Added new deterministic validators to catch early failures

tests/test_llm_validators.py

@@ -5,7 +5,12 @@ from audita.core.config import AuditaConfig
 from audita.core.errors import AuditaLLMError
 from audita.core.schemas import parse_glossary_yaml, parse_transcript_json
 from audita.framework.models import CorrectionProposal, ModuleRunSpec
-from audita.validators import GrammarOnlyValidator, NonEmptySegmentValidator
+from audita.validators import (
+    GrammarOnlyValidator,
+    IdenticalTextValidator,
+    NonEmptySegmentValidator,
+    OriginalTextPresentValidator,
+)
 from audita.validators.base import ValidationContext
 from audita.validators.llm import MeaningReversalValidator, SpokenFormPlausibilityValidator, SpokenWordValidator
 from audita.validators.prompts import (
@@ -550,6 +555,96 @@ def test_non_empty_segment_validator_allows_punctuation_only_and_unpreviewable_p
     ]
 
 
+def test_identical_text_validator_rejects_exact_noops_only(tmp_path):
+    transcript = parse_transcript_json(
+        """
+        [
+          {"id": 1, "speaker": "A", "start": 0.0, "end": 1.0, "text": "hello"},
+          {"id": 2, "speaker": "A", "start": 1.0, "end": 2.0, "text": "hello"}
+        ]
+        """
+    )
+    proposals = [
+        CorrectionProposal(
+            proposal_index=0,
+            module_instance="grammar",
+            module_key="grammar",
+            id=1,
+            original_text="hello",
+            corrected_text="hello",
+            confidence=0.95,
+        ),
+        CorrectionProposal(
+            proposal_index=1,
+            module_instance="grammar",
+            module_key="grammar",
+            id=2,
+            original_text="hello",
+            corrected_text="Hello",
+            confidence=0.95,
+        ),
+    ]
+
+    result = IdenticalTextValidator("identical_text_guard").validate(
+        _context(proposals=proposals, transcript=transcript, llm_client=None, tmp_path=tmp_path)
+    )
+
+    assert [(decision.proposal_index, decision.approved, decision.reason) for decision in result.decisions] == [
+        (0, False, "proposal original_text and corrected_text are identical"),
+        (1, True, None),
+    ]
+
+
+def test_original_text_present_validator_rejects_missing_spans_and_allows_present_or_unknown_segments(tmp_path):
+    transcript = parse_transcript_json(
+        """
+        [
+          {"id": 1, "speaker": "A", "start": 0.0, "end": 1.0, "text": "hello hello"},
+          {"id": 2, "speaker": "A", "start": 1.0, "end": 2.0, "text": "goodbye"}
+        ]
+        """
+    )
+    proposals = [
+        CorrectionProposal(
+            proposal_index=0,
+            module_instance="homophones",
+            module_key="homophones",
+            id=1,
+            original_text="hello",
+            corrected_text="hi",
+            confidence=0.95,
+        ),
+        CorrectionProposal(
+            proposal_index=1,
+            module_instance="homophones",
+            module_key="homophones",
+            id=2,
+            original_text="rank",
+            corrected_text="Hrank",
+            confidence=0.95,
+        ),
+        CorrectionProposal(
+            proposal_index=2,
+            module_instance="homophones",
+            module_key="homophones",
+            id=99,
+            original_text="missing",
+            corrected_text="present",
+            confidence=0.95,
+        ),
+    ]
+
+    result = OriginalTextPresentValidator("original_text_present_guard").validate(
+        _context(proposals=proposals, transcript=transcript, llm_client=None, tmp_path=tmp_path)
+    )
+
+    assert [(decision.proposal_index, decision.approved, decision.reason) for decision in result.decisions] == [
+        (0, True, None),
+        (1, False, "proposal original_text does not match segment text"),
+        (2, True, None),
+    ]
+
+
 @pytest.mark.parametrize(
     ("validator", "payload", "message_fragment"),
     [