feat(rules)!: check audit, test and hook SQL and enforce SQL quality rules

crates/sqlbuild-rules-native/src/configuration/_helpers/loading.rs

@@ -99,7 +99,9 @@ fn validate_raw(value: &toml::Value) -> Result<(), String> {
 }
 
 pub(crate) fn validate(config: &RulesConfig) -> Result<(), String> {
-    const THRESHOLDS: [&str; 10] = [
+    const THRESHOLDS: [&str; 12] = [
+        "max_literal_length",
+        "max_ranking_order_by",
         "min_audits_per_model",
         "min_tests_per_model",
         "min_custom_rule_test_cases",
@@ -160,6 +162,14 @@ pub(crate) fn validate(config: &RulesConfig) -> Result<(), String> {
     {
         return Err(format!("unknown rule thresholds: {name}"));
     }
+    if let Some(name) = ["max_literal_length", "max_ranking_order_by"]
+        .into_iter()
+        .find(|name| config.thresholds.get(*name) == Some(&0))
+    {
+        return Err(format!(
+            "rules.thresholds.{name} must be a positive integer"
+        ));
+    }
     for entry in &config.threshold_overrides {
         if entry.paths.is_empty() || entry.thresholds.is_empty() || entry.reason.trim().is_empty() {
             return Err("rules.threshold_overrides requires paths, thresholds, and reason".into());

crates/sqlbuild-rules-native/src/configuration/main/mod.rs

@@ -1,2 +1,4 @@
 pub(crate) mod load;
+pub(crate) mod setting_help;
+pub(crate) mod toml_string_array;
 pub(crate) mod validate;

crates/sqlbuild-rules-native/src/configuration/main/setting_help.rs

@@ -0,0 +1,10 @@
+//! Remediation text that shows the exact `sqlbuild_project.toml` setting to write.
+
+use crate::constants::{PROJECT_CONFIG_FILE, SETTING_SNIPPET_INDENT};
+
+/// `<purpose>, set this in sqlbuild_project.toml:` followed by the section and `key = value`.
+pub(crate) fn setting_help(purpose: &str, section: &str, key: &str, value: &str) -> String {
+    format!(
+        "{purpose}, set this in {PROJECT_CONFIG_FILE}:\n{SETTING_SNIPPET_INDENT}[{section}]\n{SETTING_SNIPPET_INDENT}{key} = {value}"
+    )
+}

crates/sqlbuild-rules-native/src/configuration/main/toml_string_array.rs

@@ -0,0 +1,10 @@
+//! TOML array rendering for remediation snippets.
+
+/// A TOML array of strings, in the given order.
+pub(crate) fn toml_string_array<'a>(values: impl IntoIterator<Item = &'a str>) -> String {
+    let quoted: Vec<String> = values
+        .into_iter()
+        .map(|value| format!("{value:?}"))
+        .collect();
+    format!("[{}]", quoted.join(", "))
+}

crates/sqlbuild-rules-native/src/constants.rs

@@ -22,6 +22,7 @@ pub(crate) const ENFORCED_CONTRACT: &str = "enforced";
 pub(crate) const SELECT_STAR_MODEL_RULE_CODE: &str = "SQBRMODEL102";
 pub(crate) const BOOLEAN_TYPE: &str = "BOOLEAN";
 pub(crate) const TIMESTAMP_TYPE: &str = "TIMESTAMP";
+pub(crate) const DATETIME_TYPE: &str = "DATETIME";
 pub(crate) const DATE_TYPE: &str = "DATE";
 pub(crate) const NEGATION_OPERATOR: &str = "-";
 pub(crate) const DECLARATION_DOMAIN_COMPONENTS: usize = 3;
@@ -53,3 +54,5 @@ pub(crate) const SQL_TEST_ACTUAL_CTE: &str = "__actual";
 pub(crate) const SQL_TEST_EXPECTED_CTE: &str = "__expected";
 pub(crate) const SQL_TEST_ACTUAL_CTE_PREFIX: &str = "__actual__";
 pub(crate) const QUOTED_IDENTIFIER_DELIMITER_BYTES: usize = 2;
+pub(crate) const PROJECT_CONFIG_FILE: &str = "sqlbuild_project.toml";
+pub(crate) const SETTING_SNIPPET_INDENT: &str = "            ";

crates/sqlbuild-rules-native/src/lib.rs

@@ -13,5 +13,6 @@ mod scope_metadata;
 mod semantic_usage;
 mod semantic_validation;
 mod sql_lint;
+mod sql_quality;
 mod sql_scan;
 mod sql_tokens;

