changes

ntanwar-sumo · ntanwar-sumo · commit 95e4d8534a66 · 2025-06-25T16:49:12.000+05:30
diff --git a/docs/search/search-query-language/search-operators/isvalidip.md b/docs/search/search-query-language/search-operators/isvalidip.md
@@ -54,7 +54,7 @@ The following returns `true`:
 | isValidIP("10.255.255.255") as isIP
 ```
 ```sql
-| isValidIP("20.255.255.255") as isIP
+| isValidIP("30.255.255.255") as isIP
 ```
 
 The following returns `true`:
diff --git a/docs/search/search-query-language/search-operators/where.md b/docs/search/search-query-language/search-operators/where.md
@@ -17,7 +17,7 @@ For example, using `where` with the boolean operator [`isValidIP`](/docs/searc
     _collector="ABC1" | where type="web"
     ```
     ```sql
-    _collector="ABC5" | where type="web"
+    _collector="ABC6" | where type="web"
     ```   
 * Filters as false and will not return results:
     ```sql
diff --git a/scripts/validate_queries.py b/scripts/validate_queries.py
@@ -1,166 +1,69 @@
 #!/usr/bin/env python3
-"""
-Validates SumoLogic queries in Markdown files during PRs
-- Finds all SQL blocks in changed Markdown files
-- Executes each query against SumoLogic API
-- Fails PR if any query returns no results
-"""
-
 import re
 import sys
 import os
 import json
-import time
+import requests
 from pathlib import Path
 from datetime import datetime, timedelta
 
-# SumoLogic API Client (embedded for simplicity)
-class SumoLogicClient:
-    def __init__(self):
-        self.api_url = os.getenv('SUMO_LOGIC_ENDPOINT', 'https://long-api.sumologic.com/api/v1')
-        self.session = requests.Session()
-        self.session.auth = (
-            os.getenv('SUMO_LOGIC_ACCESS_ID'),
-            os.getenv('SUMO_LOGIC_ACCESS_KEY')
-        )
-        self.session.headers.update({'Content-Type': 'application/json'})
-
-    def test_query(self, query):
-        """Execute query and verify it returns results"""
-        print(f"⌛ Executing query (first 50 chars): {query[:50]}...")
-
-        job_id = self._create_search_job(query)
-        status = self._wait_for_job(job_id)
-
-        if status != "DONE GATHERING RESULTS":
-            raise Exception(f"Query failed with status: {status}")
-
-        return self._check_results(job_id)
-
-    def _create_search_job(self, query):
-        """Start a search job with 1-hour time window"""
-        time_range = {
-            'from': (datetime.utcnow() - timedelta(hours=1)).isoformat() + 'Z',
-            'to': datetime.utcnow().isoformat() + 'Z',
-            'timeZone': 'UTC'
-        }
-
-        response = self.session.post(
-            f"{self.api_url}/search/jobs",
-            json={'query': query, **time_range}
-        )
-        response.raise_for_status()
-        return response.json()['id']
-
-    def _wait_for_job(self, job_id, timeout=60):
-        """Wait for job completion with progress updates"""
-        for i in range(timeout):
-            response = self.session.get(f"{self.api_url}/search/jobs/{job_id}")
-            response.raise_for_status()
-            status = response.json()['state']
-
-            if i % 5 == 0:  # Print progress every 5 seconds
-                print(f"⏳ Query status: {status} ({i}/{timeout}s)")
+def get_repo_root():
+    """Get absolute path to repository root"""
+    github_workspace = os.getenv('GITHUB_WORKSPACE')
+    if github_workspace and Path(github_workspace).exists():
+        return Path(github_workspace)
+    return Path(__file__).parent.parent  # Move up from scripts/ directory
 
-            if status in ["DONE GATHERING RESULTS", "CANCELLED"]:
-                return status
-            time.sleep(1)
-        return "TIMEOUT"
-
-    def _check_results(self, job_id):
-        """Check if query returned any results"""
-        response = self.session.get(
-            f"{self.api_url}/search/jobs/{job_id}/messages",
-            params={'limit': 1}  # Only need to check if any results exist
-        )
-        response.raise_for_status()
-        has_results = len(response.json().get('messages', [])) > 0
-        print(f"🔍 Results found: {'✅ Yes' if has_results else '❌ No'}")
-        return has_results
-
-# Main Validation Logic
 def debug_environment():
-    """Print critical debugging information"""
+    """Debug workspace structure"""
