mirror of
https://github.com/open-metadata/OpenMetadata
synced 2026-05-24 09:39:11 +00:00
fac395357b
* Add SQLGlot parser support Implement SQLGlot analyzer with SQLFluff/SQLParse fallback, query hash tracking, and optimized masking. * Add query_hash to all lineage parsing logs for better tracking * Better logging of parser logs * Cache db service lookups to reduce repeated searches * sqlglot query masking fallback to sqlparse and better logging to track * Consistent logs for query parsing with all useful information * Add query masking tests for all parsers * Remove duplicate query masking tests * Add specific dialect sql tests and helper methods to test/compare all parsers results * py_format * Add tests for large set of complex query patterns to validate all parsers * Add memory limits on lineage query parsers with default 100mb limit * Better memory limit handling and more tests * Remove query parsing issue summary since it's killing the workflow when list is too large * py_format * Add e2e lineage tests for oracle db and fix oracle query lineage filters * py_format * Remove SqlGlot parsing from query masker * Add __init__.py to query test packages * Better logs to track get lineage method * Disable memory limits for now as they are performance overhead * Update sql file path for e2e oracle db lineage tests * TEMP: Add local rc build of latest sqllineage with sqlglot support for checks * Revert search_cache name change in sql_lineage.py * Handle tests hanging with timeouts caused by graph checks or heavy query parsing * Complex query test formatting * py_format * Handle complex query tests with appropriate flags #01 * Handle complex query tests with appropriate flags #02 * Handle complex query tests with appropriate flags #03 * Handle complex query tests with appropriate flags #04 (final) * Update query lineage test helper for better troubleshooting * Add dialect specific query masking tests and skip sqlglot failures for now to evaluate later * Fix or skip other failing test related to sqlglot changes * py_format * Reduce sleep between proc calls for faster tests * Remove default test diff limit and skip graph check that timeout in ci check * Clear the topology runner cache in test to have cleaned state * Skip flaky graph check timeouts on test * Handle no parser in mask_query and log every message as debug to not pollute logs * Update parser logs to debug for less verbosity on default log level * Remove TEMP collate-sqllineage whl added for test since 2.0.0 is out * Log maximum 10 failures in workflow summary to not overload ingestion * Cleanup oracle db image after lineage e2e tests * Remove query parsing failures tracking from sql lineage process - Since these failures are already logged at the LineageParser, this doesn't need to be tracked via query_parsing_failures
4 lines
96 B
Python
4 lines
96 B
Python
from unittest import TestCase
|
|
|
|
# Make all diffs visible in test outputs
|
|
TestCase.maxDiff = None
|