crates/sqlbuild-rules-native/src/rules/_helpers/domain_layout.rs

@@ -1,3 +1,4 @@
+use crate::configuration::main::setting_help::setting_help;
 use crate::models::{DeclarationKind, Fault, Model, RuleMetadata, ScopeDeclaration, ScopeKind};
 use crate::rules::_helpers::evaluation::path_fault;
 use crate::rules::models::ProjectEvaluationRequest;
@@ -187,7 +188,12 @@ fn evaluate_models(evaluation: &ProjectEvaluationRequest<'_>) -> Vec<Fault> {
                             "model has ambiguous domain-root and level candidates: {}",
                             candidates.join(", ")
                         ),
-                        "Set rules.layout.domain_roots to the intended normalized roots.".into(),
+                        setting_help(
+                            "List the intended normalized domain roots",
+                            "rules.layout",
+                            "domain_roots",
+                            r#"["<domain root>"]"#,
+                        ),
                     ));
                 }
                 continue;
@@ -207,10 +213,15 @@ fn evaluate_models(evaluation: &ProjectEvaluationRequest<'_>) -> Vec<Fault> {
                     max_depth
                 ),
                 format!(
-                    "Flatten this path beneath {}/{}, promote part of it into the domain root, or set rules.thresholds.max_subdomain_depth to at least {}.",
+                    "Flatten this path beneath {}/{}, or promote part of it into the domain root. {}",
                     resolved.domain,
                     resolved.level,
-                    resolved.owners.len()
+                    setting_help(
+                        &format!("To allow this depth (the current value is {max_depth})"),
+                        "rules.thresholds",
+                        "max_subdomain_depth",
+                        &resolved.owners.len().to_string(),
+                    )
                 ),
             ));
         }
@@ -513,12 +524,20 @@ fn inspect_container(
                 format!(
                     "declaration role {:?} has bucket depth {}; configured maximum is {}",
                     inspection.root,
-                    entry.buckets.len()
-                    ,inspection.max_depth
+                    entry.buckets.len(),
+                    inspection.max_depth
                 ),
                 format!(
-                    "Flatten this bucket path or set rules.thresholds.max_role_container_depth to at least {}.",
-                    entry.buckets.len()
+                    "Flatten this bucket path. {}",
+                    setting_help(
+                        &format!(
+                            "To allow this depth (the current value is {})",
+                            inspection.max_depth
+                        ),
+                        "rules.thresholds",
+                        "max_role_container_depth",
+                        &entry.buckets.len().to_string(),
+                    )
                 ),
             ));
         }

crates/sqlbuild-rules-native/src/rules/_helpers/evaluation.rs

@@ -1,3 +1,5 @@
+use crate::configuration::main::setting_help::setting_help;
+use crate::configuration::main::toml_string_array::toml_string_array;
 use crate::constants::{
     DECLARATION_DOMAIN_COMPONENTS, ENFORCED_CONTRACT, NEGATION_OPERATOR, REFERENCE_KIND,
     SOURCE_REFERENCE_KIND, VIEW_MATERIALIZATION,
@@ -1102,7 +1104,22 @@ fn name_grammar(
                     "model {:?} uses unknown domain {:?}",
                     parsed.model.name, parts.domain
                 ),
-                Some("Rename the model into a configured rules domain, or add this domain to rules.domains when it is an intentional project owner.".into()),
+                Some(format!(
+                    "Rename the model into a configured rules domain. The current value is [rules] domains = {}. {}",
+                    toml_string_array(config.domains.iter().map(String::as_str)),
+                    setting_help(
+                        "If this domain is an intentional project owner, add it",
+                        "rules",
+                        "domains",
+                        &toml_string_array(
+                            config
+                                .domains
+                                .iter()
+                                .map(String::as_str)
+                                .chain([parts.domain.as_str()]),
+                        ),
+                    )
+                )),
             ));
         }
         return;
@@ -1208,7 +1225,22 @@ fn source_token_rule(
                     "model {:?} uses unapproved source token {:?}",
                     parsed.model.name, token
                 ),
-                Some("Rename the source suffix to a token listed in rules.approved_source_tokens at this model path.".into()),
+                Some(format!(
+                    "Rename the source suffix to an approved token. The current value is [rules] approved_source_tokens = {}. {}",
+                    toml_string_array(config.approved_source_tokens.iter().map(String::as_str)),
+                    setting_help(
+                        "To approve this token",
+                        "rules",
+                        "approved_source_tokens",
+                        &toml_string_array(
+                            config
+                                .approved_source_tokens
+                                .iter()
+                                .map(String::as_str)
+                                .chain([token.as_str()]),
+                        ),
+                    )
+                )),
             ));
     }
 }

crates/sqlbuild-rules-native/src/semantic_validation/_helpers/diagnostics.rs

@@ -2,6 +2,7 @@
 
 use crate::semantic_validation::_helpers::catalog::diagnostic_row;
 use crate::semantic_validation::_helpers::function_names;
+use crate::semantic_validation::_helpers::type_locations::locate_type_findings;
 use crate::semantic_validation::models::FunctionProbes;
 use crate::semantic_validation::types::DiagnosticRow;
 use polyglot_sql::expressions::Select;
@@ -153,8 +154,10 @@ pub(crate) fn map_diagnostics(
         dialect,
         statements: None,
     };
-    let mut errors: Vec<ValidationError> = Vec::with_capacity(result.errors.len());
-    for error in result.errors {
+    let mut native = result.errors;
+    locate_type_findings(sql, dialect, &mut native)?;
+    let mut errors: Vec<ValidationError> = Vec::with_capacity(native.len());
+    for error in native {
         if let Some(error) = context.map(error)? {
             errors.push(error);
         }

crates/sqlbuild-rules-native/src/semantic_validation/_helpers/mod.rs

@@ -4,3 +4,4 @@ pub(crate) mod function_names;
 pub(crate) mod identifiers;
 pub(crate) mod normalization;
 pub(crate) mod positions;
+pub(crate) mod type_locations;