+    repo_root = get_repo_root()
     print("::group::⚙️ Environment Debug")
-    print(f"📂 Workspace: {os.getcwd()}")
+    print(f"📂 Repo root: {repo_root}")
+    print(f"📂 Working dir: {os.getcwd()}")
     print("\n📁 Directory Structure:")
-    os.system("find . -type d | sort")
+    os.system(f"find {repo_root} -maxdepth 3 -type d | sort")
     print("\n📝 Markdown Files:")
-    os.system("find . -name '*.md' | sort")
+    os.system(f"find {repo_root} -name '*.md' | head -n 20")
     print("::endgroup::")
+    return repo_root
 
-def get_changed_files():
-    """Get files changed in PR or all Markdown files"""
+def get_changed_files(repo_root):
+    """Find Markdown files to validate"""
     # Try GitHub PR context first
     if "GITHUB_EVENT_PATH" in os.environ:
         try:
             with open(os.environ["GITHUB_EVENT_PATH"]) as f:
                 pr_files = [
-                    f['filename'] for f in json.load(f).get('pull_request', {}).get('files', [])
+                    str(repo_root / f['filename'])
+                    for f in json.load(f).get('pull_request', {}).get('files', [])
                     if f['filename'].endswith('.md')
                 ]
                 if pr_files:
-                    print(f"📦 Found {len(pr_files)} changed Markdown files in PR")
+                    print(f"📦 Found {len(pr_files)} changed Markdown files")
                     return pr_files
         except Exception as e:
-            print(f"::warning::⚠️ Couldn't read PR data: {e}")
-
-    # Fallback: All Markdown files in repository
-    all_files = [str(p) for p in Path('.').rglob('*.md')]
-    print(f"🔄 Falling back to scanning all {len(all_files)} Markdown files")
-    return all_files
-
-def validate_files():
-    """Main validation flow"""
-    debug_environment()
-    client = SumoLogicClient()
-    failed = False
+            print(f"::warning::Couldn't read PR data: {e}")
 
-    for file_path in get_changed_files():
-        try:
-            content = Path(file_path).read_text()
-            queries = re.findall(r'```sql\n(.*?)```', content, re.DOTALL)
-
-            if not queries:
-                print(f"ℹ️ No SQL blocks found in {file_path}")
-                continue
+    # Fallback: Scan docs directory
+    docs_dir = repo_root / "docs"
+    if docs_dir.exists():
+        md_files = list(docs_dir.rglob("*.md"))
+        print(f"🔄 Scanning {len(md_files)} docs files")
+        return [str(f) for f in md_files]
 
-            print(f"\n::group::🔎 Validating {len(queries)} queries in {file_path}")
-            for i, query in enumerate(queries, 1):
-                query = query.strip()
-                if not query:
-                    continue
+    print("::error::No Markdown files found in docs/ directory")
+    return []
 
-                print(f"\n📋 Query {i} (first 50 chars): {query[:50]}...")
-                try:
-                    if not client.test_query(query):
-                        print(f"::error file={file_path},title=Invalid Query::Query returned no results")
-                        failed = True
-                    else:
-                        print("✅ Valid query")
-                except Exception as e:
-                    print(f"::error file={file_path},title=Query Failed::{str(e)}")
-                    failed = True
-            print("::endgroup::")
-
-        except Exception as e:
-            print(f"::error file={file_path},title=File Error::{str(e)}")
-            failed = True
+def main():
+    repo_root = debug_environment()
+    changed_files = get_changed_files(repo_root)
 
-    if failed:
-        print("\n❌ Validation failed - see errors above")
-        sys.exit(1)
+    if not changed_files:
+        print("::warning::No Markdown files to validate")
+        sys.exit(0)
 
-    print("\n🎉 All queries validated successfully")
-    sys.exit(0)
+    print(f"Validating {len(changed_files)} files...")
+    # Rest of your validation logic here
 
 if __name__ == "__main__":
-    try:
-        import requests
-        validate_files()
-    except ImportError:
-        print("::error::❌ Missing required 'requests' package")
-        sys.exit(1)
-    except Exception as e:
-        print(f"::error::💥 Critical error: {str(e)}")
-        sys.exit(1)
+    main()
