]"), pp.srange("[a-z0-9_]")).set_results_name("name")
     data_type = (
         pp.Combine(pp.Word(pp.alphas) + pp.SkipTo("#", ignore=quoted))
         ^ pp.QuotedString("<", end_quote_char=">", unquote_results=False)

tests/unit/test_declare_hidden_attribute.py

@@ -0,0 +1,44 @@
+"""Unit tests for hidden attribute names (leading underscore) in table declarations.
+
+Regression coverage for issue #1433: the declaration parser previously rejected
+attribute names starting with ``_``, even though hidden-attribute semantics
+(``is_hidden = name.startswith("_")``) were already implemented in ``Heading``.
+"""
+
+import pytest
+
+from datajoint.declare import attribute_parser
+
+
+@pytest.mark.parametrize(
+    "line",
+    [
+        "_hidden: bool",
+        "_params_hash: varchar(32)",
+        "_job_start_time=null: datetime(3)",
+        "_a: int",
+    ],
+)
+def test_parser_accepts_leading_underscore(line):
+    match = attribute_parser.parse_string(line + "#", parse_all=True)
+    assert match["name"].startswith("_")
+
+
+def test_parser_still_accepts_plain_names():
+    match = attribute_parser.parse_string("name: varchar(40)#", parse_all=True)
+    assert match["name"] == "name"
+
+
+def test_parser_rejects_digit_start():
+    """Numeric leading char remains invalid (preserved behavior)."""
+    import pyparsing as pp
+
+    with pytest.raises(pp.ParseException):
+        attribute_parser.parse_string("1bad: int#", parse_all=True)
+
+
+def test_parser_extracts_hidden_name_for_is_hidden_dispatch():
+    """The parsed name is what Heading uses to set is_hidden via name.startswith('_')."""
+    match = attribute_parser.parse_string("_secret: int#", parse_all=True)
+    name = match["name"]
+    assert name.startswith("_")