You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 3a58d84
Browse filesBrowse the repository at this point in the historyBrowse files
fix: pass warehouse dialect to altimate-core tools instead of hardcoding snowflake
Add optional `dialect` parameter (default: "snowflake") to all 8
altimate-core tools that were missing it:
- altimate-core-validate, fix, correct, semantics,
equivalence, policy, check, and impact-analysis
Each tool now:
1. Accepts `dialect` via its zod parameter schema
2. Passes it through to the Dispatcher call
3. Includes it in telemetry metadata
Follows the pattern established in sql-analyze.ts, sql-optimize.ts,
and schema-diff.ts. Type interfaces updated in types.ts.
Tests updated to pass dialect explicitly.
FixesAltimateAI#455
Co-Authored-By: Vijay Yadav <[email protected]>
"Run full analysis pipeline: validate + lint + safety scan + PII check. Single call for comprehensive SQL analysis. Provide schema_context or schema_path for accurate table/column resolution.",
"Iteratively correct SQL using a propose-verify-refine loop. More thorough than fix — applies multiple correction rounds to produce valid SQL. Provide schema_context or schema_path for accurate table/column resolution.",
"Auto-fix SQL errors using fuzzy matching and iterative re-validation. Corrects syntax errors, typos, and schema reference issues. IMPORTANT: Provide schema_context or schema_path — without schema, table/column references cannot be resolved or fixed.",
"Check SQL against YAML-based governance policy guardrails. Validates compliance with custom rules like allowed tables, forbidden operations, and data access restrictions. Provide schema_context or schema_path for accurate table/column resolution.",
9
9
parameters: z.object({
10
10
sql: z.string().describe("SQL query to check against policy"),
"Run semantic validation rules against SQL. Detects logical issues like cartesian products, wrong JOIN conditions, NULL misuse, and type mismatches that syntax checking alone misses. Provide schema_context or schema_path for accurate table/column resolution.",
9
9
parameters: z.object({
10
10
sql: z.string().describe("SQL query to validate semantically"),
"Validate SQL syntax and schema references. Checks if tables/columns exist in the schema and if SQL is valid for the target dialect. IMPORTANT: Provide schema_context or schema_path — without schema, all table/column references will report as 'not found'.",
9
9
parameters: z.object({
10
10
sql: z.string().describe("SQL query to validate"),
0 commit comments