Added a configuration flag to set the maximum number of tokens in validation prompts

tests/test_new_config.py

@@ -12,6 +12,7 @@ from audita.core.config import (
     DEFAULT_MIN_SECTION_TOKENS,
     DEFAULT_NORMALIZE_MAX_SEGMENT_GAP,
     DEFAULT_SPOKEN_WORD_CONFIDENCE_THRESHOLD,
+    DEFAULT_VALIDATION_MAX_PROMPT_TOKENS,
     DEFAULT_WORK_DIR_RETENTION,
 )
 from audita.core.errors import AuditaConfigError
@@ -30,6 +31,7 @@ def test_default_config_allows_missing_api_key():
     assert config.validation_model is None
     assert config.validation_base_url is None
     assert config.validation_max_retries is None
+    assert config.validation_max_prompt_tokens == DEFAULT_VALIDATION_MAX_PROMPT_TOKENS
     assert config.target_sections is None
     assert config.max_section_tokens == DEFAULT_MAX_SECTION_TOKENS
     assert config.min_section_tokens == DEFAULT_MIN_SECTION_TOKENS
@@ -138,6 +140,21 @@ def test_validation_max_retries_env_is_parsed():
     assert config.validation_max_retries == 7
 
 
+def test_validation_max_prompt_tokens_cli_override_takes_precedence():
+    config = AuditaConfig.from_sources(
+        env={"AUDITA_VALIDATION_MAX_PROMPT_TOKENS": "1024"},
+        overrides=ConfigOverrides(validation_max_prompt_tokens=4096),
+    )
+
+    assert config.validation_max_prompt_tokens == 4096
+
+
+def test_validation_max_prompt_tokens_env_is_parsed():
+    config = AuditaConfig.from_sources(env={"AUDITA_VALIDATION_MAX_PROMPT_TOKENS": "1024"})
+
+    assert config.validation_max_prompt_tokens == 1024
+
+
 def test_target_sections_cli_override_takes_precedence():
     config = AuditaConfig.from_sources(
         env={"AUDITA_TARGET_SECTIONS": "2"},
@@ -296,6 +313,7 @@ def test_report_dict_includes_effective_validation_llm_config():
     report = config.to_report_dict()
 
     assert report["validation_model"] == "validation-model"
+    assert report["validation_max_prompt_tokens"] == DEFAULT_VALIDATION_MAX_PROMPT_TOKENS
     assert report["effective_validation_llm"]["api_key_configured"] is True
     assert report["effective_validation_llm"]["model"] == "validation-model"
 
@@ -379,6 +397,12 @@ def test_invalid_validation_max_retries_is_rejected(value):
         AuditaConfig.from_sources(env={"AUDITA_VALIDATION_MAX_RETRIES": value})
 
 
+@pytest.mark.parametrize("value", ["0", "-1", "many"])
+def test_invalid_validation_max_prompt_tokens_is_rejected(value):
+    with pytest.raises(AuditaConfigError, match="AUDITA_VALIDATION_MAX_PROMPT_TOKENS"):
+        AuditaConfig.from_sources(env={"AUDITA_VALIDATION_MAX_PROMPT_TOKENS": value})
+
+
 @pytest.mark.parametrize("value", ["0", "-1", "many"])
 def test_invalid_target_sections_is_rejected(value):
     with pytest.raises(AuditaConfigError, match="AUDITA_TARGET_SECTIONS"):